Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
byJ. F. W. GALYER
C.El1g., For.Prod.E.
Principal L<!cturer in Production Engineering at Luton College of 'Technology
~
.
.
l .\,'
L~,~:~71/'
CASSELL LONDON
li1'
1
t"
Iy:!eltO~ogy)
r
.
OJ' .0i1gmeers
1964
1968
1969
1971
1972
LS.B.N. 0 30493434 8
772
Preface to
",'
THE purpose of this book is that it should meet the needs of students pr:~p"ring
for examinations in the subject of metrology, or in that group of subjects which
metrology forms part. It is designed to be appropriate reading in prepaf:ltion for
the foHowing examinations in particular: the Higher "National Diploma and
Certificate examinations in Mechanical and Production Engineering, and the
Associate Membership examinations of lhe Institution of MeGllllnical Engineers,
the Institution of Production Engineers, and the Institute of Engineering Inspection. It will also prove useful to those students preparing for the fInal years of the
Mechanical Engineering Technicians' Courses of the City and Guilds of London
Institute.
A know1cdge of metrology, and practice in its application, is of increasing
importance in industry. This is true also of those techniques which are allied to
metrology and do 110t strictly form part of it. It is for this reason that a chapter
dealing with statistical quality control has been included. It is hoped that not only
will this be of bEnefit to college students, but also to those in industry who require
a knowledge of the fundamenmls of this technique.
The development of engineering production in the past has been inseparable
from that of metrology. It is certain that the increasingly exacting demands of
industry for mechanisms and assemblies generally, the functioning of which must
meet stringent design requirements, will result in even closer attention bcing paid
to the science of measurement.
The writers would commcnd to the student the words of Lord Kelvin, which
were used by Dr. H. Barrell in the opening remarks of his Sir Alfred Herbert
Paper, 'The Bases of Measurement', presented to the Institution of Production
Engineers in 1957: 'I often say th,lt when you can measure what you are speaking
about and express it in numbers you know something about it; but when YOll
cannot measure it, when you cannot express it in numbers, your knowledge is of a
meagre and unsatisfactory kind.'
This statement presents a powerful case for the subject of metrology to be
part of the studies required of an engineer. It also points the way to the fact that
if we are to further improve the control of the processes of manufacture, we must
continuously develop our means of measurement. This book will have achieved
its purpose if it OIlers to students the basic knowledge upon which the necessary
further development may be built.
We wish to express our gratitude to the several firms who have been kind
enough to as~isi. in the preparation of the book by granting permission to publish
various iIiustrations, all of which arc acknowledged in the text.
I. F.W.G.
C. R. S.
v
to
COi'<SlDER,\J3LE reVlSlon has been necessary for the third printing of III!::trology
for Engineers. This is due to two facters. ~yilith the adoption of the ,>vavelcngLl1 of
light as a standard of length, the original Chapter 1 has become of historical
interest only, and it !JUS been replaced by a new chapter 011 errors in measurement.
The other Inajor I.:hangc has been occasioned by the adoptioll of the rr:e~ric syst<:m
in Orea"t Britain, and particularly the use of the Systcmc Internation;}! d'l!nites
CST units). A(;cordingly, the whole; work has been amended to embrace the system
bused on th,~ metre, kilogramme, newton and second, as the units of length, mass,
force and time respectively.
Ch;l:lgc3 in the structure of technicai ducation have also taken place; in the
last few years, with the formation of the Council of Engineering Institutions, the
revisions in the Higher N<:ltional Certific:1te courses, und tile continued growth of
the City aed GaUds of London Institute's course 1'io. 293. It is hoped that the
revised text will be found useful to student" following courses leading to the
examinations of these bodies, as well as the Higher National Diploma students
for whom the work was originally intended.
Since this book was first published a number of people have spoken kindly
of it. Sl)rne hi'v,; pointed out errors and others made constructive criticisms. To
all of the;11 our thanks. Vie hope that the revj~ed text meets with th0ir approval as
did the
If it helps to
the young
of today into the metric
world of tomorrow, then its purpOSf.: wiil have been served.
vi
J.F. W.G.
C.R.S.
Chapter
Errors in IJ,feasurement
Page
12
3 Linear Measurement
31
67
93
115
Gear I'deasurement
127
156
176
19i
Appendix to Chapter 6
220
Bibliography
229
Index .
231
10
Vll
CHAPTER 1
1.1 SCOPE
ALL engineers, regardless of the branch of the profession to which they belong,
are con~tantly faced with the problem of measurement. f.t may be of time, rnass,
force, temperature, the flow of an electric current, length, angle, and so on: or it
may be of the effects of some of these in combination. Almost inv!.!riably, the
results of such measurements will determine the course of action the engineer
takes thereafter. Thus the results obtained by measurements provide information
upon which decisions are made.
All such measurements form part of the science of metrology. The m<;ehanical
and production engineer are, however, especially concerned with the measurement
of length and angle. Of these, length is of fundamental importance since ang'.11ar
measurement may be carried out by the appropriate use oflinear measurements in
combination.
Thus the purpose of any meClsurement is to provide a service to enable a
decision"to be made. The service will not be complete unless the measurement is
made to an acceptable degree of accuracy, but it must be realized that no mellsuremcn1 is exact. It is therefore necessary to state not only the measured dirnension,
but also the accuracy a/determination to \vhich the measurement has been made.
As far as possible the errors inherent in the method of mcaSln;ment used should
be kept to a minimum, and having ;Tlinimized 1:.'le error, its probable magnitude,
or accuracy of determination, should be stated.
," It follows that if is not enough to state that tbe nominal size of a gauge block
is, for example, 30 mm. It is also necessary to state:
The user, having this information, may now. avail himse1f of it if necessary.
If the gauge blode is used to set the datum for a vernier height gauge '.vhich can
only be read to 002 mm, then the gauge block errors are negligible and can be
ignored,,If on the other hand, it is used to set up a comparator whose scale
divisions represent v'OO1 mm, then the measured error is significant and must be
considered, and the accuracy of determil'lation of the gauge block must be incorporatedjn the accuracy of determination of the comparison being lx:ade.
Errors in
It must be pointed out here that in this chapter reference will be made to
subsequent work throughout the book.
l'yfeasurement
(a)
tIll
of 528 mill.
ture produces a
where
error:
whose expansion
to a nOH-standard
As the expansion cocff1cients are small Gumbel'S, the error will be very small as
long as b~th parts are at the same temperature. Thus in comp8.nttivemeasnre-
ments it is
that all co.mponents in the measuring sy,;tem arc at the same
temperature rather than necessarily at standard temperature.
Other ambient conditivl1s may affect the result of a meamrerncnt. If a gauge
block is
Fig. 1. t.
pressure and carbon dioxide content of the air afJ"ect the refractive irrdex of the'
atmospher12:. These condi tions should all be recorded during the lest and the
necessary corrections made.
L
(a) W:E STANDAilO AND EI~D BARS.
SLOPE AT ENDS ZERO
1.222
Errors
How accurately can a sc:~.le be read? This depends upon the thickness of the
rulings, th,~ spacing of the scale divisions and the thickness of the. datum or
pointer used to
t11e
.
As a
a reading of a pointer or datum hac against a scale division can
be taken as bowing an accuracy of 10 %of the scale division. On the other hand
the estimate of the position of a pointer between the
will be less accurate
and should be taken as ::20 % of the scale divisii)n. Thus a reading of -- 3 units
taken otT a scale whose di visions represent 0001 mm '''{QuId represent a comparative meaSUI\;mcnt of - 0003 mm to an accuracy determination of 0,0001 mm.
for Engineers
in
however, the rC:.Jding had been - 3-4 scale unitl>, then it would represent
0-C034 IThJJ_ .-::::O-0002 mm_
It must be rC:llized that when a measurement is made with a comparator
this tVDe of error occt!rs twice, first when
the instrument to a master
gauge' ;nd again when the reading is taken on the
1_223
where
kleasliring Errors
The diJfen::nt types of error discussed above are cllI11ulati.ve, and in c~ome
cases further amount THust be added to allow for semilivity of touch or feel.
This will depec1Cl upon the type of instrument
used, and in genet;]1 the effect
is eliminated with cOllJ.parators.
"
Consider now the problem of
the error in a plain plug gauge of
nominal diameter 25 mIll. The measurement is to be carried out l1:,il:g a coma magnification of 5000 x which is set to it gauge block of r:iominal
25 n1i,l
a known error of 00001 mm to au accuracy of detennination
mm. The comparator
011 the gauge block is 0 scale divisions
and on the plu,~ gaw!c - 1- 2 scale divisions.
1n this cas~ tlle ~iE~ct or' ebstic deformation can be
as the two parts
arl~ of :.;imiiar material under similar pres:>llJ:c,
conditicns of contact
are slightly dilferent. The problem can be set out in a tabular mann,;r as follows.
Amount or
Reading
Accuracy of
Deterlnination
G8.Dge block
Cml1i)Hrator setting
{~\)Inpi.lrQtO[ reading
0-0001 mm
0
-0-0012 mm
OOOf)2 nml
:*:0'0001
Totals
{lOOl3 m...."'1l
-J:::O-0005 mm
Element
(Jilt!
()A1.IS t.oe
1
I 11 '" f
parLia
Cdflercuna.!. 01 l~
T
.'
V/!t 1
/v.{easuremen t
is better solved
etc.
to a, al1 other variables
to b, etc.
ShO'Nil
D=L
Let L=400 rnm and 8L= O-025 mm
ISO? ) _
[_\ 1 - e
R .:1002 0 025
X.!
1
mm
mm
_00002
Thus the g'lUge siz~;s fmEld to b:: 249987 null but the ar:CUfHCY of det';;~111ina
tion shows that it can be anywhere between the values of 24-9992 mm and 2,1,9982
mm,
dD
[V-024+0'V31] mm
i': a functioll of a
accDr:lcies of cic:tcnnination
etc.; then the
of detcrtl.unatlvi1 of
kI, which we can denote rL\'f, could be found
the expression for
AI the maxim.um and mirltmnm values of a, b, c, ~tc., and :hus .flnding the maximum a;lt2 ~ninimum values for AI.
:11
.~
i
I
where
obser'fations is
c:lI1.;lOi
1,111
,d,!)
I .~-GRAPHICAL METHODS
Tf . n experiment is carried out to find the law relating two [nc:I:,lln:d \;lr;,thh"
1 a d
it is usual to plot a graph of the readings a.nd dctCfllllll:: In,; Lt"I. ,)1 I !lI;
1mc ;.~ we are averaging o.'ut t.hc errors til 11lt: 11:-1 IV. 11111:.11
xorap
anh 'oYy' piot
amea'1
.'.t l'""
""J.e
l
"l~""'"
I
r
'
~bservations- Usually the observations arc m.ulllpulated to glvc ,:1 ::i 1':11"; II .we
mph of the generallr,.mn y = ax.+ b, whe~~ a .IS the slope or, g::~dl;.:Ht ','[
11l1~:
g d 01 is the intercept 011 rhe y aXJs. If the lWtflS drawn on thl~ lj.I.!P,: ,Il .1 pu:,liIO 1
an ' 't d bv ~"e tlhen ttis will still be lIa
, bl c to error. J."'f a't ;;;';Ie.'n 17(,"Iflt ,t r I~~".
"'1"11,,1 0
eS t Ima e
''-J'
'
.,
.
.' j
1 .
A
.1
'-
Ill:',
",
's drawn r;prcsenting the accuracy of determination of the l!Jdr/J( 11:1 ()I):.crv,:~ions then two lines can be drawn through the extr~n1es and two .1aws.caJcu:n~c(l,
these' o:iving lhe limits aftile accuracy of determin:~tlOn of the def1ved';~lw, i lw; i.e,
obvio~sIy a tcdious process, and a better method lG kno\vn z.s the memod
squares.
1.51
M(~HlDd
of .Least S(tU~1n~s
C;.msider the slope a of the graph. This may be expressed as the average increase
in y for a given increment of x, and it can be expressed as
3cr
Statistical tables ,; no,v that a pproximatc!y 95 ~~ of all observations lie within
2u Df the mean of the observations, ;lnd appro:dmardy 65 %lie within the limits
of I u of the mean, Hence we can now state the confidence with which we give
the accura.;y of determination. Let the e~[imared accumcy of determination of a
single observation be :i:: S. As tbis represents 30' we call say that we are confident
this accuracy of determination \viII hold good for more than 99 % of aU sllch
observations. More simply. we say tl;.at :;: 8 represents the 99% confid,:mce limits,
Similarly,
giving 95 ~~ coniidcnee limits and 1U= 18 giving 65 %
confidence limits.
If we ~.pply this {O the mean' size of 11 obser'l<ltions, we see that
95 %conl1dcnce limits
L'x(Zy)
... (1)
n
Having found the b~st value fO,r a, the best v~!ue for b can be found by substituting a~'erage values for x and y In the expreSSiOn
y=a.x +b
.,. (2)
~_
where
65 ~~ confidence limits
To do this with accuracy "ve should have to tal<c a j;;'.rge number of obswlations and from them calculate the true vulue of the standard deviation, but by
and
~x
X=--
The problem is best set out in tabular form,.and from the above e:pr~~~~~n~
we see that we shall need coiumns for the observed values x .~nd y ,Lid H'".::,:~I
f or x 2 and x)'. _rih"v 'otal
values
?',;
~y2
and .Exy
WIJ also be requhC<l.
columns
t
~;,
:
Consider the experimental values for x ann y below:
9
B
3966 _ 431
91
819 -
37
Substituting for
a=329
91
x=-=7
13
from this
the measure-
,. A. J. Scurr, Pr!Jc. I. Meek E. VoL 32, part 1., no. 23. 1967-3.
/,
Errors in Measurement
10
11
CHAPTER 2
It is this definition which has been universally adoplcd hy ill,,"\' ,(lI,lIil" Ii .11' '
or intending to use in the future, Sl units.
Clearly, a universal st",ndard must be one which i;; Ii'pr, "Il/I"d I., '.\.,/' I .. I. '
degree of accuracy tbat foi all industrial and sci,;ntilic 11IU'Pl>:,'::: 'I 111.\: I,,' '\'1'
side red as absolute. By means of interferometry, r.h~~ GlTlJr 01 n'I'C".! III" 11<111 I'! 1'1'
metre is of the order of I part in 100 million. ~")imilady. ~'J'V ,;Iillth-c.",,! I'i ,I,I'
metre may be;produced, and reference):o B.S. 888: Sill! (')1: !ilPI''') : "/1":": .II!,/
their Aceessorif:!! shows thatthe"practic'al workin!:. SWlldilnl" PI' 1"11:'[:11 ',' '\ III
industry-u):e-6t sur;h accuracy that tlle'caiibratiolJ and 1"'fcrcIHT !'II'd.'i. ",I i!l"",;
must be verified by iI:terferometI'Y, that is, in terms \.1f Ull; \VavL:!ell:'il! "r li:,!it_
- ,\I '
,',
It_ is fupdam;eqtaJ to the science of measurement, and hence the degn~c of control
which it exdts on the development of technologies, thut it should be based
an
, ,,' '::agreed, aiXan- possible internationally agreed, system of stancJards:-~Foi many
yeaffthe major industrial countrie:; of the world used two syst;;ms, imperial and
on
2.2
InterteromlOtry is th<:t branch of scienee which xc; concerned will! lhe Illi\%Cr in
which ray;,: of lightl product~d from a common som:cc, ure n:l.:on1 I;incd
~! kn~
system, usu:xlly the eye. The dH-rerg!lf_~jl1 ~patb length" ah:)i1g whieh the ray:; I.f:t'ld
before being !'ecombined a.ercrnlines thclr-rdationsilip, and 11CIK~ We
sensa-tior:
of lic:htthceve.
-an unck,srandlng ~f th.'! pllen6;n:-ieIrtl associated with interferometry, it
is necessary to :,::ol1sider the nature
'
Two theories have been advanced to expiain the nature oflight: the Emission
Theory, and the Wave TJlcory. The former was adv~nced by Newton, and considered ligl1t as Gonsisti!lg of particles emitted by fuminous bodies, the in; pact of
the particles' on the' eye causing tEb-sensation of ligbt.
.
The wave theory, however, was advanced by Huygens, 0.nd consider\ld light
as a \Vt\v:; rtlml0ll ;}foDar.:ated in the erhel', : l'
It is this tbcOfY" ,,"nlits subsequcilt ;:ievelopmem, wilicb
explains
the phenomena associated with light, including that of interf;;-rence.
If, then. light is cou'sidered as <1.;} electromagnetic "lave of sinusoidal fcrm, it
may be represented as in Fig. 2. L
'
'_""r
Thf~ disadv:11:tagefJ of SUC!:I an arrange.ment are evident wben on~ (;o!i~ide~s
tho.t a very iarge -part of the world's populatio:1 :used metrIc'unlts, biifthal impdrtant industrial. coumr!Cs, notably the UniTed K.ingdom and th,~ U,S.A., used both
imperial and metric units, the'former being domiriant in the industrial fidd. Vir~
tunny the only concession made by these countries was that scientific work was
carried out in- metric unlts.
." I
' , .
'The basis for a solution to this confusion was establIshed
1960/when the
General Conkrenc,; of 'Weights and Measures, an internatiOnal body, recom"mended that 31 units should be brought into use to replL~Ge
metric units.
Sl is an abbreviatlq!l of Systcme lr;.ternntional d'Unitcs (bter:-w.tional SYstem of
'Units) whicn has gro'vn out of the l\IKS
kilogramIac. second) sy~tera and
the MKSA
kilogrammc, second, ampere) system, The major industrial
countdes, induding Th0se at present using u metric system, ha'i'e adopted the
, recommendation. Thus the United Kingdom is at present in the process
con~
ve~~~..::'1 from imperial to SI units, The proc;;:ss will take some years for its
completion.
'
1'11<; standnrd of length, therefore, will be the metre, and for the purposes of
this book win be the most important of the SI units considered,
metric.
--'-Por
in
'of
2.n
NATUHE OF LIGHT
As part of the cvolutioll of a universal standard of iength. the International Committcc'()fW'cights ,md Measures recommended in 1958 that th.; metr;:: be defined as
165076373 x.\
where J\ =1he wavelength, in a vacuum, of the orange-red radiation of the isotope
krypton 36
Fig, 2.1.
12
t
.I
13
"umber
of Viaves of eC1iual wavelength, the value of which determines thr~ coiour
.>
-
ofth~yght.
.,
_,'
".
LWO
such
~, ,
layS,
we may
' "
GO ,0
h,
uy
quency = l / T . ; ,
For any given monochromatic Jight source, these characteristics arc virtually
independent of '-lrnbicnt conditions sllch as temperat1.m~ and pressure.
if we now consider, for Gxo.mplc, the w;c of t11;; green mcliation from the
'spec.trurG of TI1Crcury 198 -1.S a lTI.OIlOchrornatic light source fer The i::lbs.DI:.lt~~
iF'
IJ'lill~1..
.,'::l..L
SO'l['('
.
t!. l .....
'
~ --~-r,"
\1 \J
.-
or
!-
to consti.tu le
;J.
I,! _ _ _ _
,,; -~'.
idJow
i Orange
I P,ed
19,3-,22-6
226-236
23 '0-25 -4
2:5 <!--,275
I
1
~~-;-
f,"J
O423-0,'~90
O"~SO-Oj75
0,575-0'(00
O'oOO-O-ein
Fig, 2.2.
Resultant amplituJc ar, of two waves A nnd B of dilfcn:nt amplitudes ale and (In, but in ph'elSe.
3ta.r~dal'd of
IViolct15'7-l67
~lru,.'"f.""~"'l
16.7-19- 3
I
II
aC8Uf(.!te
"
two rays, A and B, are in phase at their origin 0, and clearly will remain so for any
distance of propagation.
If A and l-3
amplitude
!18,VC
Cl2u=U.:\ +-0:8.
..,
In ofb.<:;r "\'ards, whea A and B are of the same wavelengtl1 and :n prllsc, tney
i ' anl~.-l t he[eJOl'~ tl'lJC>'111lrv
"'J"l-ty
0+.1. '1",
cOIJ1bine to jnc[ca~c' .
tne 2.mpltuue,
.l~) ""
l.u. . . . .l"f'r'lit"rt
,-"H... Llu. i
C
0-643-0698
R to
2, l1l2ximum.
"
, n
If now 1ve cOilsictcr the effect of changing the phase rdaLlOl1sIllP of A an~;~
byt !)e .''Yr,r'['!l''lt;S qc.' l'r: !=';r; 23 it carl be sho\vn that 'when a_\..;c.-;;::{[lJ, ..71~=2a cos, {~/.I~.
That ic;, tbe combination of A zend B no l.onrcf proJucco maximum illumina..
L
...
,_~Lll.,-,'
..
"',
{._~
. , ... -
.
-.
(' d
Further. if we consider the effect of changing thl~ phase rdatlOilShlP 01 Pc an'
D
t1l'ey
"[C"
clicl"J"'lc"d
l' (o'oa
O.
,\/2
then a condition 2,S in Fig.
2/+ occnrs,
J...J, SO t'll""
1 ~ ... l
l
I..~
_0 .'.!.......
-,
~
"
....
1
in which R has an tiil1pIitude \vhich is the algebraic sum of tzA and {lB ~,n~ IS
. reduced to i minimUll1: It is important to note that if aA ar-d an were equal 111
value. then al\, would be zero, as cos (180/2) ~;O. That is, complete interlcrcnCt;
betw~cn two waves of the same wavelength and amplitude produces darkIless;
and no sensation of light is regisw,ed on the rctit",- of the eye.
It is the abiJitv t~ split light from a single S(;1J[CC into two compol1ent TJ.YS,
to rec~moine them: and ;bser~e the way in .,;hich they recombinc, that allo\vs the
tion.
l'nnt the prinEtl'Y colours 11rrv(; sneh lH,defined ".vavt:~L~r:gtb.') 13 the principal
reason for tbe int('l!sive (:Jr'vrts 111J.ci::: by pl.i.ysicists- Qver l"nany yc~~rs to prodl1,ce
pure DO:l()c\Jromatic li;;:ht :;ac:h as that fr()iil merC1.li:y J9d en: krypion 136, havin8 a
l')rcci::;~~J Ieprcducibl~ vi/ftvclcngt~l.
L"J
j\lONDCHJRGMATIC l<AYS
(t
..
14
flleftologY}(H Atlgineers
I ~J
Fig. 2.3.
Resllltant amplitude aft of two \vuves A and B out of phase by an amount il.
Fig. 2.4.
otil
of phase bv
-
the two rays will arrive at 0 1 in phase, and will cornbine in a manner simihu lu
that shown in
2.2 to give the maximum illumination at 01'
If ,'ie now consider a point such as O 2 on the screen, the distnnce ;\0,\ is
clearly less than the distance B0 2, and if B0 2 - AO:) is equal to an odd number
of half w<"'idcngths, thitt is (217 -I- 1) (A/2), where n is an integer, then the waws
will be 18(Y out of phase, complete interference will occur and there will be dark
ness at this point.
if we consider a point such as 0a, then ifB0 3 -AOs is equal to an e'ven
number of haIf wavelengths, that is 2n (>./2), then the rays are again in phase, and
no iutcrferr::nce occurs at 0 3 , this point receiving maximum illumination.
This process could be repeated lH points both above and below 0, and would
result in alternate dilrk and light areas being formed. The dark areas would occur
wherever the path difference of A and 13 amoun:'ed to an odd number of half
wavelengths, and the bright where their path difference amounted to nIl even
number Ot half wavelengths.
It must be emphasized that the phenomena descdbed would occur only jf
the sources A and B were images of a single source. This can be achieved with a
Fresnel biprism by which light passing through a slit is divided into two identical
and equally spaced images, the rays from which,
from the same sotlrcc\
will be in phase at the images. It 15 the dil'l:crence in path lengths of the subsequent
rays which causes interference.
Summ.arizing, it is clear that the following two conditions are necessary fer
light interference to occur.
(a) Light from a single source must be diyided into two component rays.
(b) Before being combined at the eye, the componems must travel ;;aths
whose lengths diff.::r by an odd number of half wavelengths.
16
17
Light Waves as
man~facturing
~~Od~~~,l?n
,.,.
.'
.-
problem
frcquent~y encouu"tered
..I.
>.J
,.
~...
AJ'..I
An optical flat is a disc of stress-free glass, or quartz. One ;)1' both faces ofrhe
are ?round, Japped rmd polished to a degree of flatness llGl normallv en.
cou.nt~reC1 on an engineering sl:rface. For engineering pnrpose'), thercfor~, the
PHC::H flat may be consklen~d as a reference of l1atness, and l,lsed as such for
t:~lmparing engineering surfaces. Optical fiats v:lry in size: from 25 mm di,tmeter to
a:JOllL 300 mm diame::!er, t11i~ thickness being abolit 50 mm fvr the
In all
C,lses,
are relatively rigid and stress-frcr~ disc:; which, used antI stored corJectly
V/l'l1 r~."':p t l -,,!'- flat . .
.[;
e
' .
,
,
" ,,,Lu.L., -d~ , "~ 11C,55 an:- ttlererore
:llfllOst mdefin;cely.
.
,~f an 0pueal nat lS laid (ilot 'wnmg') on to a nominaIly fbt
surfac(~,
It wulllot term an intirn"tc contact but will Lie at SOme an"j" [)
I'll
'lor"
in which f) is greatly
j~ wedge-shaped 'lir cllshf~:1 n~;'1 hi; rorr~~d
between the surfaces.
dISC
of Length
If abc = AI: where ,\ = wavelength of source, then the conditions for complete
interference have been satisfied, i.e. the rrxy [rom S has been split into two components, and recombined; also, the path lengths of the components differ by an
odd number (one) of half wavelengths,
If the surface Hat, then at
to the plane of the paper it will be
parallel to the optical fiat, and these conditio.Gs wiE be salistled in a straight line
across the surface. Thus a straight dark line, or interference
be seen.
Further along the surface, and due to
0, the ray
S will similarly
split into two components whose path lengthS differ by an amount def.
If def"~ 3A1l, the next odd number of hdf wavelengths, interference will
again ot;cur and a similar fringe will De seen.
, At unlntermediate point, the path difference will be au even, number of half
wavelengths, the: two (;omponents will b,:; in phase, and a light band will be
wm
'>,'
Thus, a surface will be crossed by a p-3.ttern of alternate light and dark bands, which
will be straight, t'o:' Lhe case- of a flat surface,
as in Fig. 2.7, A dcYi2tion from vel,""",,,.,,
would b,~ a m~::isnrc of
error h-::. flatness of
the surface in a
p~irallcl to the apex of
the angb 0,
Referring again to Fig.
it is seen
that if 0 is small (which it must be),
a.b=bc
and
de
The
in
sl~paration
ef
between the
3'\
If this ,'1lTangement is l:vW placed in the path of .:! paratld bC(lm of monowe en,l C;)!1S1(ter S :lS the source of on(; W9,ve of the incidc:u beam.
any reffc1Clivc eilixts due to the light
tJlfCllh:h
glass. and
it j'; seen that th,~ wave i'w.m S is '.
r:~f1.:;cted at
transmitted c,cross the air gap, to be ren(~c(ed at 'b', a !)oint on the
The two reflected compol1l:nts ate collected and recombined bv the
traveHed ptlls whose lengths ditfer
an amount abc.
J
18
19
under test, the spacing of the fringes reprcsellting Il!'il'ilt 1'1\<'1 v.Ii'. 1"I,t" .. I,' II:,
optical flat of N2. In
2. TO, point C is the same d i.',LlIll" 1'1, ,j II Iii .. "llI'I ,I II II I,
point B, but /,,/2 farther (or
than point D. Thel,:I,;ll' II:.' ,','''. :11 ( , , ..\
higher (or lower) than D. If the fringes curve toward:; 111\: lint: ," ""nl.l1 1.11 \, II..
surface is convex, the opposite case also applying.
Practice in the use of optical fiats is essential to a lrm; lIIHic-!':,LII.d!'I" ,.I Ill,'
patterns Jw")duccd, and at 'this point it may be appropriate 1.<) indi,,;\k 111:' p"II I I.
to be observed in their use.
(a) Handle optical fiats carefully, and the mini-
'1
1 'III1
. , 'lIOln pomt
.
'7A x n = 1
otal
c.mnge
e evatlOn
mumamoum.
(b) Ensure that the work surface and the optico.l
flat arc perfectly dean, by careful ,;iping
with a cloth of the 'Se!vyt' type or with
chamois leather.
(c) 1'.fever 'wring' ,m optical fiat to a work surface, It should be kid on, so that the flat is
not distorted by lending to 8.dapt itself to
the contour of the work
thus producing a false fringe lxUlc;r.
(d) Never 'wring' two optical fh.,ts together. Separation may be difficult and cause damage:.
served
'Nh~n
opticai
To distinguish between th,ese two conditions come i1arrow~r (md mOf() closeiy
spw;ed.
if, when the surface is spherically convex, one
cJge of the optical fla.t is lightly pressed, it will
rocle on a new high spot and the centre of the fringe pattern will move as shown
in
2.9, and the outer fringes move closer together.
Also, when the surface is spher1cally concave, the fiat rests on a l.ine passing
around the surface, and if the edge of the optical flat is lightly pressed, the edge
line does not move as the pressure is 'laded. Alternatively, light pressure at the
centre of the optica.! fbi wJJl callse it to deflect
<me. bccorn,e more
lXlfallcl with
the concave
thus reciueil1g the number
of fcinges observed.
Commoniy, opticalllats are used in normal
daylight, the spectrum of which has a waveler:gth of approximately 0,5 p.m. 1hus, each
fringe interval correspoads to u ehange in ele",<,~
tion of the sllIface 01'0,25 p.m,
SUjnose an
tb.t to be, (,tid on to a
G
t......
is that
"11f
''''" ; no' t11" 1"~~',1 'li"ll
0,
b
shown in Fig. 2.10.
H;lving first determined rhe point or line of
cont:ld of the optical
whkh i:-; assumed to
be at A,A, it mmt bc remembered thfJ.t the con"
tour of I~ac;,
lks on points of ,'nual height
(in a negative direction) relative to the s" . ~'qcc
of the optiC;}! fiat. Thus, the
pattern, in
Fig. 2.9~ ivicrhoJ of testing to show fact, represents a contour map of the surface
~d,J;.
that a surface is
20
t,..
._
.....
Fig.l.l0.
palten',
fonned on a
wh(~:~~
C.L. is half a \Vil\'elength
higher than its edg<;s.
2.5 INTERFEROMETERS
AlthQ.u.g1L oPtical fiats can be used in either
or, better, in a diffused
source of JlCar-mOl1llchromati~ light, e,g. a light box consisting of a sodium disch'arge lamp behi!id' a yello\v nlt,ii;""ihey sutTer the following disadvantages for
very precise work:
'
(a) It is diJTk:llit to control the 'lay' of the optical fiat and thus orientate the
fring~ to the best advarllage.
(b) The fringe pattern is not viewed from directly above and the resulting
obliquity can cau~~
itUd,crrors in viewing.
v'"'!"*_~
This instrument, shown in diagrammatic form in Fig. 2.11, wcs designed by the
Nationd Physical Laboratory and is manufacturcd commercially by Cov~nt;:y
Gauge and Tool Co. Ltd., and Hilger and "Vatts Ltd.
convex~
"
'f
21
VAi'OlJR
SOURce
LENS
an
PiN
D'~
'
HOLE
parallelisrD
bf~t\ver)n
d )
",,~j.i
tC
/'':'"
~J
~h\)rter
~md
the
parallel
fringes
the fringes on t11e
If the
the base
is
is
to.ble is turnd
Fig. 2.12 (b), ;) ncJ
0(1
b:t~;e
PL."TE
REf-LECTOR
LENS
FLAT
(~onsidGr
TO BE TESTED
on
gauge
in
2.12 (a). If lhc
with the optiGal
J)umb(,I sf
along ils
111 Gnc
and 18
in the SCC'cl1d
In
2.12 (a) the dis1.~Lnce b\t['.~!Cell the gnt1ge and the opticalllat has 111creased by ~i dLt:,nce
over li:c
of the gauge, and in the second position
[Fig. 2.12
It has
changes by
;).. disla.ncc
Therefore! 61.
G\AIJq~
(ll,d
fkit
C-<ll!g<;;:
pctQll<l.i
c,)nv~x
'Or cor.cavl!
'.(T;r4
The
Fig. 2.11.
Op~:,.::al arrailgC111cnt
(Co:lrlesy
be
in the
)~_auge
rOI
_J-
A~-18
GAUGE
- -
I
~ROTARY eASE Pl'~~
(a) 1ST POSiTION
I! I
But due to the rotation through 180 c there is a doubling elTcot. Therefore the
_II ,
! Ii
!I
error in parallelism
.1
8 3 <\ 8 A
"'~=2x2
,\
42
T;ms the gauge has an error in parallelism of 10 micro-metres over its Iengtll.
2.52 'l"'l'!e
Pitter-N.P"IJ~
Gauge Interferometer
to
or
paim source, 1n
by lens .J
OIl
tv
th~
24
c
Fig, 2,14.
l'-J.J).. L.~lype
gaLlf~e
intcrferon1.;r.er..
,.\ a t,
i.e. G=H -h=7Z,-2.521
]1'!etJwd ~f jJ,feasuremellt
Red
(}":~67
81743 fern
2, 14
shov/s the
phCefi, slip gauge) and
ship,
lUent. F or each of i~be
vt"ill be cbse.r(!ed..
n. plus the fraction
(~re
26
obscr'/cd.
Of.
We thus have observed fractiom adb 1, Cl z/02' o:;/b 3, for each of the radi;:niol1s,
which may be expressed as
respectively,
of; series or expressions may now be obtained for the height of the gauge
above the platen.
,
1
,"
U'
"
'-'
" "
"
arJ:~tngclllcnt,
3932 d30 64
4275,[7205
"
"
"
"
wb.ik
'Nllerc
radia1k)TI [!,nd
the hdf
d{.h,~
'0
fl
27
GN=~((n-N)+(f-F)l
Assume that the observed fractions, f, are /-; =0-23; j~ ~~O-33, and j;=0'71,
and that the calculated values of Fare FI =0-94;
=0'44; F3~=O-52 for the three
radiations red, green, 3.nd violet respectively.
[nserting this information in the above eq:mti;;ns we ha ve:
0-467
=--~-. [(n3
- N~) +(0'19)J in
L.
{n3 _.
~~3
a closely similar resdx ror all three equations is found. However, tUg is laborious
and [t better method is to set the information out in tabular form as follows:
,.,
;)
7
J\1eall
Wavelength (A)
f-Lm
R,~0643
error ill
Obserred tVl). of ,\12 in Calculated
I r:"actirJiis (f)
]Omm
Fracr/o71S . li eo,. 2._('01.
023
9318-9354
094
Gauge
Coincide
L
Length
029
I,
Z'29
_i
Scales
coincidence nt the values shown in column 6. The lower scale of the slide. rule has
through the valttes
graduations corresponding to micro-met'.'es and a tine
2'29,289 and 3-19 OG the wavelength sCales cuts the
"calc at 0/14 ,am, the
error in the
or ell\:: gauge.
It is foune. in prac:tice that the values off c8.nnol be r~ad to an accuncy
greater than about 005 of a
spacing. TI~is accounts for
des in the fractional parts of the values shown in column 6, ('()m1;""",ri
fntctions sbown i.n column 5. Since the linCe'lf error re~;ultil1g
tiona! error of this order will always be much
it can
.,!t ShOlild be noted trHit the Jurf
scales are llct ll, i.c. the \vholc nUlnbcrs arc
positive. Thm; considering the cursor line shown to the kft of zero, the
mcnts arc respectively:
Red
Green
Violet3'92
0-=0508
033
III 796441 92
044
0,89
2-84
V=O467
07}
'28255165
0-j2
0-19
3-[3
.28
lengths of red, green, and violet are set out to scale, [TOm a common zero. The
values obtained in column 5 of the table are found to have a close
of
0-74 fHll
-2+0-09)= -1-91"
(-3+0'92)=-2-08"
"
"
Thus the gctugc error is equal \0 those values multiplied by the respective l:aU
wavelengi.bs.
29
'/
I
'/
r-l.52x---Flm=
.
01143
-0488
l
. '. Error =
Mean error =
')
JI - 191
l
-
am
L..
><
0'5e8
m
-----z--fL
CHAPTER 3
f
=
0467
- 208 x ---,t,m= 2
0486 }-Lill
0 -48 5 ,um
l~'~c-r
()'486/-LlU
Note that the scale cannot be read to three decimal places. In the authors'
opinion it should be used for establishing the position of coincidence and all
'approximate solntion. The 8.ctual errors should be calculated, as :;hown above,
for each ,;v8.vclength, ,1l1d then. the mean value found.
The important [,v:t about this method of mca3ure!11.(~nt is that the Q3.uf!e
length is established without reference to any physical standard, and only in tcr~ls
oC the wa'idcn:;ths of tile variolls mvnochromatic radiations employed.
3.1 GENERAL
WE have been concerned in previous chapters with the met.hods by '.vhieh engineering sta.ndaros of length are cstabli~hed, and \7v"hy such standards are necessary.
This field is outside the normal scope of engineering metrology, and is more
properly Ylithin tnt; pro\'inee of the physicist.
Such \york, however, is tl:e neccss~ry b8.sis of Lhe lincat rneU8urenlent carried
out by the t;nginccf, siDc,:; etlmost invari!lbl:; tIus takes the [o;'m or comparing tIle
size of a "0/orkpiecc or other part with the :<:il0wn size of an end gauge, i.e. cornparative lr,casurement.
To CilFY out sach rneasU! ements suCC~SSfilll;', that is, to the order of accuracy
required, often calls for ingenuity in the use of relatively simple equipment. It
always requires the application of certain simple principles, tn:;ethcr with patience,
a systernatic approach '(o a measurement probk;-i1, :md. the uc;) of techniques only
acquired by practice. These points ~-l!"e perh:lps best illllstm~d by the apparcntly
simple task of establishing the size of a plain gap gauge. There caG be no substitut:::
for experience in c2tIl"ying out such a measurement.
These are the engineer's practical, workin;S, standuds of lcn;sth, and are rell1ark:able for si;vf;fal rerlscns.
The following t:J_bl:~ is extracted hom B.S. 4311, and specifies GlC aCCUi"2,ey of
dimensions, llatness <tnd parallelism of slip r;3.ugcs.
Th01;J,ble does nal, ho\vever, reveal the c:oTIlpletc charn.c[cristics of slip g~.luges.
In addition, they have a very high degree of cii;]1cnsional sllbiJity, and possess the
ability to he \\lrunc;" together. :Cin"lenslo.i1al slabjlity is brought about by tbe
careful selcclion or the steel from wbich they ;3.re made" and the st;ibilizing heattreatment process the gauges undergo after h.:rdening. The prop;;TY of \vringing'
is due to the fl,,:"lness and finish of the defining sUi'fhces bciJig such thai "",vhen t\VO
gauges <:m: brough .. into intirrmtc contact, not oDly is the air film be,VIew them
virtually removeci, bui moiecu.lar attr<lctio!1 occurS between the surfaces. The latter
may be shown by the f~cc thill ""hen t\VO 'wnmg' g311gCS areleftfcr D. period of time ,
the difficulty of separc.tion may be conGidcmble. :n fact it is advisable not to leave
30
31
L
Up to
and
inc!llOver ding
Grade I
Grade 0
Grade II
CalibraCion
Grade
Gl"ade
25
CO
10
10
:l:10
13
20
'20
25
.15
,----.
mm
r0;11
20
-. !5
20
GO
60
80
80
lOa
5
5
8
10
10
25
12
:1:50
50
iO
15
10
10
10
15
15
:d5
:::20
25
1"
,J
15
15
50
25
20
'JO
25
JS
+80
50
25
-20
+50
25
35
!-120
25
-25
+60
25
35
-75
+140
100
-30
32
Where the quality of work demands it, the calibr:lll:d \I:IIIH"; (II' ',/II) . "q" .
should always be used. Normally 'protection slips' of 10 I 11 Ill. ;111.1 <l1.lll,< (>j " " , ' " ,. II
carbide, are
with each set. These should be uscd un III\' ,"II.j'/ .. ,1,1",/
combination to prey~nt un~ue wear of the actual gauges.
33
Linear lk[easuremcnt
standards room, locatEd in a quiet area, and not subj(~ct to vibration from outside
control to plus or minus l"C is desirable, and may be
sources.
controL Dust control, [hroUgl1 the filtering of air entering
the roOT;~, is a u:ther
agamst the deterioration of equipment.
All comparators should be permanently mounted on rigid supports, and in a
level p03ition. Givcn these conditions, and a technjque of carefulrreasurement, it
is possible to determlm; toe accuracy of gauges to a fraction of a micro-metre.
The appropriate technique is dealt witb in association with the
and ope:o...
tion of the comparators suitable for this class of measurement.
Secondary standards
(Veriilcd by interferometry)
I
I
be closely followed, in order that machines and instruments shouid possess the
following characteristics.
3.41 Tne
In addition to these
which may be
for
kinematics, a further
important
must be observed. This is the ~'ril1~iplc of Ali~nmel~t, which
may be statf:j : 771e line of'measuranef,t, and trle !me
the cllmens:on
, This principle is so fundamental to goo d CCSlgll
' .
measured, :;lIculd be coincident.
that it is rarely departed from to any serious extent.
A simple e:::lmple, <lIed one of common expetic;nec, is found in the design of
an external micrometer. In Fig. 3.2 it is clear that the
is completely
satisn;;d. Tbis applies equally to an internal micrometer.
(Verified
\Vorkpiece
suitable
method)
F:g.3.2.
Fig. 3.1.
OF UNEAR
hovvever, vie consider a vernier calliper, it is a1.so clear, F~ig~ 3.3, th.L1t the
principle is Qot observed. It may also be shown, th~lt L'.S tIl" displncf;merlt Y
tlle greater will be the probability of a di.':crepancy arising between
dimension A and the feading B.
UlU11..,")]",,)
Fig. 3.3.
34
35
This may be defined as the amount of correction whicll 11111;:. 1>,- 111;1,1" (" III!
in respect of the values of the qnaIlLilic:; !win" 111,-.1:,111,,1
careful calibration of instrument scalc'; will OWl< Ill'" 'I."
difficulty m the time of instrument assembly [lnd testing bul, iJ I w:c, dl'!"l:' '1.11:, ,;
of the operating mechanism may set in, which will require r.:l::i1il'illli"ll ,<I Ii,'
instrument, to restme its accuracy, An anaiogy to th,is iii Ih~ c;i1iilr;l1i,III "I ';1('
to determine their actual size in relation to their nominal ',i,~e,
3.44 Variance
This may be defined as the range of variation in instrument reading which llI:r,/
be obtained from repeated measurements of a given quantity.
Variance is inherent in many types of instruments, the exteni to which il i"
present depending upon such factors as the quality of manufacture or the :,U\Iability of the operating princip.lr:: for the type of measurement to be undert:;l.:.cn.
A simpl0 and common example of variance is the operating characterislies
of a dial test indicator.
il.. test to determine variance in such all instrument takes the form of 3 series
of observations of ir.:.dicatcd values in relation to the known values
slip gauges.
The dial test indicator is rigidly mounted above a suitable reference surface, such
as a lapped plate. A series of slip
is then selected, the series
111
size in increments of say 1 mm
extending to embrace the full range of
plunger movement. If then each
gauge in turn is passed carefully across the
reference sGrface and under the measuring plunger, a series of correspcndir::.g
scale
and recorded. Let us assume that these readings
are taken with the
moving
upwards, The r>;uJings an; then
but in the reverse direction, and again recorded. If the readings are I:'O'H
or
(!)
z
o
<.(
III
c::
1-.
:z
LLl
;:;:
+
MEASURED qUANTITY
READINGS INCREASING
,
It '.vill be ncted that this treatment of i;lstrumcrrc
enlarges on the
"~qU~'l'h
ac"ep'f'd
',1,
, 1 - ut:
. comp.etClj
I " (!etlTI{~n 2nd detennmcd by
r '~.
,,' C~) ',c.
,~
h
. ea t'"
fide i. t l!igy
rererence ,0 tile leas! change t11 the measured quantity which will cause an observable
in, tlot'
i-'S'l"'lm' 11
~1'
,-'
h a ("Aennltlo11
J""
,
i'
...
... -'
<..1.1.J _L.~ . . 'w 1 4Gl "" re~l\..l.tng . .::otic.
Has
no
lor the
iae:t ths.t sensitivity may not be un.ifonn over [he total displacement of the indioatsuch as a pointer.
o
36
co,
READINGS
DECREASING
Fig. 3.5.
37
Linear lvieasure?nent
plotted as in Fig. 3.5, it may be found that the plotted points lie on two cnrves
whtch form a loop, the boundaries of which represent the maxim?lnI variation of
readings for a given value within the series of values of .he slip gauges used. A
pair of snch curves forms a hysteresis loop, and is a characteristic of the instnllnent
under test.
Fig. 3.5 shows that, in general, dial test indicators are not sctitable for the
mea3Uremellt of h~~ight or length differences of a magnitude approac11ing the full
range of indication. They are suitable for the measurement ~)f small displacements,
and in one direction only.
The {(lost important t!lcorcm cf kinematics states: A rigid body in space has )'ix
o.Ffj':::cdoln.
38
any
As an
of the latt;;r case, consider the mounting of II surface plate on a
.
. iou
t 'In Sno)."1 a l'lCln11-'"
t'lP~ t,,110\,ll'I'"
stand. The mOllntll1g
must .oe carrJ(~(
< ,~, ~lllat
,
~".-"
'0
conditions ~~.re satisfied:
j
forces.
(b) The plate must be free to expalld and contract with changes of'lmbicnt
or
3.5
a body is i're(~ to
in
it
mrl"'1 have \JIIC ('1',
a
of
nUlnber of ; he rnotjon5 sho\vn~
tempemture.
(c) The piate must be adequately supported to
its distorLlon.
(d) It must not be possible to accidentally clisbclge the phte from its set
position.
11)3 ke
J.
'i)
\.,-', ....
,,'-1-....
vv
39
vertical Z
;and nne of t!J.e horizontal X or Y axes, is constrained by the
vee, rotation ahm.1t the
horizontal axis being restrained by the fiat
support. It shod.d be noted that no clamping forces an, applied, other than that
of gravity.
r
i
-4t~I
emphasize that only by the application of kinematic principles can the design of
an instrument or rna,chine be such that its accuracy in operation does not rest
entirely on its accuracy of manufacture, although it must be realized that for the
moving member of a vee-flat bail slide to have true linear motion the following
manufacture:
conditions must be realized
(a) The vee must be straight in the horizontal and vertical planes.
(b) The flat must
,.'7
, CONICAL
HOLE
VEE GROOVE
pJate.
IJ
_,_ _I..==:J
Thus, th,~ plate is completely constrained, a~d at the s~me time ~he other
three cor.. ditionc;
for its mounting are satIsfied. It Win further be noted,
t!mt the location
the Dlate is
As an additional refinement, if the three
l'J f~>et (;al'~
'"In ';d;ust i "", !'cre'v:> their relative
may be set so that
C C"!'X'j'"o'
.. l.\.<
,,,ba "..........
the plate is
to a true horizontal plcme.
!", .. ~
'~j
,,-(.be>
;,.I
'~
FLAf,
Underside of surface
.,~
LINE OF
/MOVEMENT
ILOAD/ /
'-L---~-- m
,
plane.
(c) The vee and the fiat must lie in parallel horizontal planes.
Fig. 3.7.
'(
Fig, 3,9.
3.52
If we now consider tile requirements for one degree of fr'eedom. e.g:. lin<::ar motion
along the Y
if we substitute a second vee groove for the conical
we have
a condition <13 in Fi.g. 3.9, the
being free to move oniy aior,g the axis shown,
provided its ccrrtre ;f
is contained within the triangle fOf:lld by lines conball feet.
40
/\A
Fig. 3.10.
41
Linear
out un experiment in the calibrati,oll
magnification, sensitive comparator.
considerab1y from
or slip
The accuracy of determination
such 11 method should 'JC of the order of
:to05 p.m, uut obviously it will vary with, and be dependent upon, the accuracy
of the comparator, and the observer.
DESIG1\T OF
ceD1parators are the
instruments used in linear measureSllCh th~ir.
has received much attention. Many principl:.;s
hnve. been ~sed to ob[~m slll!.a!Jic.
of magnification of the indicating ~vice
relatIve to the Ghange 1ll the dimenSIOn
measured. The main principles used
are as fGllows:
(a)
JjJl1vl1,:>,"HW."
ment. and as
(b) lvfechanical-opticaL
(c) Pneumatic.
(d) Electrical.
(c) Fluid displacement.
3.71
Additionally, comparators of
sensitivl'tyall,a"
r1
an",
suitable for li.0C in ~tandards rooms, rather than inspection
have
been
and brought into wide use for the calibration c,' gauges. These 8re:
'-LIlHlJdl
The design of 2:11c11 originated at the Natiol1<:'J Physical Laboratory, and was the
work of men to whO:11 much is owed in the field of line mcasurel11eIlt.'~
Each of r;L;se instrurncnts is now
commerciai1y, but the simplicity
and soundness of th<:ir d::sign is such that no boo~c dealing \vith
Can
overlDok them.
Nominal
Size (nun)
2
3
Known error
Reading all
Grade If Gauge
+01
,LOA
()
+08
+05
-i-0-6
of Grade 0
+02
-01
+01
-0,5
Ei'/'Or ill
Grade iT Gauge
03
-:,0,8
+03
+07
06
\'
J'v!ea:,'urClIh'W'
POINTER
(b) READING
GAUGES
~l}~~~:~ING f3~O~~~Jl_1
Fig. 3.11.
CHf2 ,,,,::n.
A:'e:lture of this
and one which is quit;; L!nique, i'l that the gauges
togethc;-, and l'.ot in isolation from cud} vthc-:-. ]'11';
of
t'bis may be judged by the fact (ilUt having 'wrung' t:K: Jaugcs to the
~Uly !h~;!~~ he left to
jn
before readings arc taken~
I min/;:nm of g:mge
recommended for thi" purpose.
The lJl~;lruTD.ent is
:,on 5iz,:;s to aGCOmmodllte gauges up
to 1 .m in length. This agaIn is a unique feature.
l!'rc
44
II---'-----lr +-=1
/
FIXED
ANVIL
(~It~c~
EAR MOTION
Of' I,jOVING SLOCK
MOVING A~iVlL
(SiHG~[
~-
~J
LINEAR DISPLACEMENT OF
:. MECHANICAL MAGN
WE8~
\)
\
~f
l==.-:"~I-~-'""~\
'\ --~
.'
ANGULAR DISPLACEMENT
d"SEPARAT!O?;
OF r'OINTER
\
\{
STRI;:>S
'\
\___
---I
BALL)
WEB
I ...........
I~
I~ -
,0
EFFE::1 OF DiSPUl,CING
--MO\,I~;G BLOCI{
----Fig.3.12(a),
45
Linear
The scr"L, is caUh;:ated to read dircsdy to 02 p,1ll and the scale divisions
are about 5 rom
It is therefore possible to read by estlmstLon to 002 ~lm.
It will oc noted that thi;:;; design produces a very
magnification in two
relatively simple stages, the first of which depends upon Idnt:mrttics for its sensitivity and simplicity of manufacture.
The contact faces of the anvil and plunger are also unusual. That carried on
blDd: ]\(l consists of a single ball. The anvil face,
consists of a cluster of
three halls
a
The gauge.: is therefore seated ':Ofreetly on the
and cannet take up r!_~ incorrect attitude rel~ltlve to the line 0f trlo'V~~raerrt of
j}1 easurement
block M.
SCALE
STRIP
ClIHTILEVER___
STRIP
BELL CR/\NK
LEVER
PROJECT'ON LENS
WEB
SCALE
COf'WENSERS
PRISM
th~:: a\~cnracy
For e:mmpk,
the pointer t:)
oe
the lilstn.l!nen
46
to be eXi;zctcd ..
Movement of Johanssonivlikrobtor.
of the reG-dings.
~teel
strins c~'xrvjng
the m~Zi11ncQ'tion
dcilciencies
.fronl the n,arluf~lctur.'3 of ::;ither of r:hcse systenlS may
for by the cal!braticH of the :;cnie of' the ins~',:.~un.l(~{~tt by rel{~rel.1Ge
to
_. ct-naiI::d on a 3e/ics of stip gauges
ki1(>'1/I1 l;{Wrs.
!,./;le '-.:'iul GCnS1(1,er nOVl the
of
of !.he niGi'(~ 'iI5U~:1J types of
comp;m'.wrs,
in which kirrerrw'(ic
are
and
b~
Fig. 3.13.
example of design, in
the
plunger.
As the measuril'g
(1
47
I"".
5
Linear !VIcasuremcf!i
,~
d8
=- =
di
.. ~.
amolrncatlOl1 =
911
---
W2n
n = number of turns
8 =twisi.. of mid.point of strip with respect 10 the end (d.;;grces)
The dimensions of the rectangular section of the tl:llistc:d strip are always
very small. and vary according to tb~ amplification of the lnstrumcm. An average
value for these dimensions .is 006 m.m x 00025 mm and the stresses in the strip,
for a given tension, may be further reduced by smail perforations along its length.
The purpose of the cantilever strip, other thiln as an anchorage, is to allow
an adjustment to be made in arnDWication. Its effective .tensth may be varied,
and if for example it is increased, then for a given lotal rnovelnent of the plunger
mOTe of this movement will be accommodated by deflection of the cantilever, and
les~; by extension of the twisted strip.
It should be noted that here again is an example in which design allows
simplicity of manU~lcturc; the final setting of the instrumem ampli:lcatioll bt~ing
made by a simple adjustment of the free length of the cantilever strip. It should
also be noted ~hat the cantilever mounting is adjustable, and by slackening one,
and tightening the Olher mounting screw, the initial tension in the twisted strip
may be adjusted.
The lower mounting of the plunger is in the form of a slit C washer as shown
in Fig. 3.13 and this completes the movement of an instmmenl which has no
mechanical pivots or slidcs at which Vicar can take place.
The instrument is surprisingly fobust and is produced comm~,;rGial1y in a range
of j~lagnif,C:J.tions :.1p to :< ::;000, B.lthough under controliedl".coratory conditions
much higher sensitivitie3 have been 2.chjcvr;d.
3.722
;jIter~"Un~:;1.y
at right angles
to ead, other. ;t ;zm be shov"l1 Lh,1! if 3.,1 '~xI~crnal f~m~ is applieJ to the moving
member it will pivot, as would a hii:ge, clb011t tb::; line of intcrsect;ou of the strips.
48
SCALE
I
fr
~?-"
./
/' /
(e;'" --'"
.,\\'(.~~
A
o~.
f
~~~/
FIXED
/%"""\
\
DRIVING
=_=~~fSLIT
DIAPHRAGM
.---~
j'J---- "i--
~
h'
,lR~1 OF
?:r;/'FJ
-SAPPHII:U: BE!,RING
BLOCK
J. .
I\~
LEI'lGTH l.
CROSS-STRIP
PHOSF'HOR-BfWI.jZE
HINGE
DRIVING BAND
Df,UM
l~---, C\
x
'-)
SCR~NS
I-~ 1---1/
MEMBER--......n --h----rJ,-- Iii
l"'--'-'u
POINTER SPINDLE
\ /~~......---./'
!J .//~,'~'-..
-PLUNGER
MOVING
BLOCK
=r
NOTE: FOR CLARITY THE FRAME
SLIT
~IAPHRAGf~
I
Fig. 3.14.
Another meclianical comparator of illgeniom, yet simple design is that produced in a range of magnific:niolls by the Sigma Instrument Company. The
movement is sho\vn in diagrajl1n),l':i,~ form in Fig. J.l4 (it mlls[ L',~ emphasized
that this is Q diagraIIl explainIng the pr.i.n~lplcs and that thl3 ac[ual 1110VCrncnt is
m.uclt m0[{~ cornp;tct than this).
The plunger, mOll.nted. on a pair of sFi. dhDhragms to give a fricti-Jn1css linear
movement., has mGlElted upon it a knj~'e ,edge which bears I.lpml. th;; face of the
T.OOV.iDg 111Clnber or a cross . . strip hiuge, rfb},:,; C{)1l0.lStS (If the rr~oving c:)G.l.ponent
and a
Attached to the moving member is an ann whieh divides into a 'Y' form. If
the effGctivc length of this arm is L and the distance from the hinge pivot to the
knife edge is x then the [irst stage of the magnification is L/x.
To the extremities of the 'Y' arm is attached a phosphor-bronze strip which
is passed around a drum of radius r attached to the pointer spindle. If the pointGr
is of length R then the second stage of the magnification is R/r and the totai
magnification is !-/x x R/r.
I
I
I
!!
I,
it
II
L_
Di~graIl1
of movcnlcnt of Sign1:l
rr~cchanical
compara.tor.
Ca) Safety. It shouid be noted t]-,~lt the knife edge moves away from the moving
member of the hinge :lnd 1S followed by it. Therefore. if too robust a plunger
movemcl,t is lllaac thc knit"; edge moves away from the hinge mcmber and shock
loads arc: not transmitted tl.irol-,!;h the movement.
(b) 'Qead b~'ai" ,.ead;I7~;s. TJle poinkr is cau~cd to come to rest, with little or
no oscillation, by mounting Of) tIle pointer spindle a Den-ferrous disc moving in
the field of a permanent magnel. Rotational movement of the pointer and cli~c
cause eddy currents to be ,;el up in the disc, which have a turning effect ill opposi.
tiOl} to the' direcf:Oll of motion and proportLonal to the rotational velocity. Thus
49
Linear jf;l'easuremem'
oscillations are damped out by a damping force "'ihich increases as required, i.e:
the amplitude and velocity of the osciHation.
7v-:rtl
AND
(c) Fine adjustment. The normal height adjustment of the viGrk strrge, to zero
the pointer, is too coarse for the
magnification of silch an instrument, and;
..is uHslLttablc for this purpose. To overcome this tt,;: dial of the
instrument
is mounted so that it rotates about the spindle axis. Thus, final setting of the
. pointer is achieved
the
rather than the
and in this case the
Instrument
rather than hinders.
UGHT SOURCE
CONDENSERS
INDEX
(~lTOnconsly
dial type
The line between the
eye and pointer is
normal to the scale when the pointer obscures its DWJ\
in the silvered
of the dial.
pnf::llIax by
the tip
tip is turned through
across lis end a small 'Tee'
moves in the slot and thus
of the sca!es. As the pointer and scales lie in the same plane the
is cOlnpit.::tely elin1inatecL
.1
'
are
of an index is
fhe
causes
th~n
PRO.!ECTION
LENS
SC~tle.
MIRROR
:t
v
PLU~~GER
Fig. 3.15.
In such a system:
Mechanical ampJificatioll
1 x 20 x 1=20
il!1itS
=50 2
=lC{) units
=20 x 100 units
=2000 :units
The factor of 2. contn [ned in the
following manner.
(~onsi.dcr a
bean1
relative: to the normal to the
form the
11
an1pEfication is
falIiI~g
on
Fig. 3.:6
rcfkctor, aed at
The reflected bearn \vill
arnp1iti~.d
of an
tilt about an
to a posicion rc1aIi'ic to a
and
t~
3.15
lil
dit:i.g:ranl
forB}
only, SUC11
;)Xl
arfangl)lncnt
Fig. 3. \6(a). ReElection from a phlnc scl!"i:IC";.
50
about in tll;,7
I)
al~;o
JVietrology for
Linear lvleasurwrcn(
If DOW, the
incident beam
plane reflector
through 2ou,
r,~'11'H'-",,'
COIFtant
j t 1'S ;;e~n
[PI'
. v
,..!.:i
V
.1. ry~
.L IJ
D~
REFLECTOR
TURNED THROUGH Ci8
It is ill effect an auto-collimator, in which the reflector is built in, and deflected
by the measuring plunger. To keep the instrument compaCt in design the
path is turned throOlgh 90 by the prism, and of course the scale is graduated \11
linear units, usually of 0-001 mm.
Such instruments have very low inertia in' the movements, as the only
parts are tile::: pli.mgcr and the reflector. The norma'! point<::r in mechanical instm
ments is replaced by a bCD.m of light which is of course weightless, and the
lever system (Fig. 3.16) gives a doubling effect to the magnification.
Fig. 3.17.
FilS. 3.16(b).
tUT1Jed
Reil~ction
If the distance from the plunger centre-line to the mirror pivot is x, and the
plunger mOi'es a dIstance fl, then the
movement of the mirror
The rElg[e
the incident beam and its new normal
Likewise, the clDgl(~ betv'Iecn normal 2 and rei1ected Deanl 2
The total angle between the ineidefIt .1nd Eflccl:eci bC3.tns
2(8 + MJ).
le i::: this 'built in'
of 2 obtained from a
which
is of such value in
COll'lparaton. Clearly
if a doubk reflection is
provided, as may be
from two opposed ret1ectt)fs, one
the other
capable of tilting
then t.'lc factor becomes 4.
~orm[l:Hy, _
. :mHkient amplification is obtained fmm a
mirror
~sed III COl1J,mctlOl1 WIth a suitabb length of reflected beJ.m,
as a weiohtless
wver.
h
x
the focal length o[ the lens then the movement of the scale is 2fcf)
4).
ae
I;;>
. ..A point of
norrnal bac~~
. , rei1ectioJ'l
.,
.
~
{.onSlUf';Tatle
care IS
.... I)-
oJ
'j
C11)plicatioa of the
Measuring Tools Ltd.
Scale movement
mo\,cmem:
x
It should be r;o(ed tb<1t this
is double that of a silPple mechanical lever in which th~ mah"1ifkaticn is the ratio of the
or;. either Sit1C of the
fulcrum. In t11i:; caGe one lever arm length is the distc1.ncC x; the orllcr lever arm
is the focal
and the factor 2 is due to
lever.
Further mngnif,cmion is provided by the
in the
Overall
rm~gnifica tioll
magnification
53
F
Linear
Industrially, pneumatic comparators, in which small variations are maGe in. the
dimension being measured vi'itb
to a referenc>.O dim(~nsioil and which are
shmvn by a variation in either
or
the velocity of air
are
becoming of
importance,
The reasons for tilis "xc that very high L"U'.'.H.'''~''
contact is nmde either with the setting gauge or the part
and. that
internal dimemicns m;).y be readily measui'cd. not
boundaries, but also
form.
the system lends itself to the
inspection
a
or a number of rlimensions
either
or
immediately ancr the operating
of ii metclline 1001.
or
3.741
Back-pr'essure CUl1iparatol's
The air prtcssUll; variaLion sYGtem is based on the ilse of a two-orioce arrange111<:nt, as shown ill
3.19.
ME;:\SUii!NG
CCN:flOL
o2
ORIFICE 01
COLLIMATING
LENS OF
LE~lGTH
f
I.VOR:\.
SURFAce
Fig. 3.1.9.
MCVEMENT
OF PLUNGER
Fig.
3.18~
the
E~scntia13
of a
bacl.;:~pressure
OVt:ral1 magnificllLion
(Co['rle~y
0/ C'Plical jl;fcasuring
AB
" x-y:
Tools Lid.)
within
for the
diameters
or orifice;;.
Linear JHcasuremc!ll
bA 2
Al ,
The
j.l9.
CONTROL ORIF!CE
LINEAR 'NlTHIN
1% IN THIS flANGE
I'
."
Fig. 3.20. Charllcteristic curve of backpressure pneuma/i,; system.
TUBE
MEASURiNG
JETS
As with any other comparator, the sensitivity is the mtio of the change of
position of the indicator with respect to a corresponding chang;: of position of the
plunger. In this case the change of position of t:le indicator is
and of the
h
' te
d,,'!~,
.senslttVlty
. .. b'
cmg dPIJ
pneumat~c
dP o
bA 2 Ps
..11
bPs
56
DIP
Fig.3.21.
Application or \)::lck-pressurc air gauging system used by So/ex Air Gauges Ltd.
Ca) Magnification Otij1lslment. It has beel! shown that the mngnifkation can be
varied by varying the di,nncler of tbe control ('riflee. This is achi(;vcd by means of
a taper ..l1ecdie valve in the control criDec and t:ilabics a
sea.,c; to be used
for all work
adjusting the magnific,Hion and zero
(b) Zero ad;ustmellt. An air bleed, upstream of the I11Jusuring oriDce and
a taper.needle valve, provides a zero adjustment.
controlJ..::a
dial
gauges.
Il
Engineers
Lineai'l}[easurement
C=:=J===:==::?---:::-
TO GAUGING HE,AD
I,e.
MR PLUG G!:;UGE
GAP GAUGE
ily
VM~ES
'1'1: ..."
!;( : \ \'
",(.\illd
! lI. ,:,[:;
<;lj;;c;li 'i':
r;~tt:
fon!) <,C
"
tag~
ST/\r.~iU
TY
() f l: .'.;:
322 shows the circuit foc such a system. Ajr fwm a main suppJy, say the
is 5ltefcd ,<mel
It pass:.;s
an
in th[~ for:n of J.
Ihe bore of \vhich is
and in which tlie
.:11e indicator tikc;$
in
tube such that the air
the annulus
the 'Hoat' and the tube is constant. The mrtllen escapes through the
crince.
If, nO\"-1 a nlHstcf g~it.~geD.f corr;~ct size und
fOrD] is used to pre vide
a 4datum' ~~Llt\~ of air ffo'l'Y
cystem, and J <LS _:~lGVv.n by the height of the
iIldi~;.r(or in ~lle tube, any vurla-:50n in the dInlensl0n of the part
\;:ill
1
"
....
,
e
the effective
prouucc
::1. va,r~:u:n
in
dl:,r2.-~t,!
?~ JlO:,.j' .
~ t1".I",
. ~pal't
C,AUSZ::
TO HOrArE FOR
b[~
system
RC::-7 Ut:\TQR
Fig. 3.22.
For
tb,,, norma;
unit: for ln~:ornJ.l di:xnetcrs consists of a
free !jttiDi~ pLug g,:.u,!;e as in. Fi;~.
having two
By rotmi;g the hC:ld ] 80" in the ~ore being
t1.'.c
be detected.
By
a
head
3.23)
orilkcs,
the <;onditi:Jl1 of a lbrec~lob~;d sh[jft detected..
the shaft through
120. The same head will also gi lie an average diar:::cter ,)1' the shaft.
that
J.n\~asured.
]}lis
58
59
Linear Ai {!aSllrnJJl'lIt
Fig. 3.24 shows a similar principle out for an a.c. circuit, in which the !llO'le
rl1ent of the plunger dispiaces an armature, thus causing a variation in the indue
tance of a pair of coils forming one ann of n.n a.c. bridge. A possible armnge;,,:nt
for the
head of such an instnrmcnt is also shown, in which till: piul1p.:r
displaces an arm mounted between the coils on a thin flexible steel strip. The ann
carries an
and the inductance in the coils is dependent upon r.ho displacement (JUne amlature reiative to tIre coils. The baJance of the circult j:,;
, set to' zero, and the amount of unbalance caused by the movement ofrhe
plunger, and hellce by that of the armature, is amplified and shovm on a scak
graduated in !.inear units. Magnifications as high as x 30 000 arl~ possiblt" WiLli
this system.
FI~1GER
FINGER
SET AT
rlE:GHT OF FLUiD
FLUID CHfl.M3ER
PLUNGER
Fig.
3~25.
.
1 D2
1\',agl1l11Cs.l!on
IS approxIll1ate Y"(i
z'
Eg. 3.24.
60
Measuring bea.d
tOf
dectrical ~omparutc;.
1\ fine bore. capilhrry tube is ar:rangcd so that ~t~ low,: ?~d ,is plll~ed in a
chaluber or reluIl'Vel~.r
cross-)ectlonaJ area. contaUll!1g a DUld or lov,.' 1./1SCOSIty_
The bottom of th.~ eil:lfnbGl' rakes the [(lIm of a diaphragm which may be defi.ec.ed
by pressure transferred to it by the n'!casuring plunger. j)japhragnl def1tction
'cnuse~ a small quantity of fJuid to be displaced from the chamber into thl> tube,
61
Linear lV{easurelnenl
The
iimit::d
\:0
.i11UCfl l~ss
?';i'ucn
u .......<J'"U!Lf',
is not constant.
the measuring force
vior!(; ho\,/::;\'cr.
rlst: to
:.1 nU'~llber
l:-;]scd
Gn
th'::
U,iC
bet',w;'cn a
othtr
~Jrce
1nvolv~::1 djrn~usic'n.s
of
of a
:-naclttnes of various
screw as a means
G1icrom<.':l;~r
gauge~
with screw
a;) tll 1 m.
Fig. 3.26.
of the
of the dial indicator. The
or
now substi1.utcd r'oj" the
bar cmd a
same
ofthc indic~lt()L The difference between the
is clearly lhc difference between ill,; kngths of 1llC
~en!ith
be measured is
is taken at d:H::
t\VO l11i(:l"Omctc~:'
or
62
63
Linear l''/leaSlIrcmc!!t
3.32 Universal Measmring Machines
Several machines of this type are available to cover a wide range of types of
rneasurement. One such maehine,
3.27, is made by Societe Genevoise. The
divided scale viewed through
basis of the accuracy of m"nsuremcnt an
a rricrosGope, and used to determine the movements
the measuring ilnvils of
the machine. It should be noted that the divided scale is positioned to be in line
vrith the line of measurement as determined by the axis of the measuring anvils.
Thus, the principle .of alignment (section 3.41) is satisfied.
or
~.,
The reading accuracy of measuring machines based OIl the use of built-in li~e
standards has been greatly enhanced by the development of the photo-electrIc
microscopc< The scale reading in such a machine is taken by causlng a
of
hairlines in a m~croscope
to straddle evenly a UIiil line on the
com bining the scale andgraticule readings. It must be
that even straddling by the hairlines is a matter of subjective judgment and is thus liable to variation.
The photo-electric microscope overcomes this by scanning
with a
photo-transistor or cell, the light contained in the spaces bctwe:::n the stradd~ng
wires and the line. The electrical outputs from these alternate scans are causea to
be of o~posite sign and their algebraic sum is recorded on a meter. To make a
reading therefore, the microscope is sct by eye in the normal way and then adjusted
until the meter reads zero, when the scale line is accurately positioned midV{;:'eY
between the two hair lines. Thus greater rcp::atability of readings to a higher order
of accuracy is oi.lm.ined.
Fig. 3.27.
Th,;;,; IDEtchine is uluvcrsal in the sense that lengths, and diamet~rsj of both
and G.lxcaded WOile, tapers, and the
vi scre'.'v threads m2.y be measured,
nne! to a higb order of ~ccuracy.
Simiiar machines an; bL1~lt by SocietrS
cmployiI).g; t'vo and three
co-ordinaU:~ measuring
a.iso based Oil til':; -rise of line Dt::tf1dards. It is of
inter'C3t d:a.t the ~ca,res 3-re enc\)sed \vithin the i13Qchines and atc tlnt~ ilot !i;:tble to
,-,,,"'e,;,;,, by physical contact and atmospherk
~i:; z:re the
faces of
.nd ~i.luges.
64
65
CHAPTER 4
~yst<!m.
-J,.
-,
4.2 Ar'TGLKS
SIr~E El~~l:{
E~~~:'~:~;:::~~~~:~,:~::~i:~:~~;i;'
Fig~ 4.i
1 ilC Sine
bar utfcI'cls
set at a
66
67
If 1 is the linear distance between the axes of the follers and h is the heigh t
of the gauge blocks, then sin fl=hjl.
The design requirements of a S!M bar are as foHows, and urJess these are
carefully maintained the Girder of accuracy of angular measurement will fall:
trial and error. This time can be simply reduced by following a logieal procedure
as follows:
(a) Sci up the sine
and workpiece as in
4.2 so that the upper surface
of th(~ work is approximately parallel with the table surface.
(a) The rollers must ibe of equal diameter and true geometric cylinders.
(b) Take readings with a dial gauge at both ends and note their difference,
noting which end of the wad:: is low.
(b) The distance between the roller axes must be precise and known, and
these axes must be mutually parallel.
(c) Assuming that the end nearest the high end of the sine bar is low, then
the gauge block height must be increased by an amount equal to the
difference in the dial glluge readings multiplied by the proportion of sine
bar lengtb to '~vork h~ngtho
For example, assuming that the end of a "'"orkpiece was 001 lTIm low,
the sine bar being 250 mm long and the work 100 mm long, then the
required increase in gauge-block height will be:
(c) The npper surfac:c of the beam must be nat and parallel with the roller
axes, and equidis~llt from each.
These requirements are met and mailltaincd by care in the fnanufacture of all
parts, which should be hardened and stabilized before grinding and lapping.
Provided that all i3 in
and with n 250 mm siIle bar the distance between the
raHer axes should be aCC'I:1;t:"Lte to witl;in 1 ,urn, then it m.ay b,~ nsed in COqjUDCtion with gauge blocks t(} realize high orders of accuracy for setting individual
angles.
O.01 x 250
100 = 0 .O~"') mm
TIlis will not give an immediately carrect s(;tting from a first approximation,
but it is mm;h quicker than by a trial and eeror method.
Finally no sine bar should be used to 5et off ftngles greater than 45, as beyrmd
this
the et[crs (he 10 the centre dinance of.railers, and gauge bloCKS,
in error, arc much mr.gllified. It is interestbg to plot a graph of angular error
nominal
if an error of 002 rD.m in an values of his as:;umecl.
angle:; should be SCL off, '.vhere possible, by subtraction from 90", I.e. set to the
complen;cnt rather tbm the angle, and to a datum provided
an
piutc
or cube known to be square to the table surface.
Thcs;; are a d'';'le!opmcnt of the sine bar principle and aF; set in a 5imiJa,r manner.
It is ckar tli;lt the sine bar is sui1able ollly for relatively small '\I(ork of
The sine table has a lnrger
surface and is much mom roi)Llst th:lJ1.
the sine bar ~U1d is suitable for
heavier worle.
A further deveJopment is the compound sine table in which two sine
their axes of tilt set at
(0 c<{cll other, are mounted on a common base.
The compound angJe to be set is resolved into its indivicluni angles in two pl::i.nes
at right
to each other, ;'Inc! each table is s;;;t accordingly. ~
In practice the sine h~rshollid be used ou a Grade A surface plate, and even
so it is desirable to support both f;;llers on gaug.e blocks s;) that the [l1jnute
irregularities of the plate may L'e eliminated. Thus in Fig. 4.1 the height h would
be the difference lrr
between the two prIes of gauge blocks.
To measure an indiv~d!;al angle Gare lImst be taken not to r~)nn a compound
angle by having the workpiece'mi~,aligned '~,ith tbc; sine bar. Thls is avoided by
lightly holc1ing the bar
an
plate or cube, The
is similarly
held
the
(lD.c1 a s.-;ries of
IDl\:;;n
upper sllrL1.ce
with a dial gauge. '\~Vhcn the readings nrc constant then the ang~c of rhe \yorkptece
cun be ohtailled from :;in ,9
In determining a workpiece
rathe;: than s;:tting it !)ine bar to a prl~~
determIned angle, much time is ofwi1;:vasted 111 finding the correct value of h by
68
For the testing of conical work:, centred at each end, the sine centres shown in
Fig. 4.3 are extrcE1ely llseful since
alignment 2.ccuntcy of tho centres ensures
th,lt the correct line of measurement is made along: tiK; workpiece.
69
f;
The principl,:: of setting is the same ~lS in the sine bar, although a hazard to be
avoided is of the work and centres not being co-axial. To ovcrcorm; this the werle
s;lolllci be rotated on the centr~s until the maximum dial gauge r:;;ading is at the
""P Tht' C1Il01t>
rraUi!c
blocks in this condltiGll, Hud then the
f:; '"'.,is c"lcl'l2ted from the c
-..'
worle turned thrcwzh 180 and the process repeated.
~.~,
. .I..i
1:.."
'_.
/......
6-----1
---- .J--J_--1
.--8-1:37\
.-~
~\---
Plvar i'-Ji - .
L
t.----r--~
I_'J\
I ( ) c ..
. -__._J
_--
G.t>.UGl=
-R:"'-'
f3LOCKS
_ _ _- -
I__ ~
.__ C=~
.. ____
The mean of the two angles determined will be tho" semi-angle of the wod::
n;,c."c,,,- <~ ~,tJcll'1'IC;11 ~-.!-,. Yl\U~t b Dc,j n terl 0 1 rl- tn8.t an~{ ',vork yv.bicL:. run~ 0 ut to a rneasurj-'.1""
able extent ~~ouId probably be considered as sllb-stac.dara ill quality aud be
rejected on this account
.
J.
'--
i:
Fig. 4.4.
Fig. 4.3.
f..~_
I.
; i
, i
/r---f ::1p--:J -
~I___._____~. _____.__.
\.t.
t;
_~-~r-
____------l;:-7.;."j
------t-.---J
ROLL~t~ (
-'
.......
-"
",
1.
-'.
_.
In triangle ABC it
r[;
...
... (1)
Thus the semi.-a.!lf:\le ()f tRper can be
deter:rnincd by direct ll1CaSl1l"en-::cnt.
1'0 find the Ill;}xiDll!lil and n1inirDtl.lYl
diameters at the top and l;uttom of the
step it is first nccess;::ry to dct.::rmil1c Dc,
the dia.Illctcr at the sInaJ.l end. Fig. 4.5
represents the small end of the gauge
during the measurement of jl1'l'
Tauer Ernlt g:l.ug~~s are not (1(Tc~nall'\! used to rneasurc the taper, or angle, in a hole
r<~ ....~fl'l)l:-j0., hnt
vI..
. - . . -\a['~"l
- . .:>r 'L'O BTla
.. - \V~1('tll;r H"le ui2."lrteter cf ;rl (Jartjcu1c~r cros:;,~se:,:tion of
-,.H_~
Into
~;1/Jrk
~~
-~...,
is
~vithin
~./..........
-1.
the
Tll(:;
m;~:l~llrcment
t\VO
sIi:ges:
allov~'ed
It is seen 1h8.1
A11 =L~s+2r+2x
and from Fig. 4.5
tan
90 - 0 r
---=-
.'. x
. j'.;11 = Dc
01
(90 - 0)
r x cotangent -'--2:----'....
(
Jf1 = Ds -+ d \]
+ cot
70
90 - 0\ '
.
-2-;
wncrc d=roller d13..
90 - o
--2-)
71
f)
and D min
where H is the
of the gauge
and S is the height of the step
Ds + 2H tan f)
Ds +2(H ~ S) tan
"
."< ;1
iiI
~1
(J
;1
~i
Fi:;.4,7.
e=
-M~
~.2-~,_':::'
90
shown in
4.6. It consi::>ts
, )f a
reading to
O'C002 r,m l;lnits, J.BC a fiducial incli
cater, which is fr'ef;) t(~ '11oat' across
:lS
the line of m~aSill-ell1ent . The !neasnriq:; head Cat, be raised or lowered on a "
stiff ..;oillnm, and tile work is Sl1P~
ported on a
base having a
0-1
J
(where d =
b~tll
dia.)
!}
J) i Fisioll
Flg. 4.6,
Co, Lid.)
necessary size
fHade.
This method at least provideS an objective process in approaching the correct
'feel' and is not simply :l method of ",uessing al a 'nice snUf" fit' \vhich l1obodv
defines.
~
~
.'
Gauges
Tl1e
here is similar that llsed for taper plug gaUI;(oS, the measnrernents
being made with
,~alibraled balls Zend gauge blocks.
72
The procedure describd in section 4,32 may only be useo whc~re the taper bore
is of large enough di:llneter to allow easy access of balls and gauge blocks. For
F
73
smelller diameter tapers, angllbr and diametral measurements may bc; made using
shown in Fig. 4.8, in which the height at which baJls of
the
diameters seat directly on the taper is measured..
h~, to giv~ rise to .rdatcd errors in t~le calcul~ted angle. On no acco~~t shoul~ ~he
balls be dropped mto the taper. It IS better l[ they are gently reIlec, mto positron
with the axis of the taper lying almost horizoEtal.Again, here is a case where
practical experience of snch measurement allows one to overcome the difficulties
involved ...
Reference has
been made in Chapter 3 to the Brookes Level Comparator,
in which the bads of comparison of the height of two gauge blocks is the sensitivity
of a spirit leveL
Essentially the
Fig, 4.8.
IO
level is an
3 radians
LM
C1
sin :::
L.
If the graduations afe at 25 mill intervals .and these represent a till of 10 sec
of arc, eben:
-hi r1 +1'2
1',IL
LM
mid-\xay
of the
tion t'or the new m,'8:;Ll1"(,d v<:!hes obtained.
25mm
25mm
= 51 500 mill
orR=515m
the caIcula
111m
or ,,>0012 111m
andOP=OL+PL
Similarly as the base length is reduced from 250 mm, so the sensitivity i;;
increased. 'l'hus, if the standard base of 250 DIm is reduced to 115 n1m: tl1en ca!..:h
scale gradui:ltion represents 0,006 lnl1l.
. ,
'
Tl1c main use Df Q nrecisiol1 levcllies, not in measunng
"r,( ill U1l'a:w!"
and IVi;ld in machine tool
This is dcail Vii:]; aWl,.' rid!>'
in
6, but bridlY, if the level is stepped along the ollde ;)\ i:'i,:,"':iI:: "I- il:
Own base length, ti1C first position
taken as a. daturn, h(;it
liC ~llj 01;,('(
points relative to this datum can be determined.
74
;}(
75
4.41
Dil'isiOIl
This effect is shown in Fig. 4.9 (b). If the ray passing tlJIOugh the gcomel,ic
centre of the lens is considered, as it is, unaffected by refraction, it can be seen that:
x=2ofmm
where f is the focal length of tlle lens.
The Clinometer
The clinometer is a special case of the appJication of the spirit level. [n this instance
the level is mounted in a rmatable body carried in a housing, one f'Jce of which
forms the base of the instrument. A main use of the instrument is the measurement
of the inclUded angle of two adjacent faces of a ;vorkpiece. Thus, in use, the
instrument base is placed on one face and the rotatable body is adjusted until a
zero
of the bubble is obtained. The angle of rotatIon necessary to bring
tllis about is then shown on an angular scale moving
all index.
'
A second reading is takeu in a similar manner on the second face ofthc workthe included angle between the faces being the difference between the first
and second
Depe11ding upon the type of instrument used, re3.dings direct
to 1 min are
and up to a range of movement of 90"
COLLIMAnNG
LENS
Fig, 4S(a).
RCFLECTOR
w
z
0:'
f,cv~ls
4.5
ANGULAR
focal
76
a Jistance x from O.
REFLECTOR
There are certain imDortant points about this expression which are Dot
immediateJy apparent. Thc~e arc:
(a) The distance between the reflector and the lens has
separation neJ:\veen source and image.
it
110
effect on the
77
(c) ,AJthourrh the distance of the re/'lector does not affect the reading x, if,
cit a giv~n value of 3, it is moved too far back, all of the Jf;;ilected rays will
I err
f
..
miss v the lens completely, and no
Wi'I'I ,-"
IJO rormCG.
.ms, .or
a WI:!;;
range of
the minh"TIUm dis:ance bet;,vcen lens and reflector is
essc~tial. This is particularly impor~:Jnt where llw principle is used in
optical comparators (Cha)ter
However, in -1tis application it simply
limits the maximum remoteness oftlle refiector if the full range of
of the instrument is to be used.
of one 1 sec of arc. With care, and given a rigid mounting for the instrument,
repeat readings of 0,2 sec arc possible.
The instrument normally has a range of readings of 10 min of arc up to a
range of 10 m. It is invaluable in machine tool alignment
or for ;JIlY
scale measurement involving smull angular dcvin tiolls. A fuller description of iIS
use for such purposes is given in Chapter 6.
4.52 The
TARGET WJRE
REFLECTED i;'1'\GE
OF TAaGET W:RE
~IICRO!vtETEF/
ADJUSTMENT
FCR SETTING
LIGHT ::;OURCE
SEcm,os
SE',~CND3
CONDENSERS
Dekkor
nl~l$ter,
a group of cOftibin:ltion
(see section
is sel: up on the
p1ate and the instl'Un}e~lt
readillg en bot}l :jcalc.s is obtain,~d. It is '10\V
ti1C
to
EYEPIECE
OF T,i~RGET \ViR:::
78
Df auto-co1:iinator. A. reading
scltin[!. wires 5tnudlc the image.
on tbe face to be
gauges
until a
~i1"',~"',,,,,1.
be held in
v;ifh elust:c bani.l~. 'Tl1c -<vork ~3 t1C;VI
rotated unti] the illuminated scale moves aCrdSS tl,e iixcd
nnd is
until the n;,,:d scale
is tllC SalltC as on the
gal~ge. Tll:; error in the
work
is I:.~l'; difference in the two
on the iaLlll~iIl:lL::;d :;Celk,
To c!v::cJ.:. tile angle bCIJv 0en f~tces \vhich are
no TIlast,:::: ncccs~ary . If gauge blocks are held
dekkoJ'
to give a readi!l!~ it '\vii! be found to con~;;si. of two mirror
(due to double ren;ction) of til; illuminated
on each other.
The
of the readings of these images will
the error in th:::
right angle.
The
this manner is sl10wn in
it is difficult to
the method provides a quick and accurate means of
:lllgle3.
the
dek.kor is not as sensitive ,,$ the Ci.l.ito-collimatcl' II: ;,':
extremely usefal for a wide range of angular measurements at short dlsi:mccs.
79
lyl~troll)gy for
E;rgineers
Conlbination angle
angles, so that they can be
realized that angular blocks can be
REFLECTECD !'A.aGE
OF ILlliMINAn::D SCALE
XED SCALE
FIXED OMUf.1
SCALE
OF
OF ILLUMiNATED SC-'".LES
~YEPIECE
.~NGLE
DEKI,OR
MICROSCOPE
EYEPIECE
GLASS SCREEN IN
FOCAL PLANE OF
COLLIMATING LENS
CAruM SCALE ACROSS
CENT!lE CF 5C?EEN
!-'ig. 4.12.
CO'IVERG:~IG RA"S
RE~lECTED;MAGE
f~CALE
FO'lM
Cf
ON SCRcE>1
COLLiMATING
LENS
~--
TOTAL ANGLE 36
___
____________._L-
4.521
Combination
80
gaugr;s.
8J
Decimal
Afilmtes
}I[inutes
3
9
IMAGE OF TILT
GRATICULE IN
FOCUS WITH _..{+---rTELESCOPE
FOCUSED ON
INFINITY
0-05
1
3
9
27
01
03
05
Fig, 4. 14.
The principal use or t11is instrument is in testing trle alignment accuracy of bearings,
loca.tion SUl'f~lCeS on
assembly futures, and work of a similar nature.
It Gon<s,s of two
unit and a focusing td;scopc, the body
of each of whieb is ground
and to a precise outsici<; diamet~r.
the oplical axis z:mc'. the mC<;;~1(;lnical axis arc coincident. Thus each unit
T:~ay be titted directly or ,hy preeisio~ bl!:;l~ng iDtc tWO
a comidcrnble
(lr~t:~LDce apart, ane! slgl!t~.ngs tnken tram tlle
unit to the colErnuting
unit.
The
mtit (;o:r:,ains a
grH LlcuJe in th,;; fcc::d
.is
as ;1 parallel
fncHs(;t3. at iriinic,: it
bring to ~] focus the
Ln
diItul'n Hae:; lit the
to
a second
1f ~he fOCllS of the
the d<ilum lines of th::;
ColHln,~ling unit of an
Most of the carly work in line standards of measurement WEtS carried out
Sociele
a Swiss firm 1.'vho have sincG
preciSe Jille
including the International
Thury, the
for this
n:aJized
precisely divided circular
and
high-class instruments
circular dividing machine
1865.
At this
there was no m~th.ocl of
~he SP:1Cillg of the teeth ar01md
the table of the
m:tchinc, and Thury set OlIt to usc the linear
as a form of ~ine bar to c:dibratc
machine, used for manufacture of line
the
The first divided scale used for circular division was 1.11 made. Since tIKi1
Imve been found for snch scales in optical rota,.y tables,
~ach othcr~
and
th,c
ThllS the Gccuracy
82
distnnc~';
i;)
with distance,
i3 in1JreRSed,
reduc(;d"
al~fo
83
I I
\ \
I \ \
:Fig.4.15.
ErmeR
(SECONDS)
Fig. 4.17.
NOTE: WHEN
OF
IN
Fig. 4.16.
i;![fO(S
eCCRJ1triG
by an amount e,
In the
84
85
T
,,"
Division
.,
~'
Correct
for Zero
(sec)
CUlJJulative
Position
mill
4
4
1
2
3
4
4
4
4
5
6
'7
8
9
10
11
12
(Repeat 0)
4
4
5
5
5
4
4
sec
Error
(,ec)
53
45
38
36
0
+025
-15
+050
,-17
075
39
42
-14
11
+ 1-00
48
+ 150
+ 1-75
01
- 5
+ 6
+15
+16
+ 8
56
+ 3
+2-00
+225
+2-50
2"75
50
+3-00
59
08
09
+ 1-25
ii
Actual
r!
Error
(sec)
0
7,/5
1450
-1625
-13-00
- 9-75
[:
!
350
l.
+ 775
+ 17-00
I:
1825
+ 1050
+ 575
., ~ ~ -\
1
pi .C~!,iPPOSeS
,;lat the
[s
perfectly accurate. None are
(lDSOl1!tclY uccnrate of course, but so long as the angular error at each face is
known then allowance can be made for tius error.
4.63
t~, 1,,,+...,.
._~ e,l.}U
eL,lUn
"""t)i(~ b
r'tc
"',0.
<".
'11'"
I,
\.1
ll"I''"
"""'<' A
dIet l~ '0
t.
l~l.L>~:;v~
,,-,
i)
anU.1
anu
,
iJ.,, - Ti.
- . f,J.1C
,
IS
T~}.~ ---.:..--~.--.:::
n
... (1)
I:Sc= -,'.,T+
[(1)
R':.
Rl
R 1)
Viewed
0- 45
lJlin
sec
mill
s'ec
min
325
487
338
45- 90
299
90-135
135-180
ii,
3
3
3
':11--/\)--'")
30-7
18Cl-225
J..'T
.--:::= ~~~
."1
86
E(R? -- R.)
,'1
315
_I- ~'.~.M_':::~-":;
11
3
3
07-2
10-2
41-0
253
225-270
Dividing through
(1)
Faces
thus
S + R 1 =.~ T + R3
If the complete
added
T is a constant, the
11
12
autocolIin:atofs, then,
<l:i
360 =T
or
l) J
")
,+-,;co1'1
r... \.,;I.'~
'v~
If the errors in a polygon an: not known th;-;n it can be calibrated using two
auto-co.E~inatGr3 set un 1'0
reflections from adjacent polygon faces as shown
dl~o'lPq7"'
L: JL
JOn
If
D
'
,ail.(I i'2
are tne
on rmto-collimatDfs 1 and :2 respectively, Sis
..
Arrangerr~cnt
Fig. 4.18.
Total
J :;,1
06-7
181
08)
J
"
-1-
1,7
- 67
47
25
87
degree of accuracy required, and parallelism can be. prodl:ced by a number ot~
means, surface grinding, spot grinding, etc.,
dependmg on the degr~c 01
accuracy required.
Let us assume that the block is accurately parallel but om of square. If a
parallel strip :Jfknown accura~y is c1an:p~d_ agains~ an
plate in sOIl1c.attitlt?c
Ilcaf the vertIcal, then gauge alocks of dIfienng SIze can
tr:lpped betvveen t,:(~
block and the pal'alld strip as in Fig. 4.19 (a).
3600
T=-- - -----n
n
25' 47'2"
-~
A-O
-I')
,,'
-J
134"
-'
l"(R -R)
,2
-t-(Ra-Rl)
STRIP
GAUGE BLOCI\S
1l.i
W'-~~~t~T
Thus the polygon can be calibrated 1;\iithout reference to any ::lilgular standard
and to a surprisingly high degree of accuracy. This is of course based 011 the fact
that a circle is a continnous function, i.c. whatever the value~ of the individual
of a polygon they total exaG,tly 360. This then provides us with il natural
stamhrd of angle and the maintenance of anguJar worl<jng standards such as
combinr,tion angle gltuges tS considerably helped by it.
4.7
8,-~--j
I . ___
YvLE).,SURElYIENT OF SQUARENESS
There is one particular angle which is probably more important than any other,
Le. 90. Over the years its importance has been emphasized by the fact that
spedal names--square, normal, and more particularly the
angle-have been
assig:1Ccl to it.
If a iathe cross-slide does not move at 90 to the spindle axis then a flat face
will not be produced
a facing operation. If a depth micrometer spindle is
not square to the !oc2.Ling
an incorrect measurement,
ar:osine error,
will b,; incurred. if the blade of c\ flttcr's square is not normal to the stock then
ma:in'~
will 11m fit togeU:;.ci' as they should. In fact i.f cl right
was unattainable to v{ithin ::1 dose: de;;ree of accuracy it is doubtful whether the achievements ()f modern science weuld have reached their presellt state of advancement.
Fortunatdy:::n angle 01'90 lends itself to accurate measurcmcllt, as in most
of the t~dtniques u:sed a
etTect is possible, An example of this lias already
been given in section 4.52 as a particular use of tb,~ angle dekkor. Often it is not
enough to test a
agaimt an engineer's square, Tbis simply shows
whether iUs a right
as near as can be
by -:ye, or not. N[G,'I,,) often the
amount of error, ei[h~r as rill
or as a line~lr m,:aSUf',;l1lent over a
length
Df
is required. III any caSe each result G,lU be readily converted to the other
form.
(a.)
Fig. 4.19.
SGR?ACE
The contact position of the slip gauge is noted and transferred across the
block, whkh is then turned through 1800 and gauge block l\ is now trapped at the
top contact
If the
is "quare a slip gauge equal in size to 13 should just fit at the
lower contact
If it docs not, bl~t ;J gauge block of size B J jU3t contacts
correctly then tb.::; difl'crcnce B -- 6 1 is double the error in squareness in ih'J worKpiece over the iength between the contact mrrrks.
A
but morc convenknt method, is to use a measudng instrument
designed for this
incorporating a dial test indicator or small comparator
head (lS shown in
4.20.
In thi,; ~ase the difl1::rence in the dial gauge
is double the squarelleSS
over a length of work
to the centre distance between the fixed contact and
the dial gauge.
88
89
Angular A!easuremellt
again to
4.:0 let us assume that the difference in dial gauge
is 012 TIun, i.e. the block is out of square by 006 mm over the length ~f
surface of the contact
and dial gauge c';;ntre distance. Correcting this value for
the complete
of surface let us say that the block is Gut
by 0075 mm.
To correct this, point A must have 0075 mm remcvedfrom it, nothing being taken
GiT D.
C must be brought 0075 mrn towards D, no metal being remQved
at B.
DIAL GAUGE READING RI
Circular Division
4.72
II
Optlea!
IVjfetbcd~1
of Checking t::.!iiMfl:re,Gc:3s
(a)
Fig. 'L20.
I'
I.
I:
r
I:
1"'
I!
III'
Ii
Ii
(b)
Flg~
4.21.
II
Ii
SURFACE PLATE
Fig. 'l22.
eITor
Ii'I
I
A readIng \S talcn ',vith the autocollimator with the square in t.he first
showl!. The square is then moved to oosition 2, and the au~o-colJjmat0r
again noted. The: angular error in squ~reness is ';al/the dift(;renc~ in tbe-two rc;adings.
It should he noted that apart [rom errors in reading the aWG-collim;: (1))'. :: p.l
human
c1e:luing contact
d;(~ ollll!
liiii!'
of error
of unequal dj"l;l';~CI. Tl;i:: .:\'1"\:1' (';, ,1
be eliIninateci
b.J"H
..
Squarcncss. lest
and
:~~T){"p ...
... -- ..'c7Y~2.';lIlr.r
1 ..... -... ;:; t~le ,/'.
"'~.
'n
--''-'oJ .. ) .
The actual error i.n squareness is the mean of the two results.
91
ikleir%gy}'r Fn;r[;u:crs
. It i.s of i~tcrest t? DOt;) that all of the methods of testing sql1areness refe .
to.m
sectIon
are
tundamental anrl~,_do not r"fer t ., "n"
sta,,'l~r,'
of UP!!I"
,....,~ed
d . . thIs
:l
1::.1
tf,
J
- J."",~<.A~.",J.
4.1.,.:; "' .. lllev
e}l"OG sO:t:,y I)U .ne l?ccuimr geometrIc properties of this important 'right' au!),' =
Other mCli1~ds of testmg are available but have purposely beea ignored since d~
to ~he fact taat the~ depend .on the accuracy of ancillary equipment, errors'm e
For lDstance m Chapter
6. dea1i;}"< :with
mc,Cj"'l'ne
tu' 01 me. j ro,e)gy
ay
be(.': mtroduced.
"
- '
:..
. - ,
' 1,' '
rClcrence IS made to the pentarronal
D"i~m Ot' 'CF\tic'll
~F1U'"
"<0' 11
c O.r ene
;,:::.
-"
i: '"' \.
.....
'-1
,Ci\.lO
t;he squarencss of machine tool slideways. It should b,~ noted that this metho~
<'lcpends on the accuracy of the prism ,wd if the apgie bt'tw~"n ;tC' ,"Q~ctl'n",
f',,,
'\"0
is not 45" 1he error incurred will hQ double the error in tho prism.
..
0
.f
J/i
..
-),.
... ..;
t.;v- ""'.
(;
........~,
,~
........
CHAPTER 5
f.<.l......
.......,..
"
"""",
.~,'~.::s
~-
93
If then a part cannot be made exactly torts nominal size how can two parts
be made to fit together in the required manner?
.
There are three possible answers to this question, each method having its
place in industry.
5.H
(a) Interference fit. The minimum permitted diameter of the shaft is larger
than. the maximum allowable diameter of The hole.
l'bJdng to Suit
This technique simply flXJ.uires one part to be made to its nominal size as accurately as is economically ,,~asonable. The other part is then machined away a small
amount at a time and the parts orTered until they fit in the required manner.
This method ma.y be useci for 'one off' jobs, toolroorn work and so Gil, where both
parts will be replaced at once. As such it is of littlE: conCt,Tl1 in this book.'
..L-.
5J.2
Sekctin Ass:mhly
llJ
N
iii
tV
I'J
It is sometimes found that it,is not economic to manufacture parts to the required
bigh degree of accuracy for their correct functioning. Inst~ad they are made in
an economic maDner, measured to the required accuracy and graded, or sorted
Into gl"()Up~ each ofwlLlch contains parts of the same size to within close limits.
'riley aTC thw assembled with matieg parts which ha'ie becn simj]arly graded.
A good e;~ample of this system o'~curs in ball-bearing manufacture. A bail
bearing consists essentiatiy of an imler ring, and an outer ring, separated by steel
balls. Both types of ring and the balls are graded automatically and when assembied the follo'iving conditions can be allowed:
iii
ltL
:r:
<.(
...J
Ul
11J
N
tI)
ItL
:r
V)
><
,
L:l
l
SHAFT
HOLE
(a) INTERFERENCE FIT
(a) LU'ge balls are ,i;sscmbicd 'r.to small inner and larger outer rings.
(0) 0,f,cdium balls ,E'e ,lssernbled in i.o medium. inner and '.)Uter rings, let cgc
inner and outer ifings, or sJl1all inner and outer rings.
(c) Small balls are assembled into large inner and SU1o.11 outer rings.
Usually a selective <.'Lsseml11y system is used where the assembiy is replaced
as a u:'lit rather thml rep"\acing separate parts. For example, if a bali cracks in a
bea(ing the whole bearing is repl.Qc,~cl, not just one ball.
~l
CLEARANCE OF<
.jINTERFERENCE
f:7.> ..
II
HOLE
5.13
17.~;pctiti\1C
. "c,.:1
F'-~~~_.d ~ ~--~-4
~_rb
94
---I
Fig. 5.2.
The cii:;.gr8.m in Fig, 5.2 also i.llustrates the terms used ill this type of work
definitions of these term:') Oei!lg as follows:
95
ivletrology far
:~-ilgince/'s
(a) Limits of size. These are the maximum and minimum sizes allowed '~n a
given component.
Further it must be realized that larger sizes require greater tolerances, and
consideration must be given to this. A working system of limits and fits is therefore
fairly co.mplex if a manufacturing organization is setting one up from ilrst principles. Fortunately there are standard systems already available.
(b) Tolerance is the maximum size variation which will be tolerated on a given
component.
(c) Allowance is the size difference betweetl the limiting conditions of, size
on the two cbmponents. It should be noted that a positive allowance
produces a clearance fit, and a negativ(} alIowance "Jroduees interference.
A given manufacturing organization will require a number of different types
of fit, ranging from light driv(} fits requiring shrinkage or hydraulic pressure for
assembly, through keying fits for locations which can be ,1ssembled with light
mechanical pressure, to running fits and clearance fits, Such a series of fits can be
obtained using two distinct policies or basic methods,
(a) Hole basis system. For a given nomInal size the limits on the hole are
constant and a series of fits are obtained by varying the limits on
the shafts.
Thus assluning a hole of dimension 25 mm +002 mm
- 0000 mm, a sh~ft of
and all of these fits are obtained with a common diarn;;ter hole.
(b) Shqfc basis system. [n the "aTIle way a series of fits can be auangcd for a
given Dominal size using a standard shaft and v8.rying the limits on the
hoIe~.
For most work a hole basis system is used because a great many
ho1.es are produced by standard tooling, e.g. ICamel'S, drHls, etc., whose
size is not adjustable, Thus for a nominal 25 mm size a 25 mm reamer [s
used to produce a standard hole, the shaft sizes being more readily' variable'
abou.t the nominal size
machine adjustments, e.g. roller box, centreless
grinder, etc.
It must be realized that a singJe class of hole <lccuracy is not normally sufficient to rr:..eet the requirenlcnts nf ~ln clacses of work. 'There is no porn!: in renrning a
hole just to utilize one S(;[ lif limits \VhCfl th;, Qccuracy of l]. drilled hole wiII do,
111L,3 a given SY~l,c.m WOUld have
four cl,!::ses of hoic, the tolerances on
vlhich al i<;Yf ror fOt~r dilfel'cnt
method,;. [I. ;night also reqtl.lrc, say,
nine diJrcrell~ d3.~;ses of shafts to giv;,; nine ditIerent types Of fit when associated
'{lith a
dass of hole.
96
5.131
Ii
Ii
I
t
This system is very popular because of its simplicity. It is a hole basis system
providing two cl:sses ~ of ~oles for diiIerent degr~es oJ- precision ~f pr.o?uction.
Limits are also gIven ror SIX classes of shaft; two mterIerencc, one cransltion, and
three running fits. These limits arc specifIed for a wide range of sizes of work, the
whole system being set out in the for~n of a table which makes it very convenient
to use.
5.132 The British Standard System (B.S. 4500: 1969)
B.S. 4500: 1969 is a cOillprehensive system designed to cater for all classes
of worI< from instruments and gauges to large heavy engineering. At first it appears
to be an extremely complex system until it is realized that no one organization
should attempt to usc all of it. Iilste[td a company selects from within the standard
system a 'sub-system' to suit its own requirements and manufacturing techniques.
The system provides 28 type;; of hole designated by a capjtal1ew~r A, B, C, D,
.. " etc. and 28 types of shaft designated by a lower case letter, a, b, C, d, . ,., etc.
These letters define the position of the tolerance zone relative to the nomin;]! size.
For instance , all class H holes havc limits of +
x. Similarly class h shafts have
-0
, 0 f + 0 and -so
nnIts
I
-x
011.
Holes: H4
Shafts:
s6,
H7
H9
p6
1116,
L_-----.;---'
Interference
L-~ ..
Hll
k6
"'-.------..J
"--------v---------..J
Transition
Clearance
(used in
association
with H7 hole).
97
G.AUGS
~"r~
",
., 'shafts.
\l i\OMI~L
Consider first a hole on which the limits on diameter are specified. It would
appear that quite simply the 'GO' gauge is ,~ cyliad.er whose diameter is equal to
the minimum hole size, and that the "NOT GO' gauge is a similar cylillder equal In
dIameter to trle m,lximum hole size. Unfortunately it is not as simple as tillS, for
the same reason that limits of size are
for the 'l{ork; rrothing can be .made
to an c;,aci: size and this includes gauges.
the gauge maker requin:s a tolerance
to which he
and the positioning of this gauge tolcnwce rdative to the
nominal gauge
requires a policy decision. For
If :he gauge tolerance
increases the size of a 'GO'
gange, and decreases the sizet; of the 'NOT GO' end,
the gauge will tend to
good work which is nC:lf the upper or lower size
limits.
Similarly if the gauge tolerance increases the size of the 'NOT GO' plug gauge
and decreas('.s Lhe size of the 'GO' end then the gauge v,rill tend to accept work
which is just olltside the specilled limits.
Obviously a firm which is qu,llity conscious would use the system which
would reject doubtful work but it must be realized that the gauge tolel'llJlce is
approximately 10 % of the w06;: tolerance. Therefore, using tllis system thf; production department is deprived of 20 % of its allowed tolerance_ This pwblcm
has be~ll dealt with in B.S.
which lays dO\m the
pciicy for the
0,'
OIA.
\
V
II'
!
,.
NOlvilNAL Dlt\METER
CF'GO'GAUGE
MAXiMUM
HOLE
D!AM::TER
\!
II
!_._~_~J
(NOT GC\ PLUG
GAUGi::
Fig 5.3.
I!L!
_I
Is--_____________----lI'
l)i~posltion of tolenu1c<:-s on
plug gauges.
RiNG
(;A.U~;l::
The tolerance on the 'GO' gaug\; shall be within the w'Jrk tolerance zone.
The tolerance on t:1C 'NOT GO' gauge shall be outside the work tolerance zone.
This poEcy
fOf the
Thus a wear allo\'\'8.nce is made
which is ~quat to
20 ~/~ of
gange tolerance and is itc1ded to, or
subtracted frorr}) the n0L11inal gauge size in
to the direction cf YVt::ar~
It ~hould be noted that the vvc:u~ QUo~.vance bccome$ teo slnaU to be useful 1:vhen
che v/ork toleraI!cc is less than o t'r:i1m and no al1o\van:.::e is 111aJe for
init~,ll
ll"W
\v!1osc tolcran.cl'; is less than this. J'\Tor is the ,,\fear aHo\va.nce applied to
ShOE1 J
littte "\-~'car in use~
~:NOT GO'
t\OMIN.:\ISIZE OF
'GO'
GAUGE
StZE OF
'NOT GD'
Oft
t;ic
\\7 car
allo'Nanc(:!~20%
of gauge tolerance to
n~afcst
o-oot
tJ.lrtl
unit
99
98
""
In B.S. 969 it ,vill be found 'lilat tolerances on plug gauges are in some cases
less than those for ring and gap gauges. This is due to Lhe fact that plug gauges
lend themselves more readily than do ring and gap gauges to measurements
where the order of accuracy is 0-001 rom or less_
Thus the dimensions of simple gauges may be dealt with. However, in any
design problem, the geometric form of the component, in this case a gauge, must
be considered, and under conditions in which the diai1Jetral tolerance is small,
this assumes considerable signnkance.
"\."}It,;"lC""
FU;"'L
MAXIMUM
8READTH,
The 'GO' gauge is used to ensure that the mm~imum metal condition is 110t
exceeded and that metal does Rvtencroacn into the mininn,m allowable hole
It should therefore be rcl'adc to the maximum allowable m;~tbl condition
due alloV/ance
mad!;) f:)f wear :.tnd the gauge r:olemnc';~ as in
:iection 5.2.
Now consider 'he 'NOT GO' gnuge. If this was made to gauge both dimensioJls
at' the! minimum metal conditiOlb: (maxirrlllF:i ~lOJe si.~c) a (;o7lriition :would arise.
where the breadth of the hoie is withill the
limits, but (l1e icngtl1 is oversiz',:;, as in Fig. 5.6.
~',',.
100
'
This principle should be applied to all systems of limit gauges, and where
possible ~his is o.one. In the simplest case of a 'GO-NOT GO'
gauge it would
'not appear to be so, but to a large extent it is. The 'NOT GO' gauge is always
relatively short, and approximately
equal in lengtb to ille hole diameter. Thc 'GO' gauge should
where possibJe be equal in length
to about three or four diameters.
In addition to more readily dis"
tinguishing between 'NOT GO' and
'GO' ends of the gauges, the length of
the '00' gauge ensures 1hat the mal:i.
mum !)letal condirion is not exceeded
Fig. 5.7. Length of 'GO' r:1ug gauge pNvcnts
it entering 3 non-straiLtht hole.
due to geometric errors in the work,
e.g. straightness as shown in Fig. 5.7.
A .similar situation occurs when gauging shafts. ideally a nIH form 'GO'
gauge, I.e. a ring gauge of reasonable iength in relation to its diameter. should
be used, in ;;onjunctiol1 with a 'NOT GO' gap gaugc. In practice both ,~o' and
'NOT GO' gap gaus.;s arc frequently used, but it is advisable to SUIJplement these
. gauge, to be used at frequent intervals. This is particularly true
;"'1'tJ1 a 'GO, nng
In the case of centrdess ground work which is liable to a conditiOi1 known as
lobing.
101
clears the metal. This chec:(s the low limit of size. The gauge is then rocked back
until the lilgh point prevents further movement and thus the ;:naximum diamet:r
is tested. If the raised portion passes tbough the hole then the work is rejected as
the hole is oversize.
J1iJetrologyfor Engineers
The simplest lobed condition has three lobes and is based on an equilateral
triangle as shown in Fig.
but it may occur "vith any odd number of lobes.
From each corner of the triangle let radii of rand R be struck, the lar"e
radii blending W.itll tht: small radii struck from the other two angles.
/;>
THREE-LOBED
j\
" \
R tr IS THE
ClIMENSON
/
,0
"~~
WhiCH WILL
)~~y ~~~\g~~~
~~"
'-t-------
'\.~/
~
,~,~~~ ___~_.L
I
Fig. 5.il.
., .
531
ie-to Gauges
These
afC a particlllar
Co. Ltd., in Sweden.
normal
th,~ 'GO'
r.oler~Ync(;
limit
Oil
It is inserted into the [lOle by tilting tbf, gauge forward so that the h.igh portion
102
..
It is ofillterest to note that the 'NOT GO' gauge conforms to T<lylor's principle,
and only g3.uges one diameter at a till~e, i.e. it can check
However, th.e
'GO' gauge, unlike a normal plug gange, cannot reject a hole wbose diameter is
within the allowance
out whose \L;k of straightness caus::S the maximum
metal condition to be exceeded.
SA SCREW THREADS
The allowable errors in commercial screw threads arc quoted in diameter 0111y,
i.e. for a thre:.l.d of a given pitch and form, tokrances on
o..nJ. flaILk
are
not given, but Qrc inch;.ded in the tolerance
on
diameter. BrieHy
the reason for this is that errors ill Dicch and fi::lI1k angle
'lbom a virtual
increase in the effective diameter as' noted in ChqJter 8. Simiiaf errors on an
internal thread hin.!! abolll a virtual decrease in effective diameter.
The effect of these errors is more fully dealt wiIh in Chapter 8, dealing with
screw thread measurements, but the fact tl{:lt a pitch error, or a Ylan.,1< angle error,
can cause an
chJnQ"c in effcctive di;tm~ter is most imlJortanl in
con
sidemtion of the
ofscrcw Emit gauge, and calls for' lhe applicat
or
Taylor's
Consider first a 'GO-NOT GO' type screw gap gauge in which both 'GO' and 'NOT GO'
gauges are made to the full form of the thread, and both of which engage the full
length of the thread. It must be emphasized here that such a gauge is incorrectly
designed, but a consideration of 1tS faults will lead to an explanation of how these
faults can be corrected by ccrrectly designing the gauges.
1f such a gauge is to be used for inspecting work threaded by means of a selfopening die head, as it frequently is, it wllI also be of assistance to the machine
setter in the initIal setting of me die head. Consider in this case that the die head
is causing a pitch error.
If the diameter is cotrect the work will not enter the 'GO' gauge because of the
pitch error, although it a.ppears that the diameter is too large, The die head would
the1l be adjusted to reduce the thread diameter until it enters the 'GO' gauge, but
not the '"OT GO' gauge. The work thus appears correct, not because the diameter
is correct, it is in fact under:;r.ze, but because of the diametral compensation by tue
pitch error.
- Tris is an outstanding e."l:arnple of gauges not conforming to Taylor's principle.
If they arc redesigned accordi"~g to this principle, the gauges required would be
as follows:
(a) 'GO' gauge~Full fm:m and full length of thread to the maximum diameter.
of the thread.
(b) 'NOT 00' gauges-(i) A gauge for the major diameter.
(1i) A separate gauge for the effective diameter and one which is not
influenced by pitch (o::rror8.
Ideally the 'GO' gauge shotlld be (l full form ring gauge, but these arc cumbersome !n use, and in practicer a full form
is used, a ring gauge being kept
for the periodic checks. The plain ';;lor
gap gauge is oftcn omitted so that a
simple 'GO-NOT GO' screw g<tp gauge can be used,
To prevent the 'NOT GO' g:~'1;lge beillg influenced by pitch errors, it gauges on a
short length of thread only, and checks effective diameter only, by
cut away
at its crests and roots. Thus the anvils of a 'NOT GO' screw gap gauge appear as
shown in Fig. 5.10.
Consider now the w{lrkpieoe haVIng a pitch error,
gauged by a gr,p gauge
having anvils of tins type. The dtametcr is reduced until th,~ dlfcad passes the '00'
gc.uge. When offered to the ')lOT GO' anvils it enters :J.nd passes the11l because the
diameter is undersize, and tiJ,.e
length of thread is not long: enough for the
pitch error to be seriously effective.
Similarly, for gauging rnt'.;rnal threads, an ideaL set of plug gauges \yould
consist of
(a) '00'
i:
\'1
i,
:.\'
g~lUge~Fun
~,
AT CREST
AND ROOT
5.411
GO'
screw..gap gauge.
diameter.
.'.,
.- . 1 d;~' oa wear "llowance added to Ie> nomiHa .. ,m,. o.
Thus a screw P1ug gauge lla"
"..
.'
. ,
,. t
.
Beyond tl1i" )'S " tolf'rance zone for simple effective dmmeter wrw:;[l C~!1I:(jt De
SIan
,,'
<..
~
11
f
,. - ,,-d
,~. ded 'DU'" r.oa'cd to this is a further tolerance zone to 3.0W or PLC;l .,',
eXL;C~
.
d'
. Tl
tolara[1~e
flank angle eHOTS in terms of an increase in effectIve tamewr. .lese "".. 1 0.;
zones are shown in Fig. 5.11.
. ' ~ 't' nal
Similar tole'ance ZOiles <,xc
for gap .find nng g;;u~es tor t"X Lf
", d"iJ a"l "i'lt
set out In tabular IOun.
t1le magr:'.lLIl
. v tolera'-'c,'
_
.. v b"':n"
Y.< b
A
ana1
-.\,'
l.llClr
.
'. d~ for highly accurate
rnagmcu.v
. "
d
q
aU"t1 t esf mg
'~-~w
~Llv
!Zap :raurre
.'"
l>
.~
-,
an'
.,
< ,
'"
'''A
i,4.J.l 1..';)
',-
c,', '7 __
105
104
fOlERANCE ON Er"ECTiVE 0
~O
cr-!
IAMETER
Vl"'R
CJ _ ANCF. ON SIMPI.E EFFECTIVE
-=f\Jl\I
~---i--
f~;;~~g~~~~~.:..-:;:;:D:'A~METE'I'?ALLOWANCE
LJ
=:
Fig. 5,:2.
Fig. 5.1I.
small diameter of Ii taper plug gauge, These measurernents ,He fuHy 0xplained in
Chapter 4~
tl~js rneans that
the manufaci:urc of su.ch ga.ug t:; it is
not pO$sibl~ to 'neaslIre the required sizes directly,
Consider a taper plug gauge. If the
diameter D is det~rD1incd, rh!~ diamct'::Is
D ms " and
afc functions of the h;;igilt of'
of taper .
gauge and the
~):an~
Further, the
of til2 ,'/OTK to be
\vill have a tolerance, and so vvi11 the
gauge. If the
tolerance on th'~ war:,
is say, :1: 5 sec, lhe tolerance 011 the
should be of the order of 1 sec (Le. 10%
the work tolerance), or A sec.
If D is correct, then over a height
this angular gauge tolerilf).ce -w'ill make
HiI
Fig. 5.)3.
be used.
(a) ]vIakr; the gauge o'verlength and to an
nOJninar -\vork
(b) Remove the gauge from the machine and accurately determine
D and 1he angle
106
diiillld:1'
107
will
give the correct diameter Den,x' Similarly calculate the height of the step S
wrJch will give the correct diameter D min'
Grind the length to the correct values of Hand S to the required degree
of accuracy for Dm and D min
A similar procedure may be used to produce taper ring gauges to the required
degree of accuracy.
To use taper gauges to check the accuracy of ang[e of the workpiece a line
along the gauge should be smeared lightly with a marking agent s]lch as -prussiau
blue Of rouge. The g'.luge is placed in position on the taper and rotated slightly.
The registration, a transfer of the marking agent on to the work, gives a slibjective
indication of the an2:;,'ular accuracy of the work.
however, will not give a
measurement of the angular error and the effort WOt:id probably be better spent
in <::nsuring that the machine alignments necessary to produce the taper are
correct: this techniqIJ.oc is obviously not suitabie for a production check and can
ordy be u:::l~d for setting purposes and as a spot check during production.
5.71
G~mging
rM.e gauge used for this purpose is a. 'pin' gauge whose lengL~ has been accurately
J.JJ.
"
,
~'~hirl" ("ce Chn;l)ter 3).
d"termtneo mang a iengi.ll oar mc"sulhlg me..- '~", ~,l' " ' w f" bore
v
T'le l~rl'Jth of Ule
"auge should be smaller Wan t'.1e o.lamel<:. 0 dle. A'
D '(1":
'"
b
t .. "lr about pOlnt
Ttl." m'oort o "; jn the bore and the amoun 01 l'Ov ~ <
.i
to b e galLge ..... ""
~,,~~. > - .
'
r:;.'. ~ 1 measured with a llexible steel rule or tape; see J. If? ) .. ).
1
"
,.',.
F ...C
5.7
'..
~.~
[~~1TT'
:
--11. /
From
Be
VI
I
I
Fig. 5,14.
The g;:uges used for dete~'mining the diameter of iarge workpiece!> are not, strictly
speaking, lirrtlt gauges altnongh they may be used 2S such.
Or
109,
108
5.16.
as ]n
a
where L :is the length of the pin gaug;} and W is the a.menn, of rock
and as D=L+3
W2
then D=L+8L
or W=V{D-L)8L
If th('; Ihaiting values of ,0, and the known length L are inserted into this
expression, then th'~ limits allowed on the amoLlnt of rock m<:\y be derived from
this final expression.
The aCCl!.T!lcy which carl be achieved by this method is of a much higher degree
than can be obtain;,d using an imtrumcnt sl1::;h as a vernier calliper to make the
measurement direct. Consider a gauge of 400 mm
exhibiting 50 mm rock.
Then
2500
8 x4ll0
,=400 +078 mm
C\
D='h,O+---.- mm
The
1S
me~snremt;'nt
Me"su"emcnt
of larg() external radii.
a!
If
(51 )2
D = 400 +--~--.- mm
;5 x 400
=4CO
To carry out these Il1(;ast!remcnts an angle; piece m2cy be mude up, the Induded
.kno'l,vI1 . T'lris may be fitt~;d \:Vittl a n1icl'o"
fr0rn
Fig. 5.16,
cos (90,8)
5,72
(J
an error of
Thus 2m induced error of 1 mm in the amount of rock
O'03 mm in the d~am(t'Cr measnfcn:,ent, and on a diameter of ';00 mm this would
be difficult to improve upon with simple direct measuring equipment.
110
~.
F :19....."16
(1)
f)\'
In this expression for R the term cos V(no -.)
15 a
r'"
~')l,.
('f'7tn~rG
,,' ,co
HK
R=(l
IC)
111
LItnits
H=h+Kl
III
and
A high carbon steel is fully hardened by healiGg to 730 lmd quenching in water.
This will
a hardness of appwxirl1atciy 64 on the Rockwcil 'C' scale, but it
will also make the steel extremely brittle. It is J:ccc:;sary to
tl10 gauge to
reduce the brittleness,. but not to make it so son as to :'cduce its 'Ncar re.si.stlwr:;C.
At the same time the tempering tre:tlment can be used to stabilize the IDGuerlal
and relieve any intern::l stresses Wllich [nay distort i~ over a period of time.
A tempering temperature o( 2UO"C vI/ill reduce the brittleness so then the)
C>
'
'1 ' L4, 5b. li',. ~,m!.:>
.
gauge not
to
iI.nd the har d ness va Iue Wl'Jl. 1oe h.oeb:weJ
temperature of 200C is maintained over a period of 8 to 10 hours it will also
make the gauge extremely stahle.
Screvi thread g;mgcs arc particuiady fragile and pronG to damage if rOl.1J.;hiy
handled, and these gauges should be 'let dm,vn' furti1:.::r at a temperatur,;: of 240"j::::
to
a Rockwell hardness of 'C' 52.
'.,.
.''!' '
R 6
.lvrucb. can be done to reduce the initial wear rate of a gauge if its surface finish is
good. A poor finish with a small number of high peaks is prone to more rapid wear
than a finish having a large number of very small peaks, giv:ing a large contact
area. Thus a gauge should be finished by high quality grinding or lapping to give
a C.L.A. value of not more than 010 Ji-m.
114
115
'Pilie tests applied to machine tools, regardless of type, fall into well deiined !TroUDS
whlch may be summarized as follows:
<:>
(a) The level of instaiIation of the macrJne in L'1e horizontal and vertical
planes.
(b) The main spindle is tested for axial movement and in the running truth of
its axis.
(c) The bed-ways are tested for straightness and parallelism.
(d) The line of movement of members such as saddles and tables alorlg bedways is tested.
(e) Practical tests in the form of the machining of test specimens, followed by
line, and it would therefore be impossible to turn a true geometric cylinder on Lll0
lathe. Similar effects are obtained 011 other types of machines when the accuracy
ofinstallation is not of the required order.
The maintenance of the initial installation accuracy is dependent upon the
, type and thickness of foundation on w1:1ich the machine is set. Many machines
require no more than the norma! thickness of workshop floor, say 150-200 mm of
concrete laid on hardcore, Lo
cases. such as jig boring machines and high
grinders, it is necessary to prepare a special concrete foundation of
considerable
which addiLional1y may be insulated from the surrounding
fioor.
6.1 in which is outlined the
The method of testing for level is shown in
bed and bed-ways of a centrl3 lathe.
BRIDGE
by
6.21
It
116
SPIRIT LEVEL
117
'l'j.I'
lvlachine
accurate work. The equipment required [or this test consists of:
(a) Hardened ground paraliel test bar havillg a concentric
is a dose fit to the spindle nose taper.
shank which
of the
plunger is then arranged to lie on the
axis of rotation.
The slip should not exceed 001 mill.
.
. ,
It is imr;mtznt to distinguish oet';'reen the necessary end play)!: [he spmole,
.:l e to
,,... ,,;
><'-"'1'
'oet"\""~l1
hea1'ing
and axial slip. The latter is due to
til!
,. t 1I...J
I,..,J1Y
.. ~~o.. !'"C"
...
' c ' ......
~~
errors in
of the ,clbutm<;;nt faces or locating faces ?f end
,:vltn
respect t'G
axis of rotation, and produces a
cnd-wlse movement Of thf"
spindle.
I...t
10,,<
,j.J
""
"
6.3.
Fig. 6.2.
Two readings, at I flnd n, are taken of the running t;:uth of the test bar,
which obviously must itself be a highly accurate one in all respects. Over a distance
sllch as JOO mm be~ween I and II, a difference of running truth is allowed. At I
tl1'c' ,~rror should not exceed 001 nun and at II, 003 mm.
j\ most
test is that of the paniklism of the spindle axis with i'cspcct
to the
in both the vertical and horizontal planes. The method of test is
shown in
6.2. ,<\ny error in p:tral!cEsm whi.ch is revealed \vhen the saudlc is
moved along the bed should l10r exceed 002 nUll/300 mm in either phne. In the
vertical plane, shown at
the reading ma~' be in the plus directioIl at the free i;nd
of the test bar. Ia the case of ihe horizontal plane, shown at (0).. any c.rror should
incline the <est bar towmds the directior. of the rool pressure. By permitting only
these directions of error, the tendency is for the !ccl pressures to correct them.
The axial
or fioat of the spindte is teSted as in
5.3. the dial test
indicator
firmly mC1Jn ted at any suitab;e position 011 the machine.
must be a snug fit into the spindie no,:;e laper, and have its
Outer eIld hc\.: ground flat and sqmlf<~ to the taper axis. The Ene of movement
118
Chapter 4,
with 11() 'v.
Assume
of (l lathe bed 2 In in
i.nGiISUreG.
the auto . .
'.)1 measurement 'would be as
or the lathe
the
projectcu
oEth~ bed~
A
of the tripod type,
fDr tIns.
AS'TPli'1. 7
:'c':
havc
the
reflector is set 011
b~'d jn~E:r~r tlt;' i~sirument and u rcHectioil obtained [!'Om it such tl;at t,h~ ,image
'u,' "P5 01' ill'" c(~J1ima[or ilpl)e8~r nearer tIlt: centre of L1e fleIG, The
.t t;"~ C-()-'"
.0,)'n.l.....
0 ..
reflector is thel moved to 1he other end of the bed, Clnd.
the
line of moveme;lt of the rCJ":;~ctor has been reasonably
. , .'
r'
of Hw
then the
of the cross-wires
:,~ppear m tDe fIeld 0[. '.Ii.'::
r 'c''',
Of~ tl", ,"rl'p~tor aise Tills ofoceclure ensures that rettt~Clwn;,
ey
. e)l
l;-C e<..i.. !-.u1-:>
~
a:t i~terme(bte positions wiII ;:e with.in the field, and is thus an
ened;
on the level of the bed in the horizontal plane.
.1.1 ......
1.
"C'; ".',.
the
'-~
,_, _ . . .
.l'.... LvI
......-
.1 _
119
1\11etrolob~
0" V for EllO"ineers
'"
A straight-edge should now be set down on the bed, to ensure that the reflector
is stepped along it in a straight line.
Assume that the distance between the support feet of the rdlector is 1035 rom,
and that LTJe interval length at which measurements are taken is also 1035 nun.
2-;r
=-~-----
ra d'lans
___-_"rr_x 1035
mm
=0-03 mm
That is each tilt of 1 min of arc of the reflector as it is stepped along the bedway corres~onds very closely to a rise or fall of the guide-way suri~tcc of 003 mill.
Having ensured that an
of the croGs-wires will be received by the autocollirn<:!tor when the reRector is set at the end positions of the bed, the reflector is
now set at th,; forward end ofille bed. nearest the instrument, to begin the series
of readings. This condition, and those for subsequent readings, is "hown in fig_ 6.4
in which the rise and fall of the bed surface is greatly exaggerated.
,,'_~I
be done. Column 3 gives the differences of each reading [rom the first. In column 4these diuerences arc converted to the corresponding linear rise or fall, on the
basi~ of 1 sec of n.fC =00005 mm per lO35 mm. The second zero introduced at the
head of column 4, when associated wilh the previous zero in tIlls column reDresents the heights of the t'1iO feet of the reflector support mounting vbe~ if; its
original position. Column 5 gives tlJ.e lleights of the support feet of ihe refl.ector
above the da',um line drawn through their first pOSition. Thar is, the values in
column 5 ,.re obtained
successiveJy adding, algebraically, the values in column 4.
This is necessary because the individual heights obtained ill column 4 am tbe
heights of the b3,ck feet of thi': support relative to the front feet in a given
and nN relative to the datum.
Position
Adjustment
to bring
both ends
lO Zero
Reading
on
Surface
AUTO-COLLIMATOR
I
2
2
2
2
2
2
2
2
.2
2
2
4
5
Fig_ 6.4.
\Nith the reiIecror set at A--R, the setticg wires in the auto-collhnator eyepiece
are moved to straddle symmetrically the im:lge of the horizontal cross-wIre, by
th;~ suitable rotation of the micrometer drum, and the micrometer reading is noted.
The reflector is then moved 103-5 mm to the position B-C and a second reading
is taken on the micrometer drum. Successive reading.s at CoD, D-E, E-F, etc.,
aoe taken until the length of the bed has been stepped along, A second set of read
iml:s should now be obtained by stepping the reflector in the reverse direction
al;ng the bed, to fl;veal any serious errors in the first set of readings, Assuming
none have
the mean values of each ~et of readings ;Tny now be reGorded,
and these
the anrou!ar positions of the ref1ectof, in seconds relative to
the opticalrxiJ
the auto=collin;ator at each of its posilioll::; along the bed.
The merhod of tabulation of the results of measuremem are shown OIl p. 121the position of the plane reflector at 1035 rom intervals along
Colum!J 1
t
the bed. CoiGmn 2 gives the mean reading of the auto-cQllimator to the nea.res
second_ In practice it is possible to observe wb-divisions Of seconds, and this should
8
9
10
II
12
10
10
12
15
17
18
17
15
13
9
12
14
15
i;1(:
mill sec
l!rrors
from
SrrctigJlt
-,-"
,,,;
-:-5
-,-3
+25
+35
+4-0
+3-5
+25
-1- 1-5
-1
-05
+5
+7
~8
-7
0
0
+ 1,0
+ 3-5
+ 7-0
+ I I ,0
+ 145
+ 17-0
+ 18,5
+18,0
0
2-6
40
6-0
- 8,0
-10-0
-12,0
14-0
-160
-18-0
-200
-220
-240
-2
-3
25
-1-0
+ 10
+2'5
+3-0
+2-5
0
1'0
-1-0
0
or
120
The total rise in the surface of the bed over a n m length from a chtum
along the line of the fir~i reading is 24 Itm. In cGiun;n 6 thi; total rise is moportioned o",:r the t'Nej\'e readings taken, i.e. in increments of 24/12 = 2 ~J-m.
. These values lGolulUn 6) ure subtracted from the values in column 5 to give the
errors (column 7) in the bed from a straight line joiniDg the end points and wilhin
w~ich the s(:rJ!;s of readillgs were obtuined (I.e. it is as thongh a straight-edge were
laJd along the bed prorl c and touching the end points of the test surface when the'!
are in a ho,izontal jJ18.nt). Th~ rise <wd feJI of the 5urfac(~ relative to the straidrt..
edge would be the values given in column 7
~
A graphical 1'0prescntatioll of this is silOwn in Fig. 6.5 in which the values
rp
1. .... j
. lvlachine
5
POSI;
O~
8
7
6
ON SURFACE
10
II
12
6,3
XLS'IS l?()R
nl~ichilJ.c
.j.~~ C~l.se
or(~~;r
tHb]<~
Fig. 6.5.
122
of 0-05
n;i.n/:~CO 111Dl,
(: f: <l rnllf'! g
a::!cl in
(l~f~c~ion
is inc;inGcl
th\~ farce
v;/ard travers':: of the d~'iIllends to correct tile error~
In the CELse of v!~rLi'~D..l
the flin t~st \vould
ib:ccl
hC:ld nnchinc:;. A test ill d. direction
11al tab1e axis on:y j.;.:; D.\~rr=.1ficahle to
0,02
:rnrn, .1s appti(~~.r.bL:;, 'che front
of rhe table
tills l1atllr<; is knciwn as tIl;;: 'turn rOlE.ci melhod'" ancl proves t<,l 1')1; :1
very sensItive one.
In some classes of machine
the accuracy of constn:ctiun cuHc(; fer
straight Ii is necessary only to take two readings; one with the reflector at position
A, and a second at position B, the optical square being set down at the inter~
section of L'1e two surfaces when the reading at B is taken. The liinbrenc~ between
the two readings would then be the squareuess error, note
made of its
direction.
The test becomes more
however, if the straightness errors in t.he
two surfaces are considered. In this case, it is necessary to carry out a straif!htncss
test on each surface, as shown ill section 6.23. The "angle f~nned by th; mean
straight lines
through each surface profile
then indicate the mean
angle between the smfllces.
A squD.renes~ error, which reveals itself as an f.L'cial mO'lelJ1!:mt, is that of
the thrust face and collars of a lead screw. Unless these are perfectly square to the
axis of lead-screw' rutation, a cyclie end "wise movement is set l~p which is of the
same nature, and due to the same reasons as the axial slip in a main spinJIe.
The condition is shown in Fig. 6.8.
The two methods of testing are:
requires optical methods for its testing. Trus applies especiaLly to jig"boring
machines, but is also applicable to milling machi,lies of both vertical and horizontal
types.
6.7 illustrates the conditions of te.st fOI,the squareness of the transverse
table ways with the face of the column.
wm
<'.
COLUjvlN
GUIDES
Fig. 6.6.
Method (a) is sim!lar ill all respects to that used for the axid
(if main
spindles and Lllc error which wouid be tolerated in the case of htlJ.:': lead ssre'iVS
would be of the order of 001 mrn.
The importance of this test inay be appreciated when it is realized that axid
slip will canse a tlu:;ad
a periodic pitch error to be cut on the lathe.
Method (b)
an'
applic<1tiGll of the auto-collimator. The
axial oscillations of the kad screw are convened to angular movet.Qcnts oC a plnne
reflector ball located inlo the centr';) hole in the end dthe lead screw, and mounted
on a cross-strip hinge
6.8). The angular movement of the rei1e:ctor during
one revolution. of the
screw is observed in the auto-coHima-tor eyepiece.
\X/hen
an optical method for such a test, and bearing in mind that the
axis of the incident beam from the auto-collimator forms thc measuring datum,
it is clear that this must be turned accurately through 90. It is dOlle by means of an
square. This is a sp~cial case of a prism, such that, rcgardless of the angle
at which the incident heaiH strikes the face of the prism, then by internal reflection
the beam is turned. tllrough 90".
Assume, then, that the transverse ways and the column face are perfectly
I-MCHINE
END FRl\ME
THRUST
L.
Fig. 6.7.
to
test the
Fig. 6.8.
j:.ead screw with angled tlmlst faGes and possible method of tcst using an
auto-collimator.
125
.sA
i
1,S
TT
C 1'1
'7
I
PRACTICAL TESTS
~rpv:ouslv
nntpd
these
are of a tVDe
desirrned
to reveal ~ the combined
eiIects
l.Hv 1
_. v
~
hi
d.
of s~veral
. , ..
As
from the Industrial Rcvollltion to the
the ncr:d for closer control oyer the accEracy of.
used for transmitting ,he
power made available has also
the iilOSt used me;:m3 of
pc\ver and InuJtiplying torqne is
the mediulil of g.ear trnLns.
It is obviOl:S tkt the ,strength of gear teeth has had to
to me;';[
loads, but this is Q
problem which is not a
COilC',;l'l1 of tli.i,) book.
it "is,)
of a ge2.r train that it shaH hCl:vc a ccr,:1r:)nt v'~i'>'
city
rDtio~
VgriL:.ti.Jlls in velocity
r~ltio
l~Yld",
ing which
rise to
fatirrue, leading to tooth
.
The nOJse
of jnt'~resr if one considers t!Ji~
mobile.
futomobiles ha.d rudir;:entary exh:lust silencers
engine noise caused most of the other mechanical noises to b~ G\/:3:r1ook:;,:rl .
EiTiciellt exhaust
Inade m(:chanical nois:~~; 1"r0111_ th~
E~Ol(~
apparent. This was silenced
t~le use of helical gC1rs and C1CSCf cent.ro! in tiocir
m8..n 7Jfacture. T'he Gear noise \vas redlJCed nod carburcttor il!takc :'1oj~e bc~aTr:e
when reduced by efficient air cleaners and intc,ke
enabled rear axle \vhine' to rnakc its presence fell. 'The use of
'bcvc!~~ and
hypoid gears, '~2D.in v/ith closer
reduced this aDd the
valve tiIning gear;;
required :lttcntion.
th~~s
G~(hD.ust and iU::I.kr;
silencers were improved and the vvhole
started again.
Thus a
item of development in the motor vehicle has been the
ment of efficient gears, an.d tLl:; only considers or.e
If one
this work applied to all of the l1lcchanlsm~ which rely on
nut pO\lier; the in1portance of th(~
apparent.
ft.... fe\v of the dHTcrent types of gea.rs required by modern industry ha"/e l;e~:;n
mentioned above. Witbjn the confines of this work it is proposed to deal
with
tooth (spur) and helical
These cOl1~titut;:; a
invc1ute gears of
127
126
:'
I )
\a
r]
' rat'm 01" a paIr
. o.:nvo
f'
1ute gears.is constant, regardless
~ le vel
oeay
of
A? involute is the locus of a point 011 a straight iine which rolls around a circl~
wl~nol1t, .
An alternative definition is: the locus of a point on a piece of
st.Eng wilic:h umvound from a stationary cylinder.
/0,.(:"Tlle curve is therefore as shown
/),,:)9:../:';
in Fig. 7.1.
~'-<)
! It';! <..i
From the
it is seen that the
length of the generawr is equal to the
"'"
arc length ()f the base circle from
",,the
of tangency to the origin of
""
the involute at A.
7.1:
arc AB=BC
'"",
'~0..,
/ /
"
128
But from
angle is the
hetwe(;!/Z the
of lhc
involute geo!l1(;try
br;;w deo.It with separately as it 1:; of
in the work to follow.
7,5
A single toctl.l of 8. gear is made up of portions of a pair of opposed involutes.
The tCC;:11
a pair of geors in mesh contact each
along a line
which is the: ,:omnWD tangc:nt to their base circles as shown in Fig, 7.3.
this
is the common generator to both involutes, the
or pressure betw;;C;l the
gears is transmitted along this jine. The
between !:11c line of actio;, a.ad lhl:)
common tangent to the pitch circles is
known as the pressure angle, '/J.
129
jVfetr%gy for
Engineers
Gear l11easuremelU
,1.=
OC=COS'I'
Ro
Rp
the
Circular l)itch CP is the arc distance measured around the pitch ciIcle from
to a similar flank in the next tooth.
..CP
Ba,~'e pitch Pi; is the arc distance measured around the base circle from the
origin of the involute on one tooth to the
of a sin1ila!.~ ilr'iolute on the next
tooth.
Pb
CPcos.p=1TMcos.p
Addendum is the radial distance from the pitch circle to the tip of the tooth.
The nominal value is:
Module
Addendum
0157
or
0250
or
0,400
=0,157 M
01
0-2501v[ or 04(;0 Ai
Dimnerra/ pitch P is the number of teeth per inch of pitch circle diameter.
This is a hypothetical value wbjch cannot be measured but it is most important
as it defir;.es the proportions of all gear teeth.
'
De(~endum
D
N
'b
130
"_.
,mere 11'1'1S rnade a w>h01e number of millimetres.
=Addendum +Ckarance
0-250
or
}.{=--
Dedencllli11 is the radial distance from the pitch circle to the b,:Hom of the
tooth space.
hS'[;cr
B, C, and D gears,
or=p
1250
1-250 Iv[
0400
blanl~
131
+2 111
but D=N1Y
(V L
,', .Blank diameter=iVlVI +2M=(N +2) x .Module or' "
w
if)
0::
*'I~
z
Tooth thickness is the arc distance measured along the pitch circle frou, its
with one flank to its intercept with the DIher l1ank of the sarne tooth.
<!;
0::
17
np
=~
k.U.
LEAD A'lGlE
O!17X
<!;
-'
CL
lIrI'_~A~~
i
~".
HEL,X OR
" ._______..
~ SPiRAL ANGL E
~.
0
:)1'-------.. I
~
\
NORMAL
PiTCH
In the VlO1'k on gear me~i,3UremeD.t which follows th'~ expressions derived will,
"vhere possible, aU be reduc;~d to functions of these dimensions.
AXIAL PITCH
HEJJIC/"L GEARS
A helical gear has involute teeth which are not cut parallel with the axis of rotation, as OIl [,;~;ur gears,. but: at an
known as the helix, or spiral, angle to it.
Helical gears arc nomdly used to transmit power between parallel shafts.
They pcovide a much smoother :md quieter ,:.etlan than SDlll' gears, owing to the
fact that at ;:my instant a l1iW}ber of teeth are engaged. FUrTher, each tooth is
i;lljl,dY,GU over 8. short lcngtll of its jIanlc at a
a.nd the engag(!Ulent is
up and ~c1eased gradually.
Thil:3 llny measurernent~; made can be in one of three pianes:
t.
a.
AB
.. AB=AC cos
CPn
Helix angle (j
Transverse pressure angle
Transverse circular pitch
Hl1d
denoted
CPt cos
(j
(1
(j
(j
133
Gear jyJeasurement
Afetlology for Engineers
Now the angle of th;,: flank face of the rack in any plane is the pressure ane,]e
of the system in that
0"
Considering
7.5:
tan
For any section DF is constant and ;;qual to the tooth depth but
DE=D1El cos (l
(f
ADDE;'JDUM~
If in the design of a gear train a pinion is Imtde too small, interference will occur
between the mating pair. The tooth of the gear or rack will tcnd to 'dig in' to the
root of the pinion, causing rough running or fracture. If a
is Cllt by a
generating process in which interference would occur, the teeth become undercut
and weakened. The condition occurs when the line of action extends
the
point of tangency to the base circle.
Consider a pinion being cut by a rack as shown in
7.6, in which the
of the rack cutter tooth extends beyond the point of tangency of the line of action
by an amount S.
f.
RACK TOOTH
SECTION TI
IN TRANSVERSE
PLANE ~ ~-.-(. .-
YEUX
,,,NGLE
d
,I
Fig. 7.6.
SECTION NN
iN NGr,~fAL F'U,NE:
F:g. 7.5,
NOl'tl1al
and
transv(~.(:e
Thjs condition would give rise to undercutting and can be overconl'~ by dis
placing the rack outwards 2. distance S. Thus, to
for
tbe whed
addendum mnst be reduced by an amount Kw and the 'pinion addendum increased
by all amount
to compensate.
The values Kp and Kw arc called the addendum coefficients for the pinion cUld
wheel and are directly related to by S=KM, where jJ;f =Modulc.
B.S. 436: 1940 gives two conditions for the calculation of Kl ) and Kw.
135
134
Gear ldea.I'uT::'lIU.:nt
$ .
or
checked by optical projection and this leaves only inaccuracy of setting to cause
the form errors.
In the manufacture of gears of higher class, e.g. precision-ground gears and
master gears, it is necessary to determine the accuracy of individual elements.
Apart from tooth thickness, these include (a) pitch of teeth, and (0) form of teeth.
and
(b) If (N +n)
S(;(;3
and
tn th\; above e:cprcssions IV and n ,tre numbers of tectll1n wheel and pinion
respcctivdy and a is the helix
for nelical gears, thus for
tooth spur
gears a~; 0 ilnd sec3 u = L
GEAR TESTS
7.9 R
3
fOf
Pinion addendum
=Mn+KpMn
VVheel addendum
+ Kw)=iYfn+KwM"
A common form of gear testing machine for performing these tests is the Parkson
Gear
shown in
7.7. It consists essentially of a base on which is mounted
a 'fued'
whose position can be adjusted to enable a wide range of gear
diameters to be accommodated, but which is locked in use, and a moving carriage
which is spring loaded towards the Ji'\ed carriage. On the fixed carriage is an
arbor made to suit the bore of the master gear, or a similar spindle to suit the
gear under test being mounted in a parallel plane on the moving carriage.
1':Qte that in each case the n.orni:lal addendum lii ie, cbanged by 'In amount IGYJ
w.bJcb. is the amO]!T(t the rack is
to avoid undercutting.
The methods of
,wd
gears depend largely
tb". l11';t.(}.od of mailllfacture, and tb~ equipment available.
rfJ:~nt can consist of:
811
MAo,TER GEAR
General tests.
or
Note.
diameters of spindles.
(b) Set limit mark~ on dial gauge.
(c) Mount the master gear and the gear to be tested, and note the variation
in the dial gauge reading when the gears are TOt:ued
hand. If it falls
outside the limit marks, the gear is not acceptable.
137
Gear Measurement
Tllis is the standard form of test used industriuIly llnd,~r production Con_
the gear tester can be used to carry out mDn~ comple;{ tests.
Fer
lq
the moving slide at th'3
centr~
of the,
gears, "wtI by fixing dle master gear, the backlash can br~ determmed by setting a
at the
line of the production gear.AIw, at this setting, the gears
,;.hecked for :)1TIooth
pr'Jduce a trace, or
It is usual for such errors to occur together and a compound trace of cccent~
rieity and tooth form errors occurring on one gear would appear as shown in
7.8 (d).
It should be noted that when testing helical gears on the ,Par1cson tester the
vertical component of the normal force between the teetn has a tendency to cause
one gear to lift, or ride up its arbor, For this reason, the master gear in these
cases should be mounted between centres and rotated in such a direction that it
absorbs the upward thrust, the equal and opposite downward component holding
the tested gear down.
If no master gear is available, the instrument can still be used to check a pair
of mating gears by running them together. It must be ensured that
do not
have compensating errors. For example, equal eccentricities, if mounted in a
particular al1gular relationship, would cancel out and show no r::rror. If this
method is used, the gears should be tested twice at relative augubr positions of
180 to each other.
A frequently used instrument for measuring gear tooth thickness is the gear tooth
vernier. It should be realized that, as the thjckness varies from the tip to the base
circle of the tooth, any instmment for measuring on a single tooth must
f'lSqM;.\U.'{ u:-~ru~
C()"i,,'C, 'NHEM
(c)
'M'
i\,DJUSTMEN
\j
(d)
Fig. 7.8.
:::.~ROR
test
il!st.pJrrl~i1L
~-w-j
Fig, 7.9. Jaws of gear-tooth vernier.
139
Gear )1;1eaSUrCI1Ul;Z!
In the above expressions it millt be remembered that hand fV arc ideal
values. Allowance must be made for h if the addendum is modified as in 7.10 and
for W to accommodate the baekiash required.
In the case of helical gears the tooth thickness in the transverse plane at the
pitch line is not eas:,! to measure but the normal tooth thickness can be detennined.
Within the limits of accuracy of this instrument, the normal tooth thickness W"
is the same as the tooth thickness of a corresponding virtual spur gear in which
the number of teeth N v = Nse(jl(}' where (}' is the helix angle.
W=2AB
D
AO
=2
NM
In the expressions for tooth thickness at the pitch line .it is seen that the dimensions
hand Ware both dependent on the number of teeth. If a large number of gears
for a set, each having different values of N, are to be tested, the separate c:rlculations would become laborious.
Consider an involute tooth symmetrically in close mesh with a basic rack
from as in Fig. 7.11. Regoxdiess of the number of teeth, for a given size of tooth,
i.e. value of J11, contact would always occur at A and F. AF is known as the constant chord.
sin
AB
A.B=AO sin 8
/90)
NM ,.'
SHl\N
~jg.
C;J.oIGl11
But W=2 AB
OE
OB
BASE
CIRCLE
+ Addendum NII;~t-M
anciOE
190)
N!k[
co:, ( -
OB=OA cos
.... h
'" (I)
PITCH LINE
OF RA":"CK-'-"'-~=
Fig.7.11.
vV ,
+-.a.X- Nfli'
,./., GOS (GO)
~__
i\!iV!
.!t ...
In Fig. 7.11:
\N
BD
II
NMf
2
I 1+-
190\ 1
- cos [ - j
G~)
\N I
'" (2)
In tliangle ABD
AJ3
. ill
. I
em;; , ar PltC..l
'lTl,,:l
=4
= cos ,p
7T1V!
..AB=BD cos 1f=- cos if!
4
140
141
Gear I'lleasuren1em
Metrolcgy for Engineers
\"
".. IC h.DC
\. U ~i,,'
... =
[lang
""".
In
AB
cos 'f
cos if;
.. AC
and W=2 AC
AI
'so,
EG
.
=sm
if!
7.H13
'f
., BC=AB sin
The abov0
IS
norn1a!
are for
lOath ~pur gears. For helical ;:rears th""
r.>
'v
ootamec1 oy sub:;tituting values of normal module and
.;
,1
.'
' .
pIe-::sufi-~
Fig. 7.12,
based on
tan
1, "0,,2
'.- ,) ,I,
r
and hn={l
r'v. If reduction
. _An interesting point about the coustant chord method is that it readily lends
ItSCit to
'H~rnier~
'/"
a fo;:m of
which is 1110[e sensitive than the gear tooth
Such D.n ins:xtnnent is rnanuJacture.d b'tj ' 1;'<1. E.
-~ Id " an'd'IS
L(
L_1Ldrated in
7.12.
Tt \,.'
h(']J':""
"'"
"
,.
1.1(~L~, (
,JI
pmr
or
having
or 20,
()fi tilG
'NidI a dial l1 aurre
down to the tip
',+
"'h~'0 toct~l
U
.. " J .I..,
,.v . :.."
,....,~h
~
.0
~.-.,~) iiu.v\ViL 1 u.e lTIstnl1nent 13 Sf~t to a
- -
- (..,
.,
<."
'J~
llT"
1"0"1"
O'tl"'4:::;".
.1., Ll "" {lh'!
4-........... b(
Fig. 7.13.
It can be seen [hat this uses a single vernier callip;:;r which overcomes the
disadvantages 1,1en[ioned above,
those inherent in the use of such an
instrument, and this can be overcome by tl1.C use of more accurate equipment
than a vernier calliper.
143
Gear Measurement
.Metrology fo/'
r"onc;d~~'
',~ edge
.',
-C
,#
i'"
'Jl a St
Illig!.u
CC-U)
"
A r ! ,ceing rolled back and -'orth
i-1C,
along a bas(~ circle as in Fig. 10.':4. ftc; ends will sweep out opposed involutes
~'\l.j\., and. CZ.C.C 1.
lvteasurcments nUl-de across these opposed
mvo.lmcs oy spall
will ~)e constant,
.
i .'~..... T'r V =.t:~1"_",
,1.1' =,LXl
A C
A
{'
,
n
1= . 2~2 =arc ;:.t~(i:bo
i.e. t~t': am,
01
To determine arc AB, th~ tooth thickness at the base circle, consider
which shmvs a single tooth and the relevant data.
7.15,
arc AB =2 arc AD
=2 (arc AC+arc CD)
AC
I
l'
arc Rb =Inv 'fJ rae.lans
.. arc AC=RI; (tan if; if;)
NM
arc AC
cos if; (tan if; - if;)
(} radians
arc EF
Rp
." (2)
arcCD
RIJ
=i circular pitch
1
arc EF
TflU
Fig. 7.14.
common
Do
.'.
ge~cmtor.
()
1T
1 mV! 2
x--=--xRp
4 NM
radians
:. arc CD
'-f'~":'t~01~or~
"1':11. " . . f .....
I...u"-'" v
.\,; \, iL puvldOu.
t_
\).t'" tu.c
."
taces
is
(1:) lon.g as they
are
and on an opposed
of true involutes.
this to gear teeth
f'<lJ't
;)rl",r-j'er"b
i'~""v +. . . . ror.. ,"'I"lJS~
. ,
1 IneaSllfC111cnt
.....
..'
it is
(."'''' ~....,
u
a nunl ~t_0
;Cf 0 [ teet1~: sue I1. t I1"rc{ tne
IS
m the
eirel.;; of the gear, for it is at tbis point that the
to GonJorm to a tnlc .involute.
'
f\pplying Ihis
to t;~le .,;onditions in
7.13, it is seen that
IT
.,. (3)
<bx-
2N
arceD
{.U I"'U'J
F;"-=arc AD-i,arc BC
in whicll arc .~\..B is tb.; tOOt~l thicrGless '2.t the base drck and arc BC is (.'1 x the
base
S '-.,.~: r.t~.. J'-: le -'11
:--.' ,.:;,. O..t t,ODt h .5j~aces over V/tItC
.. h (TleaSUrelnent JS
. ' rnade.
'- r..;~L.l6 IL
L.. ~ruo.....,r
.j....
211
by
d'
rQ wna
and base
But arc AB =2 (arc AC+arc CD) so substituting from expressions (2) and (3)
-:f
arc AB 2 [ NM cos
=
Ni}!
D x-;:,;- cos
... (4)
~.
2:'1'
(Tf.) ]
2N
Be
.1.
'I'
Jr
arc
NM.COS
.p (tau,f;,til)
.. , (1)
2N
0- 0/ +
7T
L N
r.S]
+N_
145
Engi."tZeers
Gear .iVfeasurement
Measurements by tIlls method can be improved by the use of either a micrometer with flanged anvils, or the David Brown Base Tangent Comparator shown
in Fig. 7.16 which consists essentially of a micrometer with Hmited movement on
either side of a zero setting, the zero setting being made with gauge blocks i)f
distance pieces. Further, the method lends itself to measurement of small tooth
gears by projection, either by direct measurement or by t1'1l.'/crsing the table of
a Toolmaker's Microscope.
I
j
DIMENS!ON IN SET TO
GAUGE BLOCKS
LIMITED MOVEMENT
MICROMETER
MOVIN
BASE CIRCLE
ANVIL
Again
tI10
;} b =
bacldash allowance
.pn
,.L
in which
C~ [vfoente
(,b.,,!
'II'
-'-
_'FrS]
. 1'!
cos o
in the norrrl:ll
tan
W=Ni\J!nCoSif1nrtan 4''/Z
L
146
7r
'TrS]
d'n-'---+-, '2N N
147
Gear Afeasurement
7.n
. ular SIze
. d ro 11er Whose ra~1US,
J'
'lTM' cos '1,.
' . case reqmres
. a partie
---,
-', IS 1ess than t he
T illS
4
addendum. It will therefore be shrouded by the
of the teeth and measurement
OVER
d'er
~( "
~onSl
oetvl~eil
;).
1PiQCCG.
'
'J'III tIle
1
a l'DJ' e~
spB.ce
of
ge.:;~r
may be difficult. If this is so, readers are directed to the work of Earl Buckingham,
who suggests a method using any convenient size of roller.
For gears with an even number of
a direct diameter measurement may be
made. If the gear has an odd number of teeth, a radial measurement with the
gear between centres can be carried out, using a comparator with the gear, or a
measur:ement over a pair of
provided allowance is made for the angular
relationship ofihe rollers relative to each other.
.'
cos Vi =
01\
-----
on
The measurement of the pitch of gear teeth may be made by (a) measuring the
distance from a point on one tooth to a suitable point on the next tooth; (b)
measuring the position of a suitable point on a tooth after the gear has becn
indexed through a suitable angle.
+r
NM'ffM "
+.4 C"I'~
v"
~fhusJ
Ivfe~lSUrCl)1e.Dt
over a
"
,Ir
r "'i'I"p
." --.".- >
=radll1s
of rollers in
tcoth
S]:J:lces,
One method of carrying out this measurement is by a portable hand-held instrument which measures the base pitc1I.
In Fig. 7.18, the base pitch i:;
between the involutes with origins at
and B.
It can be shown that
r l110d
_;,
and D t! = l'l,M +
7t11~f
11'
cos
,p,
to R g
!,lg
149
ikfetrology for
shov/u is
i)f
at
th~
',",''''_!,'''_,C;
It would appear that the simplest method of deterillitIing pitch errors is to set a
dial gauge aga.inst a tooth 2nd note the reading. If the gear is now indexed through
the angular pitch and the reading differs from the original reading, the difference
between these: is the cumuiative pitch error. The problem is to index through the
exact angular pitdl, as an error in indexing of 80 radians induces an error of
Rp o8 in pitch, 'yhere Rp is the
circle radius.
However, if the angle through which the gear is indexed is always the same,
and not necessarily the angular pitch, the induced error can be corrected. Such an
indexing device has been developed at the National Physical Labof2.tory and is
InstnJl11ent
incorporated in an elega.'1t pitch measuring device made by the
Company.
The indexing device is basically as shown in Fig. 7.21.
point
i'f!.ngency..
adjusted so
l)etw;~enA. and C is
to the base pitch,
J;cate the
I,
If
is comtant, th:; reversal
on ,:nch tooth should be the
St1In(~ yvhen the
is rocked.
Anolher
;]S0 tV(O dial gauges on
[c{~Lh
\v.~th
g',::8.r JI1ou.n.tcd .rIl ce:;1l:reS, as
7~20~ rThe gear 1s rndo,xed
to
a
A. Any
i),(1
Gear }.Jeasuremer.i
dial B
Fig. 7.19.
(Courtej}'
0/ the
Tl,le gear is mounted on a mandrel attached to the disc and rotates with it.
Rotating concentrically with tI:e disc is a SillC ann. Clamp A. leeks the disc to the
base and clamp B locks the ann to the disc. To index th~ gear, and starting with
clamp /\ locked, and hence the disc: and gear locked together, the procedure is:
G.L,UGE
A
I
Flg. 7.7.0.
150
to dctern1ine
(b) Hold the stylus down on the ga.uge blocks and lock
B.
151
~l"".
,-,
Tooth
(O'Oltnm
1
2
3
4
Correction
Induced Elror
Cumulative
(0'01 mm units)
(001 mm 1mits)
.i'f?etJding
19
+ 25,
+ 60
38
+ 80
57
+110
76
6
7
8
9
10
+ t3'O'
--'''9-5
11
12
+175
13
14
15
16
17
iO
.0
19
20
Error
l~
-- 11-4
-133
+2~6
- 152
-02
+ 16-5
171
190
209
-0-6
-15
-24
22~8
~':24'7
J.3
-2,7
+21
+22
+23
+26
+2B
+31
266
-3,6
-2,)
--24
+36
+375
304
323
342
- 361
+38
-,331)
radians~
35
30
20
-H
+-18
+ 14
0
~ooth numbers
15~
I
10
CUMULATIVE
Fig. 7.22.
Graphs of readings and cumulative Ditch error. Note [bt tooth-lo-tooth error
can be obtained by subtracting adjacent readirlgs.
wm
152
O'Olmm
+12
~285
.C
40
0
+06
+22
+23
+34
+35
+JA'O
+ 145
+ 15,:')
-I- 185
Gear lv!easurement
153
Gear
Do
cos,~
=.rTi',/! cos !P
~fhe I)a.vid Brav/n IG.volute Form iT\;~;ter ~)hO\Nn in
7.241s of this
the styl!].s,
the In. iolutc ferTn, being accurately locat-ed above the
1
l"WeaSUrement
straight edge. The stylus is mounted 011 a small vee/flat ball slide constrained to
move paraJlel to the straight edge. Thus slide movements measured by the dial
gauge are in fact errors in involute form measured along the line of the straight
edge, i.c. normal to the involute form. Motion without slipping is caused by a
lateral movement of the rod at the back of the instrument.
all
a tokrancc factor S,
eN + 60)11-'1 ]
1'5= [ -~~~--+1
10
Fig. 7.23.
The only other symbol used which has not been dcalt with is concerned with
accumulated pitch errors, and is: La =leng:th of arc of the pitch circIe on which the
accumulative error exists.
-------+--...
n,.,
..
L a = N X /-77 radIans x Rp
HENe:::
,\IOI~HAl
TO
')
x _77 X
=nJVlu-
DiSC
NM
B.S. 436
tables of allowable errors in tooth thickness, pitch und pfoIi1e
in terms of and La. These tables, in association with the calcuhted values of
S and La, enable the actual tolerances to be calculated.
154
155
,
8.1
Il'n'i~O.Dlr:CT10N
EXAMINATION of B.S.
dealing with the tolerances on commercial screw threads,
shows that these tokrances are relatively large. Such thrca.d" are normally
using Limic gang':os as was described in
7. Howev,~r, certain threads
must be hold to much closer tolerances, and this
particularly true of the limit
gauges llsed for screw thread ins'Jection. These threads must be measured, not
to ensure that they arc of a degree of accuracy to
successfully
threads from the bad when used as tools of inspection.
fvleasuren1cnt, ?.s distr.nct frc,ffi
of ,t screw thread can he extremely
There :lre a number of elemems to be measured
as will be shown,
some Hf(~ illt~;:rrelated. A vee-form thread is composed basicalIy of the following
m-
elen"lcnts.
(a)
or outside diameter.
,'/
.."
Pitch.
n~[Lj(
angle calJ.scd ~!
157
2ft Inade at the same pressure) and this indIcator ~!.as ;),
incr~~a"sGs
not
.instr~Ui.lent t:at
for -it
be direct
si:andard. 'rhus YOI'
fnstead it rI1ust
:.TIQ.kes
used
and set to a
at the
on tile
consi~~-t8
m:d
th(~t
obtained.
diiICt-;:;J1ce in n.1.t;se
and t1H~ thfe~ld at that
tLucad,.
OJ,'
j,1te
:~he
dirf:~rencc
is ar,Jded to
th~
tVi"(J
l1J~
U"0'.Ut;'U.
diameter of the
::::::calibratcd dialnet~r of
If
cylinder
GyJinder
~,;: ll1icrornetcr
read.tng
Fig. 8.3.
thread
then
the
rnay {)e
(.1~D.D,,:d
ClS the
It l:::
of :tn
pIOC:~;S;; sjn~ilar
~;:~;.;8.surGd
h~t{d.c:1;=J :~,~'ld
R~'e
uSed to
g.3~
st::'O\Vll .:n
It
duced
to
:~hreud.
ali
but
dh~.rne~Gr
~hread
Nm is pro-
C~t:i
pi~ch. Thi~~
vvould (;-;n;] to
rO:'Gt;~
dia1Y~cter
rnD.Grdn~
of
bcn;~'h
LnierCIneter 1:;
cGnD1r~til1(;d
b~~tv's'.:!en
Cl;fltrcs Hnd a
to the axis of tile:!
:0 l11.C'Ve Dt
jI.J
and
.on
as .i.n
158
and:
159
and tr,j_per.
-.~:.,: -- -
"-v
but
screeIl, or
.;
"ml
p, " : '
prohl ')1,'1
'h most. ef"Iec t'we projector
.
this class of work is
l . [.0
fOf
' as It
. was ('~veloned
l
prolcclor
at the" hr"t:o"nl
Dl.IYSlca
. I
-'
r
... ,a.!. .d,~ L
)C
knO\llTI
es
"1" lSLJ.D .~.
I
"'c'..
T .,
1,,1
...
LaooratofY
. I!: co~:sits. of ~ lamp-leo use whose optical outlet contains conLiel1:3er
to be projZ':cted i;; mounted on a sta;;:e ''1i'iC.,,,.~_
glVi.; even. nlur,uoatlOl1. The;
"{he condensers and the
lens which throws an enlarged
~ on the
j
screen as in
3.5.
'"
(a)
PLANE OF
SCREEN .AND LEl'JS
PROJECTED
(b)
LENGTH
--li""o:..c.~~-,,-! PLANE OF
SCREEN AND LENS
By swinging the thread through its helix angle the thread helix is parallel with
the light rays [Fig, 8.6 (b)}. However, the Hank angle is deiind a~l i}ell1g measured
on a plane section pa:, .Hel to the thread axis. This being so it '::1.l) br~ seen th:lt ];lJe
set-up in Fig. 8.6 {b) 11::.5 3. foreshortening effect and will induce i);Jrrowi l1U of the
thread
as projected on the screen.
This problem can be overcome by swinging the lam[J-llOU':t; illlOP!.", till helix
{60
objectlv[~
\_~--,
1,
ANGULAR OCULAR
IN FOC,lI.L PLM'IE
OF OBJECTIVE
LENS
MICROSCOPE AXIS
TILTED THROUGH
HELIX ANGLE
OF lMAGE
Thread flank
Elli.y be mcastm~d
arnieroscope 'Nith a
'1':1113 consists of a clear
screen in t 1;lC focal ptane of the:
datunl lines 'vh.ich can be rotated
360, the
JUr;il!mred dlrcc[ to I" and
estimation to fraction" of a mil1ure.
OF TRUE
PITCH
PU\NE OF seRE':"i
lUI::! L;:'IS PARALLEL
TO THllEr\D AXIS
--.--.--""'~-_. ___ -.-----0."....---
-~----~--
Fig. 8.8.
~c.:;:is
and screen
rays t~1.rau,'3[l
The datum lines in the microscope head are set to zero and the. table
rotated until the crests of the thread
image coincide with the horizontal
datum. The table is t!len locked :md
the datum lines in the microscope
eyepiece rotated until they coincide
with the threadfianks. The flank
angles are then read off the eyepiece
scale.
TrLis equipment is normally
among the attachments which may
be set up on a toolmaker's micro. can
scope. The readings and
either be made through the [mcroscope
or viewed 011 a screen
as shO'vn
8.9.
It should be
that the optical axis is not nOnYJal to the thrcfi.d
axis and some
of the:
still OCGurs. However, tius tcchniqui;
is cuite easy to set up and. produccos
res~lts of a reasonably lugh order of
accuracy.
w!richever method is used. errors
on both right- and
:hnk2,
should be determined as each can
cause interference with the mating
tnteIf,~rc!1ce
is
Distortion is
i.G a n1lnuLurn.
nlouat~d Otl
~:ltO\V.:l
in
The
,Jl~~12r~l1n
Cf1Jti.':.::S
::tre l1J.ountcd on
by
{:CrC-f(E.:lr:,,~c: dimcn~~ions
cn:;.~ble th(~rn
~I;ii;rGn"l(;ters
ll) be nlov~:d
to 0002 nlra ;lnd on a
i~D!.~-tCy tat)i~.
on
:~n.d
the
Ln !J~ct\.;ren('{;.
$1,"':(GD.g
bar
thxough
Fig. 8.9.
163
AB
But
3A3
oEI/,=--J!--(88 1 + 88;)
sm 2{)
where W and 08 2 are the worst flank angle errors on the r.h. and l.h. flanks in
J7),
[
mULI~G
HlUWiG
GY
I\
t;JT\
t~UT
I'
t:! \
I~/- \~---f-/_\\---~--\
~,
ZONt:.
II;'~ ,)
FLAN:( AHGLE
ERROf~
'-
~--rl
'/
08
1/
CORrlEO
1-
radians.
Applying this expression to a Whitworth thread form, the value for h may
be determined in terms of pitch, from Fig. 8.11; 8 is known, and 88 1 and 08 2 may
be converted from radians to degrees.
I
'-l
'--~
SCREW
DEPTH
06403 p
Fi~~~ 8.10(0),
A[~ enlmgen-icnt
dd,.~r~"1m(;d.
or ...
"I'
, '.
PC',
s:.o.U =
aD
~:;;:---8"
>id
...
Fig,8.11.
(J
165
164
.f:
,...
Jor
The I'lleaSurement of
~nglneers
IS
= --:-----(
08 + I) {J.)
Sill 28
1
0
Threads
train between work and tool1ead screw as when producing a metric tluead with
an inch pitch lead screw when no translatory gear is available. JYJore commonly,
it is caused by pitch errors in the lead screw of the lathe or other generating
machine.
If the pitch error per thread is op then at any position aiong the tnread the
cumulative pitch error is nop where n is the number of threads considered. A graph
of cumulative pitch error against length of thread is therefore a straight liue
[:Fig. 8.12 (a)].
0::
cr
a::
'"'
'-J,J
l-
e;:
>
in
:r:
u
a~,;
the) flank
errors
Ine:J.3ured.
~I
MAXiMLM
Bf
3 .. j
(a)
by a single
tool its
depends on:
TF'
--~
ot 0~~L,:jn\...,--,
'''1,',-/.'", ..1 (.-,~ ..
.>-leh
,?.fruch ~_. ::' the abov,~)
~l~1.C-.C;',~
(;('.i'~J.w~r10.'11~ "' ... ,.. . .
"
~'*' ~ ,,' _
" : : "tI,:,; 11.
cLeU::rmlned
of
'.;(1-1),
.1\
'::11'Of3 Vl1.t
condi~ionf)
oecurJ
or too
,"
'dH~
type a f
b\:; too
of threud
of tbIcad is
ti1,.=;
tOL~l!
'_'L.''''-Llii.'~ fruHl
constant.
e:rrJf
CUMUlA,IVE
v -,....,.....,.-.-.-+..,..-'1-'9"'
CRR'~><
,I h -'-'---i
i I I I P'TCH
'
'-,
,.,I"
1O,li21314
THREAD r~UM8ER t.
This type of error occurs when the tool-work velocity ratio is not constant.
It may be caused by pitch errors in the gears connecting the work and lead screw
or by an axial movement of tl:e lead screw due to worn thrust faces, Such a
movement would be superimposed on the normal tool motion to be reproduced
on the work. It will be appreciated. that errors clue to th'~sc CCl.U3CS will he eydie,
i.e. the pitch will increase to a ma.ximum, reduce through normal to <l minimum
and so on.
,A graph of cumulative piteh error will tl1U~ be of appro~:jrn;lldy sinusoidal
form as in Fig. 8.12 (0), and the maximum cumulative pitch eno "vill be the lotal
error bet,Y;;:en the greatest positive and negative peaks within the length of thread
engaged.
E'~"ror
166
167
NUT OF
PERFE~T FORM
,_,
S'-Rf:"N ""TH
PITCH ERROR
AND PITCh
}12
(-\ /
,..',
,--.
:1~~f\:i~
(
c,. -
"I
tEd, OF SCREW
\. Ed OF PERFE::T
w'TH "'Ir--H ERROR
NUT
"
c,
I \
np +
;n(Fc~
H
"t~r
"
llOe(l
".
Fig. 8.14(0).
.'. ,
.
011 l)lcC,l-measunng
machine.
8.14 (b).
op
It is seen that
tan 0
2:
Sp
= OEd
where
.. oEa = op cotan e
op is the cumulative pitch error over the length of engagement and OEd is
169
168
M
EfFECTIVE
GIAMETEROF
PERFECT NUT
j
Fig. 8. 14(b).
.'
1
'th the cvlinders enrraged in the thread. The
cylinder. A readll1g IS. then. ta ;en ~1 .
,- > ,
EFFECTIVE
DIAtAETER
fd
, 8.J5.
Fig,
Ea=T+2x
BC
In L1 ABC tan 8= All
Enlarged view at A.
The pitsh error is therefore almost doubled when the equivalent increase in
dlcctive diameter is calcnlatcd. A sere'x plug gauge having a cU;-;:'.Jative pitch
error of 0006 mm wiii tbus reject aD work within 0,0[2 mm (approximRtely) of
tile low limit in tlle case of "iVhltworth
and within 001 rnm in the case
of ~Ietric threads.
,', AB
In L1 ADE sin
." (1)
DE
.. AE=DE cosec 8
efTr;ctive
diameu~r
to be used in
170
But
.. AE
Now
cosec
x=AB AF
and AF=AE -EF
but EF also
.. A 'R=' 2 cosec
AF
e--2d
d.(cosec,R 1)
.. , (2)
171
ANGLE
cot
p -=llominal pitch
d =wire dilctmeter
8= nominal flank
semi-angle of thread
I
I.,.
5.61
I"iormaity v. hen
for :J.
rather dum
to rtank
Fig, 8,]6.
(Hi';
th,'~ad.
US(~
errors.
In Fig lU 6 AS is
AB
pitch.
Be
Be
cos
d
2
INSERT
MEASURE
COMPARE
172
Aietmlogy for
OF SCREW RING
!l.3
lL63
Cm:T:!ctlons IiO!:
The expression for simple effective ~iameter derived ill sect.ion ~.6 is based on the
assnn,lpllon that the measurement IS made over a
sectIOn of the thrc' d
! 1
t
I aXlS,
. l.e.
. measurement IS
, over a senes
" of annular groov<l
parade
to tJle
lnre:lc
~.
I"
rr"(US 01"" course IS
. not so an d t,1e
1
"1"
es
01 tilrcal lorm.
Cllcct ot measunng over a helix
is to throw the measuring cylinders out slightly and increase th,) measured value
by
amount C. At the same time measuring pressure aPl'Eed to point contacts
causes elastic defon-nation of the ',vires and the thread flanks and reduces the
measured '1alue by an amount e. The correct value is obtained by applying Corrections of C and -!-e.
Q
and dimensions.
a normal screw pitch measuring machine
The nitch may be measured on
f' t1 s'tvlus ~'ltherjike a boring bar.
'.
1
t t
'1 W "ltlY 0 Ie
' , 1<
. '
~
using a speCIal attac Imen ? a. 0 t;,c d 1 1aking u, plaster cast of the thread,
'f'nean
fI k "nole
mav
on'"
be
mea.~ure
)y
n
~
d
!'
e
d
t not
u. '"
'
'J
c'
d'
f tl-e thread aw ljte ou,
1.
f'l
so tIl"n hah Ihe 1::tmvter 0
>
d - h
whIch must De 01 e.' d
d T"
ay then be proJecte m t e
serewed out or the C'lst may De dlsLOrte. iUS m
0,
_ "
normal way.
.
, l ' d sine: b-tll-cnded stylii. of best wire
The simple effeetive diameter IS mel~sUIeb .ul'~ 's"'et 'to 'a l'11~ster c;nsisting of a
.
1h
acrne ell,
". ,
Size, in a measunng maC,1ll1e,I ('ft - ,e~. m(; t h ""1'1
t 1feadOC'1. e' , t11e Inn ''L'er is built up so that
pair of vee notched arms ana o",lzm b , v
., i value for
the thread to be
the effective diameter of the vees IS the nomm",
'Cd
C+e
The magnitudes of C and e are both normaUy very smalJ, and as they are of
sign their toU.! eft;,et 13 3till less and
may be disregarded. If it is
requlrt;Q to determine their values for the measurement of ref,~rence gauges,
renders are referred to the National Physical Laboratcry Notes on Applied
Science
'Measurement of Screw Threads" obtainable from l:LM. Stationery
'
meas:lred).
'I" "F "EASURING MACHI~IE CARRY:NG
)1
I '"J E
AR f'..) v IV! ~
BALL-ENDED STYLII O:="BEST WIRE S,'
Office.
in V/l}ich
+ 817 cot
simple
el1(~ctive
diameter +0-0105
+ 58;) + l'9208p
,
d N t the vee l<l'''S OIled
di;imcter of an imcrmll t[lIea " 0 e fi t' : to sw'l1"
the fioHting \\'ork ita:;e altov'lUlg t he x -:~~, .~
L 0
taken up by the t111"llu.
174
etr~ctive
diaIneter + 001
+ 1732 op
half pitch
~l.nll
the attitude
:hi'ead
simple effective
v'
175
eH
1!
R 9
I
1
of
9.13 Wear
It is a well-known law of physics that friction does not depend on contact area.
However, the rate of wear is dependent on the areas in contact, the larger the
area the lower the load per unit area and hence the lower the rate of wear.
'
(evcmpment
of modern technoloaies has called not on 1" ~ .',
-v d
".","~,M
b' t !.,
5,,' " ''''; c 'J 01 l,nprv e control
a
working: on"
,
,.1 >d ,:0, at, t lC sU,rfa,ce textur,es, bofh wori(ln(Y and nOll._,
tl'i(
[new :trn 1"lre 1. ' I "
, .",,,, _ _'0
control ()t'
r
,'~'
v , v,,
e lJaSlC e.:tIccts wllich hElve l))',de tIle
.. . ,
tC" Dr p nop Ohf'.l1,
-i
t naml~ 1y, f atlouc
','
..llfe beari!1 a pl'opel't'
d
, "vear.
'"
,~o
/
lCS, an
~'f
HIE
L'
".
'
"']
,,', '
315
!
",G
>
'f
9.1J
fatigue Vfe
fatigue and
an eqtll va[el~t
at a given itresi-i~s
SeCITIS to SGlrt at n sbar1) corneoccur, such as the root of '1 ~t'-f:1U'
i
1,
~
;;;ven
". _ _ _ "
I;.
,_
..
.i'
,,b'''.''!':'';Y'll
1., _
L ___
L~,.-_J-'-1""':~~AI"'''''J-;'':'-''.-L-~,..'...'~ J.l
mClAeJf~~~
-~--I '-_..._--"
003 025
oa
OSD
2'5 mm
LONG WAVELE'lG,H
CUT-OFF
Fig. 9,1.
finish.
(Courtesy of the Rank Taylor-Hobson Division ane! R.E. Reason)
177
176
-,.~
20 cmci"eqJand
C1.1t-o
>
\1"
h
....
-ass It
h moug smootl1 it IS ~Qt flat. It
that if the field was
the resultant earth mo+/cments may
. l'
l'
1 C
'l(\
f--~ Would
lll1Clu.atloil ana so it IS v""c;'''''''' WILli urrows .;. ~.cm ueep but only 30 c
fr5mcrcst to cr~st. If the ,vork
the field wlll DOW be /fat but
.
1 I
0 cncKct
. I l
'
0t ", ~ I ~,' I.t
"h'
. S 111
unsUll::w,e
ror
oecau~c It
}I:: ,he; '.k.gu or L, e lrr'~gularities
t:n
c'
.
(l!st~nce or the wavelength.
WOUld proba:Jly make matters Worse
[is the field would Jim!l!y be beth
and rougb, i.e.Ule roughness Would
expended WOUk1 ;lflve been better
a wavy one. However, the
prim is thal the difference between
::md wavineSS is one of wavelenoth
",cher than depth of irreguhlrities and ar some point il must be decided Which
changes 1n elevation consIitutc roughness al.id \vhich constitute '.vaviness.
1.000/25
2000/25
01 K)lm CLA
conc,::;rned.
?OOC/25
04 i~]JmCLA
The manner in which the two may be superimposed on a machined part can
th~ requirements of a lathe to produce a truly
cylindrical component. I: is simpiy ncceSS<lry that the tool shall move parallel
Ie lb.::: axis of rotation. COIlSider a lathe in which one centr'~ is higher than the
olher. This would c::wse a bobbin shaped surL1.ce (hyperbolojd ot' re'iOiution) to be
if the tool motion is not controlled ii, a
line a series
c{ shorter Vi~~/es Illt(1 be $upcrin:'PQscd on the
\vave curve) fl!1d over all of this
will b~
th,~ f:ed marks of tl1e tooL
sliil shorter wavelength may be
and the surfhcc thus cOlltains four types of irregularity
causing it not to be a true (~yli nele!'. These are:
b",' illnGtn.tccl
1, Row:d1l1ess.
I
wm
179
;I:CTS
<"','
~...,
should, be noted that all such traces are distorted, i.e. the vertical and
horizontal magnifications differ, to enable significant vertical featmes to be
observed on a trace of reasonable length, If a sample length 0[08 mm was
magnified 5000 times the resuIting trace would need to be 4 m long!
",
1'''
."
::.
.....
".-..
.L
(.l,
....
....
.t..
(e) A means of analysing the profile thus obtained. This may be incorporated
in the instrument or done separately. Instruments s2.tisfying these characteristics can be produced with mechanical or electronic systems of
magnification,
#
"
'-'
9.3
FINiSH
T, 1
Ile' "1ctl1("'s 3.'1'11'1'11,)1" f'- ill
'"1
fi' 1 ~
.
.;.
LV,!,
' . c_
'" Ot
easuflllg
L lC . ms 1 or a machined part lnay lJe com, _~"'_ " o,r oy ,dm;ct m~asu:ement., The ~ompv.i'ltive methods arc attempts to
"b""::>,, d,e ,mFace texture oy ooservuuon or ieel of t11e surf3.cc, Thev ~l.re
li1 t!:at
can be
if cOJilpai'ison is not made wi"thV"'''''''~J
,
sliIlllar
The appc::tn:.ace of a surface depends
on
trLe scrat;h.p~tr'~rn, 11: the dire~:tio:l oE'the scratches. and is less influenced by their
,"
,llls IS equ::i.,llY [rue 01 mlcroscopic eXClmil1:~ti(ln as is shown ill Fig. 9.2
\\-'11IUl rCpl\;scrrts mlcroscopic eplan!cments of six different
an1~ t11e[
sllrnlce proCHe as 5110\vn by the rnlcf;.
'
~ ." L. r..
l~ouch is
a better rnethod of
than visual
it can be
and
ShOl:.Id oilly be ruade
proch:ced by similar precesses.
standards are available
by lvlcss[s.
~~{: Co. of
being for dllTcrent machinin~
\Nhilc the l'{orton Grinding Vy~heel Co, produce cylindrical ground
st~Lnd:.lrJ8.
,i"i(e"[
~. ~;~
value
'.0 Lk'
mSL'Ufl1Cllt3
In
nO""SI"'e""'ont
"",~""
' 1opec I to enable a nurrerical
_~;~l
.<.u\.,i.llOUS
jjuVe b cen Q.::lie
011 the surt~lce
and tiles.:: ,:Ire almo;;t all stylu~
"
ImerferoTIltcl'.lc methods are
for reflective surfaces.
_.L .. <.1
p.,~kid
"
9.33
an cases
(,c)
,1,
-(.j,
"
~ I
>.
over t'.ne sunace
anu1 wI
O\"f"'"
ij~
"0
dn-;Vll
"
'1
datum.
~'"
,
nf'j1e,'al
<.:;;;'"' -
and yerri-
180
181
FiLTERED
WAVEFOR~l
V\/\/VVV\/V
~/===/-1
///:./~:j>/l
METER
l~l[t
~-=i---,J, ~ Li~R
t0."/
l
PR~05CWT~O_R ~
_ _ ____ .J
---'>
HORIZON Tr\l ~fIOTION
OF Ir;STRUi--IDn BODY
A ('J\/\ n
A v f\/\ ,.'
/,-,
vvrv
v
CARRIER
MODULATED CARRIER
Si'-1O,'\ED GLASS
DIAMOND-TIPPED
srn_us
" The
as on the
consists of a
and shoe
";Vll,~h arc: drawn across the surface under rest by an electric Inotor and
cn
/
an armature
about the
stampings around e.2ch
outer pole
. p'
:l.C. current as ll1 19.9.4.
;\, t,le arm:lture
about
M it causes the air grtps to "af'! and thus
modulates
the
ampiitude
of
lh,;
'i C C"~~<>ll' j'10H'1'j'''~l'n tl"'"
J
...
_ . . .
".....
..,.;,1."-.....
L......
i \ S tlle::;;e
Ic>rm p:.ut c:t' ~,~,)fjdge,circu\t the output ccnsists of the ml~ciu;ation
Thi~ is
fed to an am pl;ner anu c:wscd to oper1tr.; ;J. Dell recorder to pr()dllCt~ a permanent
!1nd to a me~cr to
a num.cr~cal assessment direct.
IT;c l~)en rccord'~r 7~, ',)1"
t' t t'
.
d
v. "
HI !1~L iet: trace is pro llced \L ca;'bon-backed
a {)ointc" who!;c
arcs
the paper. This
R much .finer line
ink
due to
In
tillS
CC;:':tf0
I..
J.
t V !. .:..,
....
.132
183
for Engineers
---j
.-
'II
9.43
7 7\\--------- ./--1\l
II
,J\
!-..,\ I~r-.
I
-<1
1--.
..........!
-I'
'\1\,
. ]\
!\
Pz
-1
."'---
i'-
JI
--l 1-'J:.V
Fig. 9.5,
,.
/
"'i
0-1 L
T!lis measure was standard until 1955 l' n thp USA af'er ",'[ll'ch' 'h-' dand:lrd
0..
was
..
-.
. ""
hV"'U,""" to Fig.
1.."
I-
.,. +hn
2ft
n
To determine a C.L.A. value by the erection of ordinates would be a laborious
process, and if an unfortunate ordinate spacing was chosen a sigpificant point on
the surface could be missed.
However, if an irregular area is divided
its length then the 'value obt3.ined
is thc average
of the area. Such an area can be measured using a planimeter,
thus"considcring an infinite number of ordinates, and every p0int on the surface
is considered.
to Fig. 9.7 it is seen that
\.>
to an, (J.'leragc nelght (see sectIon 9.43). It is defined as the square root
~f (he mean 01 the squares of the ordinate:> of the surt~Lce mer;.sUIed
a mean'
+ .. , +.111.
C.L.A.
Ime.
Consider
9.6,
h,'/4'
I...~-----~----L-------O-O'iL
The C.L.A.
is the standard adopted in Great Britain, and since 1955, the
U.S:A. It is defined as the average height from a mean line of all ordinates of the
surface, !"";U.lI.m;;,~"
C.L.A.
--J
v)
2A
L
9.6.
=--
/\.
r~
I
-;--- ~\ r-"'!!
I
I
~
Fig. 9.6.
ordinates
aI'''!
erected at 1, 2, 3, 4, .. , n, whe.se
are
Fig. 9.7.
by
1:A
10 6 p.in
[84
x ___,.-___,c_ __
lTICign.
where 17A = SlIm of areas in mm 2
L
length of trace in mm
. th
so
. at
(a)
by eye.
(0)
i.e. the correction distance for the trace centre line is hj4 downwards (as
shown by the minus sign). This brings it to ~ from the peaks, which would be
correct.
TIus process is geometrically correct for a regular profile of this type, and is
approximately correct for an irregular profile of the iype encountered in a surface
recording.
The technique described above is that required for the Tomlinson instrument.
The Talysmf average meter incorporates an
device which does this
work while the probe is being drawn over the surFace.
.
the position all amoun t equal to ;:he difference in areas divided
tile trace length.
i.e~ (~oricction
EA (below)
..- 'r
,R_efeI~ncc
This
triangular
.~
over
of mono-
~:lr'y",t1u:e_
VEE GROOVE
'---'--r1{5
'
I-....----::,,;----i-~
------
hb 9hb
It
4
,/
"~l
-l]
~k'-,
Correction
.--7
~~----------.~~
Tb~n
LJ
INTERFERE'NeE FRI~IGES
ON SURFI'.CE
Fig. 9.9. Interference fringes across a fiat surface having a shallow vee
groove along its Jength.
-~'-..;--
]36
187
d A
.. Depth of groove =- >, 2;"
s
and in this c.<~~~~depth "i= O ! ,urn
.1."=0-5 p.m
d 0-5
.. 01 =-~<
s 2
,-. ~=oE;il:
S
,_) = 0-4
__ _
(b)
(a)
the total
. it fonows that the
i.n terms of the half
/ fringe
indicates"
the light bi;;lng used.
. 'oJ
'.
ci as in f,'jg. 9.10 has a !~ene]'a~ fringe spaclng
A surnlce appcarin
- ,-----;;::;
-.:::. of a'with a
~cratch indicated by fringe irregularities of hdgllt h. "-~-'
"-'
~,~--
Fig. 9.11.
Microintcrferograms of (a) the surface of a steel bali; (0) a scratch on a flat surface approximately "io or 025 jAm deep.
(Courtesy of Hilger & Watts Ltd.)
".'.-
(,)
'
Vt1\
1~\1IMt~v1vl1'''OO
L- f~~~t~-T---1J
I\
in a l1nt
micminl<'!rf~rogram
" lnercj'ore
b ;(?
A
. 1..,
The; depth of the scratches is
a _ I"om wnere.\ IS Le
;"j',f
j ,,;;;,;,,<>,
"Stv-j
Lo,,- "":""'
"',
th 0 f
Wil'iClCllg.
}/~
,.
'rl.l.is work cannot lie done with the normal opical fiat_set-lUJ.? bnt It requires
an"l e LL) (he onticd flat.
veri iin') fringes, and a micro-
resolvmg.
systems to b c m r i ' a c e s of
indicating the roundness of the bails,
II
~urf,Kc."An
~r--T-'I'
IK
I i q\ I~ \~ tOOO
j N ~I
~--Ti-- -I' H- I) ' ]
Fig. 9.12(b),
188
j.,
189
the Interference
An intcrfenm~e microscope
&
mcorporatinl1
wor'< is inade by
records to be Il1ade~
tel:'feror~nlf(lf: pro'(iu~~d
afe shown in Fig. 9.11 (a) ilnd (b).
T'JJCA METHODS
v.,;hcri~
the slJrfac(:
te~(!ur~
CHA
10
,,
of a corn,oonent is required
and
not reiIectiv~
a trace by
is
10.1 INTRODUCTION
THE previous chapters in this
the fQct that althcugh
accuracy is of
importance to
accuracy is unattainable.
Even gauge bloc.:ks of the highest order of accuracy have manufa<.:turing loiew!lces
which are as small as a few millionths of an inch. These tolcrances,wilose
tude depends on the function of a component, arc necessary to allow for the
jnhf;rent vf,riability of the production process.
Consider a simple operation" such-as a
lalhe set t.')
cif
bar stock. WitrUfl limits, the lengths parted
will
the same, but
around the length at which the machine is set will occur. These V2.riations fQ"ty be
due to a number of caUSe5 anyone of which will be of "
but if nil
Ciluses tend to produce a size increase then a piece of maximum
will be
produced .. Silniiarly, if all causes tend towards Q decrease a piece of minimum
length will occur, and if all causes work
each other their eiTccts wi!l be
cancelled and a piece close to the set size wili
cut off.
These random variations are due
non-assignable causes. It is the
of a system of process control to distinguish between these and
canses such as the machine setting 'cliimglI1g, a grinding w11l::e1
machiac failure occurring. These "ssignabLe faults must bc deter:Led and corrected
before defective pieces are produced.
Apart from component production, faults can occur during
and a
large manufacturer buying in
or sub-assemblies must comrol the quality of
the goods he is buying. This aiso can be n.chieved by the intelligent. application of
statistical methods.
Thh chapter will therefore consider the ways in which. statistics can be used
to perform these functions in ::tn economical manner.
it should be borne
in rriin{f that there is no substitute for 100 % inspection, a:ld even
Imman
nature being what it is, faulty parts may be accepted by the inspector.
It has been pointed out in section 10.1 that aU manufacturing processes, and
indeed many natural ones, are subject to random variations. Tills may be shown
190
191
"I
1.
siZt~
__
and as each
l",v
42
measurlnO'
'
marK
on t hIe chart"
'
"\
FREQUENCY
5~1
hJI-HH
Ij
3l
+.
rllH"
o-JHir
I
OOOlmm UNITS
0-001 rnm
18
i+# 'fi-Hlif l
35
\,
oj-
"
5 SIZE
DEVIATION
35
15
<] 001
mrr.
UNITS
Jl' 1'1
<;
STANDAHD
DEVIATION
5.j
'It
'TGtAL
Fit:. 10.1.
FREQt.:ENCY
Tally churt.
shews some of the "har:lctcri~tics of ~he process variabilto ~')ca1:0 as" a ".
() f Irequency against
Dolv.~~orl. is S~10Vj n in I~ig. 10.2.
"F.,:.c;:~'~'t'a-~l'O"l of:J
"""fJ_ln",-,'-'-o.""
1 f c~ln
these
t...l
is 1../ery
si~nih1r
25 '
.-
13-5
U1
Z
<l:
::;:
5-7
04
309 (f
....
and the
the
.
~
SIZE DEVIATION
192
!
'
7l{ otr
J,v..." ....... ,
o!oa-y
fio!'
/0
r:>.
OF
in which JI' and ,; arc the mathematical constants; .:1: is the me;'.n
x is the
size of the indhljJluai
:'.y is the
\vith vvhich a
,.x ('c~uro'
v~:~
:;', and (1{S'
the stanchc,rd deviation.
of i.his WIve which are important in qualit/control, the
mo:,/:
being the Ltct tha,t 998~/~ of the ar:.:!a--12:llder the curve
/ ' 1
1
T
<'
. }
'vVqJCfl represents
;;OLl1C SC~LlC tnc tor;}l nUlnoer or parts
!1eS INithin
the range of 3'G9,'~~ rron:. the il1eal\ i.e. all but 1 in 1000 at each tail.
of St:;! 2000 p:.lrts is taken, and the standa.fcl c{(:viation calcu- '
Thus) if a
lated, it can be
that ,1;:;
QS the n1nc~1.i.ne
in contiol, no change
cD-king
Hl
or precess
praduced except 2, one at
each extreme, \-vili !ie within 'Clc 3090- from the mean SIze,
t
6-001
6002
6-003
1
2
6-004
4
5
6
7
IS
9
10
6005
6006
6007
6008
6,009
6010
Totals:
an
Fxx
-5
-4
-3
-2
24
60
35
42
l75
252
252
144
54
"
20
166
988
25
16
9
4
-1
0
1
4
9
16
+1
+2
+3
I
I
I
F(x-x)2
I
\
8
15
36
18
6
25
48
72
60
35
0
36
72
54
32
434
_.
Fig. lOA. Tabular calculation of standard deviation based on lrequency drstnbullon m fig. 10. L
Mean size
X=
The stand~,rd devia((Gil may be ddined "s the root mean square of the individual
deviat1cn fr~)rn the mf.;an :;ll:G of thG group and .is
by the expression~
EFx
.EF
. _ ;\,=02506 in
(7=
in which F is ihc
"fIle )TIcst c~"'>~'.tvc!lient. ",'lay to c(1.1,~u!ate (J
a tabalar TI1cthDd as shovvn
in scc'~i()n 1O.4~ th~ v:.~ILi(:s of)~ ar:.d {;taken fronl the
chart in
10~ l.
L'~\.ssun1e th~lt
~onlpl:rator \vas set 'whil gauge blocks \vhose size ,vas
6005 !nrG. ~rh~ cQinrnn~; in the t~~blc tire as fOtloVlS:
is
CO[Wrm ]
b:..!'Her
Collimn :f shows the deviation of e:ich size group from the mean, i.e. (x
vutu~
t.'i gi'l..:t1 by
(f
x).
. _ a=O0016 mm
5ho\'15 n E:>~ of the size groups used in which, for simplicity, the last
*A
u=.J1434
166
:t-, is computed.
Then (J is obtained by civieiing the tot,ll of column 2 into (he total of column 6
and finding the square root.
In this process the standard deviation =00016 mm ~
then mocess variability = 309(1'= (3'09 x QOv16 mm)
.
1
= O'0049 rum (to nearest 00001 mm)
195
194
7S
i. f,'.
~
;?
c'J
-r
I
(a) The mean size of the work produced may change, i.e. the setting has
moved, e.g. tool wear has occurred.
(b) The proc::ss variability may change, i.e. the size range of work produced
'has changed. This is usually due to something more fundamental in the
machine.
'!J
,,~ "':''''l i ,.
,.J~./I:)...
......1
1022
If a
parts was take':} and
of, say, 5 parts,
then the [rverage size of the sub-groups could b~
:~hought will
show that these average sizes would b,: grouped more closely around the mean
than "mula the sizes of the indi"j,dvai items. Typical frequency distributions for
the averag(l Slzes and the individuuL; arc shown in Fig. 10.5.
-.~~:-.:-
f-
_---SCRAP
UPPER WORK liMiT
!~
---~--'---r
+
I
I
__k:::
__ _
__,~ _ _ ~~GJt'A_L____ ,
OJ
~ ~~.
U1
.---!\
MEAN SiZE
r-Fig. 10.6.
LOWER
---,--,--<>-
TiME
during a proc<,:ss.
rU"'-,::: SCRAP
~_-1\ER we",
ab
(j
WORK
TOLERANCE
__ :..
~.AN
LIM"
SIZE
SiZE
~-:~-.~
......,
SCRAP
TIME
;.:;r;: .~
..::~:
If a
mach;nl~
196
to
a given
t1VO
1.
iii
197
,n
m~
However, the process can drift from this mean setting an amount Q before
scrap is produced, Q depending on the difference between the process variability
and the work toierance.
Thus the maximum distance a sample average should be away from the mean
-===~_r11io (W
'1
parts
l'I"'r'
size is
T'
average a
I He
(Q + 3'09a
).
__ ._C~';;'.~.:.:~
or non-assign8.hle, crruses.
t<)9 rifL
1.0.3, vlhich
for individual items
i l l ! P?(R COI)"OL
UHIT
SIZE 2
: (,..,
3'OSd.
~
tAL
:::
,.,
SIZE'
!
I
~.
INSPEC,ION iNTERVALS
Fig, 10.9.
(Q
Q before
averages curve
Immediately the points plotted show a definite c1rift to one side or the other
of the mean
the process should be treated as suspect. Before the trend reaches
the control limits it can be stopped and the process reset. Note that ir. a process
where the drift is lBEaHy in one direction, as Oil. a grinding operation where wheel
wear occurs, the res~c(ing can be purposely towards the other controilimit, thus
ensuring larger rUllS between reseUing.
198
199
C:;"
-.r.
,filmp!!jzed
iv!ethod of Setting Control Limits
appropriate sample size to see whether the process is of low, medium or high
l;mit~
rel"'l've t, .t,
~L , () ,~.,e mer.J.J1 Size
. would
- lOtH , orev,ously
c1esr-'!. d .
I' ureA ' i "
.
,
. .., (uC lllvolving
.)
".u. <~owever
It
0"0 Ii;
.
a
'l;)pC'lr to oe
' 3. 1"
,v
!8.DOnOllS DrOCeSS if the m"')..,od'
','"
1 . J t'
f .
'
"aCll a , JOll d
0 tile standard
t1
"
.-
deVl'lt;oIO
I;
"
,,11S . 1: tc
q':i-f;
ri
,L.
,
tne range 011:' \vork sizes whic71
v. "111 ~ ,
'l ,_eV1U"
gl\Cn l!rocess :It one
h
'
.
11, 0'_ oroduecrJ .
,,'
. -:.
ulUS ~ lere snoll)J be no ' 'c " " ,
,,~.
etwIII a
scundard
dGVtatlon "crl t].= ci e'
.
mt: IddtlOnshlp
h
.. -t 'T
'
, - Jc
Iv 0.Z fdng'; 111 t,ie sampl~s
", 1
v,
een
ella 1" LhIS IS so, ?,nd it Freatly ,'ilnpJ;.cnes /1 _ ,v.
"nd plotted on the
''/V'
' . ""
~L_"
Ie WOll', 111
. '
" ne~ a cha1't :, m operation the [llspe ' ,
a conlrol chart.
u~ "I" , I
c,or
a"d
. its
~l1Vt'f'1
. , b'-' " ~c. S110U ct carry out his calculation on a
., p IottlDg
dot,an oda scrap of paper. if this is d ' r]
.
data sheet
settl~g
up {he chart.
ODe. l<,:;n the same data sheet can be used i~
..:
....
v.
!'"
rT"
.)
~'
" . fo set
up the chart the inspection
1'1,vql
' ' , IE'-fLy
,"
1"
'
aeclded
U')Ol1 ~ no till'., f.
~ ,
ana t.1~, sample Size shOl'ld b
'.
'
,<.",
:; OLnI or msncctlOn i'arrii,d
t'
'.
.
e
~Xlstl!lg
method. Having noted t'l"" "'7~ [ - : , : ~l1," m !/,:Sf)Cwtlon with the
,~,
t
~
Lv SLxS 0
,ne llld"/'du"'"
t'
or computes the aV~1'8g'" 01",1" "'f l'
"
- d
1.., lil lIe sample the
h'...;;pec
1"'"
.... .~d" L~
He
''1n "loc [I' ,
/
we (,merellce DetWf'>c'n
" . '.
u"
l' >'
v" tn'
~_\.,;,> m~";~
"l'"",Alll.1Um an d
mllUP"q'~
s
.
I .J.e.samp e ranoe
b , .......
a lx;riod
of
time
a
IIUJ!l bel' of values 0" S"1"I"P)" ".~'" Izes m tIle !,ample. Thus over
'11
.l u .
<;; [nfl''" de'10t~rl by'
.
relative precision.
If low, the process is unsatisfactory and likely to produce a proportion of
defective work, i.e. Q is negative.
If medium, the process is satisfactory and, if well controlled, will produce no
defectiyes. In this case Q is zero and the process will require frequent adjustment
to main rain satisfactory control. The contr01 and warning limits can be set by
multiplying wby the appropriate values of the constant A' in Tabie 2 on page 202,
If the process is one of high relative precision a drift can be allowed to take
place before the setting is adjusted, i.e. Q is positive. In this case the control chart
limits arc found by multiplying IV by the appropriate constants in Table S, In
this case, to alloW for the process to drift, the limits are set in /rom the work limits
by an amount A"w.
Thus in aU cases the control chart limits are set by multiplying the mean
sample range wby constants provided in tables in B.S. 2564,
'
';;
.Il~
If c
,,'.'
b'~,
"
x,,
lV,
are avaIlable
"
It was shown in section 10.3 that, as well as a control chart for averages, a range
chart to control the process variability is required. The control limits for the range
='= Ew
ns
1
'
,Yncr,e
tis 15 the number of 5amp1es taken
.
It has been shown that a mCaS'!fC> of o~, "'~, ., 'l'. .
DOll, but to con1rol th" "-oc r , ' ' , ! ,,?~~S mnaOblty IS the standard devia_ "'LI.... IJl
t..:,..ss
.~s
tonda ~h
control Ghart. It follows
thflt'
,,1"'0 . . '
,0
tne
,dn
.
1 i:", I.1J.LJ v 1$ a corl"r':f'''''+1on hel'~':'p,>,,",-,'1
h sample range
t e standard deviation l' e
~ . ~ "'' ' ' ' ' ,-' 1.'(~ c,.A"u tHe
sarllple range
,
chart are set in a similar manner to those for the average chart, i.e. constants
have been arrived at which, when multiplied by the mean sample range, iV,
the positions of the upper and lower limits for the range chart. These constants are shown in Table 3 of B.S. 2564 overleaf.
..
a=CiV
(Text cont.
and
[{.F.!.
it
OJ!
page 204),
\vork tolerance
200
T::ik~~i1
20~):~
i,' ~V"!'ll'~cd
A"J
..... r.fol- l.~ ~ l "
. t t'h e
agal!1S
201
o
.?f the
British
the complete
"
I
ongmal.
JI
i
Class
Low relative
precision
lvfedium relative
precision
High relative
precision
i
I
Sample size
(x)
R.P.I.
R.P.I.
40 to 50
30 to 40
25 to 35
3
Sample size
Foy
inner lindts
C
A \'
123
067
048
0'38
032
5 and 6
.1..' U.001
0.025
3
4
For
outer li.rnits
194
105
(}75
059
.
0,)0
60 t.o 70
Unsatisfactory. *
Rejections
inevitable.
State of
production.
R.P.!.
greater
greater
greater
greater
than
than
than
than
70
50
40
3'5
Satisfactory, if
averages are
within control
SatisfactoI)', if
averages ari;
limits.
limits.
within modified
$ Not n",,""r.~~,rUy. if the specification permits a small proportion of (he product to bo outside the
limit. In such C:Jscs the limiting values for low relative procision can be Og of those given above .
TABLE 5
(x)
T/\.}3LE 3
Control CaD.rt Llmi.ts fur
(w)
f sa~'l'lei
Slze
n
2
3
4
":J
6
i
I
-I
For lower litnits*
D' 0,001
D'1).025
000
OO"}
0-18
029
004
010
(H6
{}21
For standard
Number in
deL:iation
sample
D' 0.975
!
,-. !
1) l).n99
193
41:2
299
258
181
236
172
222
23l
217
:.l.n!
eli!
2
3
4
5
6
113
169
206
233
253
For tnner
modified
limits
A." Q-025
0'3'7
0'42
To obtain the limits, mUltiply wby the appropriate values of A" 0-025 and
A" 0.001> then aod to the lower drawing limit and subtract from the upper drawing
limit.
,,
:
151 (0'83)1'
116 (0-71)
102 (065)
0-95 (0-62)
090 (0'60)
!
!
For outer
modified
limits
l\." 0.001
080 (0'12)!'
0-77
075
073
071
(0'32)
(0'J8)
(0,41)
(0,42)
For altemative'*
Inodi{ied limits
b1J1er
,f.
I
I
i
I
I
Ou.ter
!I
r\ 0.025
AM 0-001
2-32
170
161
1-31
146
1-34
112
126
108
1-19
\Vhen the inner and outer limits are close together, one of them can be
omitted. (Sec also Part One, Section B4 c, page 24.)
manufacturing rejects.
The aitemative ligures in brackets may be llsed when the buik is permitted to contain a small
proportion of rejects and the R.P.L exceeds the value given in Column 2, Table 4.
202
203
" " .
'1
h
e e
U11provemem
III tuture S1ll1har operatlOns and t e overall quaLity improved.
:n-us sl:m::-Hlry s,hvws, step ?y ste~, the setting up of a cox:trot chart, and shOuld
De read w1th reference to FIg, 10.10 and
10.11 on facmg peges 206 and 207
The data used is that from the taily chart in
toJ and the data sheet,
10.10'
may be used for both
up and ruuning a system of control charts.
'
It is recommended that the data sheet be printed on the back of the control
chart, which should be of squared paper. They can then be kept in a transparent
envelope at the machine [n front of the operator, and filed when completed.
10.62.
Let x=expected number of defectives in the sampJe
Step 1. Decide O1f the sample size n and the frequency of inspection, These
sho~ld. be chos~n to give a to~aJ of 10-20~{ ~f toml productiun. A good sampJe
SIze 1:, f, but 5 13 more converuent for computmg averages.
Step 2. Take samples at the decided frequency and record their sizes on the
data sheet.
3. For '~ach,samp!e calculate sampie werage and sample range w.
SEep 4. 'When enough pans (80-JOO) have been inspected nud recorded calcu~
late the mean sample runge w.
Step 5. Calculate
Relativ'~
n
IJ I
ties. Consider a process running at 25 %defectives, a sample size of 150 being used.
x=2'5% of 150=375
and e-"'=0'0235 (from tables of e- X )
'work tolerance
Til'~re are mallY ca;:es in inClllstrj where the il5C ,)1' cCl;,,:rol charts is Jcsirable, but
is not
in the fOtTn ,)hO'-NU in section 1O.3~ Such C~t.,~e3 are tllose v/here
i;onno] of a n:lurrGt~"cturin,~ proCeSS by controlJiJ;" a fe'.\' dimensions is ,lOt con"
vel1ient. Consider a produc(iun lIne producing an assembly which is to be checked
204
ofn
2 defectives, 3 defectives, etc., appearing in the sampJe. The warning and control
limits are set at 1/20th and 1/200tll pro babiIirics respectively, and we therefore ,vish
to find the nllmber of defectives that will appear in the sample with these probabili-
'Precision Index.
f~ . .
=E%
It can be St~en that the 7th term, representing the probability of 6 defectives
appearing in the sampie. is 0'091, and the 8th tenn (nrobabiIity of7 defectives) is
0049. It follows that the warning limit, set at a probability of 0'05, must be set at
6 defectives, as more than this number can only occur with a probability ofless than
005.
Similarly the i 1tll term shows that 10 defectives wiII occur only with a pro bability of less than 00' 5, and therefore the control limit is set at 9 defectives. These
limits are shown OIl the chart in Fig. 10.12.
Note that thc warning limit has been exceeded at sample number 5 but the
process then improved and no action was taken. This happeneu again al rH.llllk:r 14
c
3:
0::
.C(
c')
r:o-r
~~ED
These charts would nonm.lly be on the back of their data sheet (Fig. 10.10). Note the indication of drift
on the x chart towards tho U.C.L. Rcsettin~ will shortly be necessary.
Fig. 10.11.
LO'S[R UL!iTS
n ,;$ CP<;E
Fig. JO.lO.
n::L.'HIVE
\'lJ\RUlil8
i::l
......
..r;'
E"~
V')
~
-..
;;;S
(")
"':!
~:
-.
te)
eJ
11;Jetroiogy for
Statistical
and the process th!;n deteriorated, actIon being taken after number 20 when the
control limit was exceeded.
SampJes ar.o; tak;':11 at
intervals so that a total of about 5-10% of the
total produc:icn is
and the fraction or pe:'cerrtage
in the
sample is platld on the chan. it should be noted that the
size must be
such thac, based on the
average percent defective, it should contain at
kasl Ol1C
i.e.
a P.A. of 2~ % the ';mallest
that should be
used is [00/2'5=40. [I' the
size is increased the system 'beco.l1lcs corres.
pOlldingiy i~ore sensitive. If the control limit is exceeded then it means that the
prodaction
out of control and has exceeded the allowed process
J)ercent
If the normal process average percent defective is unknown it can be calculated by taking a number of samples when the pro("'Css is running normally, and
finding the mean value of the
defectives in the samples.
,,';.~ simplified means
tbe control limit, along with a very full explanation of the operation ,)f such a control system is to be found in B.S. 2564-Control
Charts for Fraction Defectives.
Confrol
Consider a limit gauge being used to check work, the limits of size used for
manuf.:'lcturlng the gauge being well inside the work limits. In effect the work tolerance has been reduced and the gauge wiIi reject a certain amount of
work
at ejt11er extreme of the. tolerance zone. If such a gauge 1S uscd in conjunction
with a fraction defective control cbart then the apparently defective components
are in fact
but provick an artificial basis for running the chart.
These gauge dimensions are normally equal to the limiting dimensions of the
work less 20~~ of the total work tolerance on both the 'GO' and 'NOT GO'
Hence the tolerance used for
purposes is only 60~';; of th(~
tolerance. Such a system need Dot be confined to direct dimensional "'~'~"'.'. "'"
For instance, the calibration limits of a speedometer, the pressure limits of a
pump, or the limiting values of a resisrance, couid ali be compressed and such a
system used.
(a) Spot
This consists of inspecting a small sample here and there
and hoping
the results ,lre a reflection of the quality of the batch.
The method is
but risky and furthermore, the risks are llnknown.
(b) 100% inspection. In terms of quality this method, w]:Jch c01lsists of
inspecting every
item, is undoubtedly the best. Due to
opero.tor boredom, and distraction, it iG not 100 % reliable. It is also very
expensive.
,
;)
Fig. 10.12.
Control chart
['Of
fraction defectives.
(c) Statistical
This tcchnique requires a sample whose size has been
carefully
to be selected at random from the batch. lfthe
contains less than x defectives, the batch is accepted. ff there are more
than y
the batch is rejected and returned to the
for
100 % inspection.
Sampling
cannot replace 100~~ inspection in many cases, but where
it docs the system is
less expensive. Furthermore, the risks invol<;;~d
are known and can be
for in costing, expected warranty claims, etc.
Such systems are based 011:
(a) The sample
representative of the batch.
(b) A knowlt:dge of the probability of acceptance (or rejection) of a batch
containing a
percentage of defective items.
gauges.
208
..
2()l)
)0.6
R+B
R2+2RB+B2
R3+3R2B+3RB:l+B 3
Prob~bi~i~:: is
Y O.
"j.,:~,. ..... <LU:. ~ , ~,., '
lave
01 ,~r~Ci a ~lln OIl the football pools
lOf""otten to pu:;t lhe coupon, has a probahll1tv or O. All other Prob:lhTf '
fajl between these extreme:;.
'
c.,A lIes
.Furthl;_~: ~f an ~,!G1Jt. th~l.t occurs can do so in different ways, then the sum of
the probabwt1.t:s ot the ll1dlVidual methods of occurrence is 1.
"-
t.)
.!.
T];~
10.61
Binmnbl
(R+B)l!=Rl! +nR(n-l)B
.
IS
oJ It
05. Let us
R or 13
the individual probahilities
and
'NC
get
05 +05
RR or RB or DR or S:S
=0.0148 +0065
RR+2BR+ BB
(O5xO-5)+{2<O'5 O5)+05xO'S)
025 +
0-5
+025
=!
_. Tlv;
0-5 3
-j-
0125
(3 x
0375
HBR
+
+
(3
BRH
R3B
-r
0-142 ...
"
=0080
The probability of 2. defectives or less =0015 +0"065 +0142
g:m
BBB
0-5'l) +O-5 a
Q-:}75 +0125
210
10l:!
c=--_ _ _ _
'
Further to this the probability of 1 defective or less is the :lUm of the first two
terms=0'015 +0065
____
of both cards
red is 1 in "1 or 025
" both cZlrds
black is 1 in 4 or 025
" red 2nd black cards is 1 in 2 or 0-5
"
"
40 X
~",0'222
1
10.62 The Poisson Prr;oability Distribution
Calculation of prot' bility by the binomial
is obviously laborious. Another
expansion which relates very well to trJs work is given
the expansion of c"'c- x
in which the successive terms are the probabilities of 0, 1, 2, 3, etc., defectives
ill the su,,'1lple, and x is the expected number of defectives in the sample.
211
'1,
1;
.....
...
't.
.i
,. x = n x
"
x=lO% 0[40
=4 de!ectives expected in a
4 2e-4 43 e- 4
e-Ze X = e-4 + 4e-4 +
" 3 defectives""
"
of 40
The best way to show the characteristics of the system is to plot a graph of
probability of acceptance against percentage of defectives in the batch. Such a
graph, shown for this sampling system in
10.13, is called the operating
characteristic curw.
The points are best found by tabulating the functions in P =e-:C -i- XC X as
shown below.
Percentage
DefectiFes
in Batch
=0'192
=0090
The probability of 2 ddcctivcs or less
and,so
F A.."11"~t .. "
f' tf . :JrotJ~
. b'l"
. , this
..
, ,7
'"
:d<d?L .?;tCS~
utws shows .
now.summy
d~stribution is to
{!Ie ,)(~o:TIJal drstnOiltivn m secnon 10.61, the results for the same problem beina
very slnldar, bUi much more e:lsily arrived at
'"
SYSTE[VIS
Assume that a sampling SystCill has been decided llpon in which the following
inst.ructions are given to the inspector:
'From an incoming batch tak.e a random samDle of 40 compOll"nts~
-f~ 1
1 tne sample contains 1 def::ctive part or iess
the batch.
If thD sample contains 2 defectives or more
the batch:
This system is denOted as 40-/~'
l' 'l'
1
'I
"b b'/'
C',
1"
. . ,. ~ .
LO KnOW 1 lc pro a l.lty 01 a oatc 1 oe1l1g accepted (or rejc<.:ted}
If It Gonrams a
percentage of de.fectives. Iil section 10.62 it WQS shol'm that
the probucilities of 0, 1, 2, J, '" n, defectives are the successive terms of tJle
L
ex:prf.~ssi'Jn ~ -- Xe:~.
212
.C
<l
vJ
x
=Bn
0,5
02
10
04
-1'5
06
08
2
3
0:1.
the batch
"
%def;~ctives in
1'2
082
067
055
0-45
030
0,20
014
,6
5
6
24
0090
2g
0060
0040
7
8
20
J'2
xc-x
e- X
Probability !
of Acceptance
p =c- z + xc-x
0164
0984
0268
0,938
0-33
036
0-36
088
0,81
0168
066
052
0-42
0-306
0228
0128
0168
032
028
0216
I
,
j
Plotting the first column against the last column we get the operating characteristic
curve shown in Fig. 10.13.
It should be noted that this operating characteristic curve does not depend
on the batch size. In fact a given sampling plan will
the same O.C curve for
any batch size as long as the sample does not exceed about 20 ~~ or the batch.
If a larger sample, and acceptance number, is used the plan becomes more
discriminating, i.e. a greater number of good batches are accepted and a greater
number of bad batches rejected. This alone tends to keep the supplier on his toes.
The difference in discrimination is shown in Fig, 10.14 in which O.C. curves are
compared fol' sampling plans for 40]/2 and 180 3 / 1 (i.e. sample size 180. Accept
the batch if the sample contains 3 or less defeclives).
213
..
]I;Ie,troiogy for Engineers
O'B
01
ro
Fig. 10.1 J.
05
1-0
15
2-0
30
Fig. 10.14.
is given belo-V!.
e-:"
xc-x
0,0
18
27
36
0-407
0165
0-067
0297
oun
0-163
0267
0-245
5A
0-C05
0-097
0-027
0-175
0-073
DOn
0366
Two problems now arise. The above two sampling plans \vere selected arbitrarily but in pro.cdce we requir3 thc most economic
plan. Further, we
need to know what percentage defectives are passed into the stores through a
receiving inspection departmem operating a sampling system.
x2e-:~
0985
0-869
0709
0503
0236
x 3e"X
these
.,
I
. ,diagr~ms it is se,;J1 that if a batch contains ,~-'/
- , defectives
we::. m (Jot) cases It 11as a prJb~Lbi1ity of 098 of
acceptd. If a 'J"tch ha,>
defectivj~s 'tcll-L-(~j"'. I- Po
I
~t
r
I 'l~
'" S") . . , .
... ~
~
'-tVj/2 glV~S 1. a prOOaJ11t~' 01" ~.}'t 0 ot aCing
2.cG.';';}ted while the 180'1/;
a probabiEty of 0-71 of being
;wci, as the
pcrc;;ntagc
In
batch increases, this divergence in the probability of
acc:cptance grows.
cl
n'.
-It
f'j
215
average
percent
defectives,
known as the orocess average ~ Tbe rp'oduc"'" t,aKes
.
-1''' 1
_1 n+
","::'),
~
~
a
.IS\{ tllGl.~, '.rillton n.nnmg at proce;;;:; average, a peSSImIstic samule will ~-~'e
acceptable batch,
..
-- ,.JJ ct an
.0.:.
.;:.
..
",".l
..... '"
Th:::n
the.. total inspection per batch J in the ionp'
run, for a. '"iven JI'-l,
.. ~ ...r.,apI'In~
~
"
O . , j
plan IS gwen Dy
~
I =n + (N
size
n
R = prcducer's .risk at P,A.
",it is seen that [ is a fundi?u of batch size and ~ is a function of sample
".
.
,sample ::;lze for a
batch SlZe a curve of the type shown
In .t'lg. 10.b IS obmmed.
This
the most economical sample size for this batch siz-=:.
This
is fairly complex and space limits a mo.re {!lll explanation
'--''''.'''.'"_T the mor,t econoraical plans for L.T,}'.D. and caI1sumer's risks ar~
pubhstlcd by Mt)~srs.
& Romig, pioneers of this type of
in their book
Sampling inspection Tabies pubiished by Wiley, 1959.
_0
~I 1..I.S
Probability of acceptance
Percentage
DeJectil'CS
in Batch
I
Fig_ 10.15.
Fol' ')artklliar
ba~ch
size and
2!6
05
10
}5
20
3,0
Probability
oj
A.O.Q.L.
Acceptance
0985
0-869
0-709
0508
0236
0493
0-869
1064
1,0160
0708
217
..,
jVJetro!ogy for Engineers
-; ; :'21
&~
--,:Ii 10
j .!::i~~~o':"7Q!:':---~
'
:J~
~,~
:J ~
C1 c'>
I
t
oa
Many book:, some extremely lengthy, have been written on the subject of sta.tistical
quality controL This chapter can only condense the essential ingredients into what
is hoped is a digestible form. It is intended to stimulate interest in this subject
which, like those discussed in other chapters, is a tool of1:ho inspector and mctrolo.
gist, the importance and complexity of which is constantly growing.
It is further emphasized that the appilcation of statistics is not the cure to all
industrial ailments; but the procedures, and their results, suggested in this chapter
generally hold good in the long run. One sample does not give detailed information about one batch, but over a period of time a series of samples gives a great
deal of information about a lot of batches.
(') V)
2:0':
8 Vl
':' w
is?::
10.9 CONCLUSiON
>-0
06
I BATCH QUAliTY
GIVING WORST
~~:: /
.q:~
,{.-........,~-
AOQL
05
~--L-,
10
15
20
25
3-0
ThG
?It.
(e) If the total sampk (nl +il z) contains deii;;ctives less than y rrcccpt the
bat~:ll.
the batch.
218
219
TO CH
R 6
O!
lines should be whole multiples of the length of the base of the spirit ~~vel
;eflector s;and, whichever instmment is used, and it is advisable to select 51 e any
diagonai lengths in the ratio of 3 : 4 : 5.
The same principies may he u:,cd to determine the deviation from a true plane of
surt~tce such as a surface table or m[tchjne rabie.
A flat sun~lce is composed of an i.nfinitely
!Hnnber ofUnes, OT gel1erators,
and for it to be truly Bat the foEowing conditions must be satisfied;
reduced III
but whose ddc:: <1,'10
If the box Is filled ','lith
of
',vhkh is U.lC1} levelled
Nith a
which is
parallel to one
end, thm all Hnes across the sun:OLCe must be
(they were pTGduced by a
as
Similarly all 11:;11':$ at 90 G to rh,::se generators must b0
the straight edge Ivas coo.trolf,:.:;d by tV!) otiler
lines, thf;se being the edges
of {he box. 7hus ii' such a surface is te:,ted [or flatness
Ilnes paraLlel to it!:
sides it will appear to be flat. That it .is :not is dearly seen from Fig. 1l.1, it being
concave [i.cross one djagon;:d al1d Gonvex across another .
It is
seGn th;;.t iI the surface i3 to be verified as being truly fla.:t
then it is necessary to measure thi':of the dir,gonuls, ill additio11 to the
generatoi's paL'alte! to the sides.
The;;; il1c,\surement of
of alJ of these lilies of test may be carried
011t with ;tn a.uto-collimawT as
described in Chapter 6, but 11Uving made tJlco;e
mC3surelt',ente It IS necessary to ,elate each tine of test co ali of the others, i.e.
w;Iit'ying :onditiol1s (b) with whic;l this appendix is concerned.
Consider the surface shown in plan view cf
U.2 on which the eight main
gellt;ratofs ..Lre set out. These sh.ould. be chosen just rnside th~~ edges of the table
so that the
area, which i:,; prone to
is avoided. The length of tile
on
220
(c) Correct the ellds of AC; AG; and CG; to zero. This gives the 11~ights of
oints A C and G as zero and these thxee pci~ts then c?nstltute ~n
~rbitrary' pJduc relative to which the heights of ali other pomt:; may be
determined.
(d) From (3) the height of 0 is lcno\v~ reI~tive to the ~rb>it:ary P~~c
ACG=OOO. As 0 is the common mid-pomt of AE,.CG, UP, and A~D,
all points on AE axe now fixed. This is. dOD~ by 1eaVI~g A=O and correcting 0 on AE to coincide with the mid-powt 0 on "-..G.
(e) Correct all other points on AE by amount~ proportionare to th.~ mo~e
ment of its mid-point. Note that as E is tWIce as far from A as llle mIdpoint, its correction is double that of 0, the mid-point.
(f) As E is now fixed and C and G are set at zero, it ~s possi?le to put in
eE and GE, proportionally Gorrecting all intermedIate pomts on these
generators.
.
(g) The positions of Hand D, aud Band F, are known so it ~s 110W poss:.ble
to fit in lines HD and BF. This provides G. check ~n prev.lOus evalu~l:on
since the mid-point of these lines should coincide w1t11 the known pOSltlOn
of 0, the mid-point of the surface.
Thus the heig..lJ.t of all points on the surface are known, relative TO an arbitrary
221
It is convenient now to consider these lines of test on n. plan view of the surface as in Fig. 11.3 in which lines AC, AG and CG have been corrected to zero
at each end. Thus the plane ACG IS fixed with the points A, C, and G at zero,
~ and points on these three lines are all known relative to this plane.
It is seen that the mid-point is positioned at + 6 units (J.bove the plane, and
the mid-point of line AE must coincide with this position, while point A is known
to be O.
+2
-13
+2
B
+1
-2
"lj
+31
+6
-3
-1
~:1
r
,,2
s-jr-----__+_----l---l--/-c7~c__+---.J------!-__r 0
H+
/'5
+2
Fjg~ 11..2.
nunlber
A-E
,G-~E
A--G
o
o
o
o
+2
-4
-7
-1
+l
+2
12
15
-4.-,
-15
"- 13
I)
-2
-6
+5
+1
3
-5
+6
-8
;5
-1-4
--12
+2
-11
.~
-- 1.. J
\)
-12
-21
-24
-2
2
-0
+4
C-E
B-P
Cumulative
0
0
1
0
0
+1
-~- .J
+2
+5
+3
+2
+2
-2
+7
+9
-5
+9
-7
+6
+9
+ 10
Error
'.
0
0
0
-1
J)
-2
4
-8
-17
-l?
Correction
Height Relative
to Plane ACG
+2
+2
+4
+5
+4
+6
+8
+10
+12
+14
+16
+6
+6
+4-
+2
-1
-21
+ 18
24
+20
-4
I,
222
223
;
i.1
44
"AMi
Cumulative Correction
Error
Error
0
0
1
+2
+5
+3
+2
0
1
0
-1
-2
-3
i
+1
2
-4
-4
-5
6
0
0
+1
3
0
+1
+2
+3
+4
+5
+6
+7
+8
-6
-8
-9
11
-12
0
+1
,,)
"t".J
-2
-3
3
-4
4
Fig. UA.
'lll
arbift:l.ry
amount
below.
It. remains now only to fix all points on lines BF and HD relative to the
plane ACG.
Considering: line TIF it is seen that relative to plane ACa. ]Joint B has a
valu.e of +. 1 m:Jt, but the value of poim B in the table oi'r;umu!ative errors is 0,
so that inicial1y all pGints on liF must be increased by + 1 unit.
Error
+1
+1
+2
+3
-I
4
-8
0
+1
+2
+3
+1
+2
+4
+6
+3
0
0
+3
+7
+9
+9
+6
+9
+10
I Initial
i
Cor..
Car-
I rection I rection
+5
+5
+8
+12
+14
+14
+[1
+ 14
+ 15
-2
4
-6
8
-10
-12
-14
16
Error
Rel. .to
ACG
+5
+3
+4
+6
+6
+4
-1
0
1
It should be noted that the mid-points of both of these lines of test coincide
correctly with the valt:e of + 6 units for the mid-point of the surface. TIlis provides a useful check on the calculations up to this point.
It may be thought that this is the end of the matter, but this is not so, because
the plfu'1e ACG was chosen entirely arbitrarily, and the definition of flatness error
224
225
Plane Surfi2Ce
(c) By inspection select the closest pair of parallel lines which will contain
all of the points. It should be noted that one lIne will have two
on it, and the other line one point.
(d) Draw a ccntre line ZZ between these two and refer all points to tbis line.
o~ v~
~~ OF FSURcACE IN DlREO"ON
"\<KO
~,
O/~
;;;-------------71
',~
'\~
(a)
'NIL~
r'NHICH
.JUST CONT;\iN {,LL
POINTS ON THE SURFACE ARE
SEPARATED BY 5UNIT5
'~
'\
,
DATUM
(1;)
It' this
is to
shown in
11.4 <led select two points,
Wh0!,t; vc1.1ues are the masimum positive and
JrJ.xifnn~n
relative to the arbitrary
jn lllis case ACG.
Connect these paints lmd pioject at right acg!es to the !ine ;(,'{ connecting
ably on
theUl.
(b) ~::;et ,-,1'f to 3cale the !ldght of all points rdative co a line YY, parallel to
)~X,
\'vhlch
plane
.:~CG .
o
Fig. 11.6,
+10
The tme flatness error of J. 5 units, obtained by tilting in Fig, 11.5(b), can also be
obtained by projection.
226
227
FROM Ol<lGiNAl
DATUM
BARNARD,
BARRELL, H., 'The Eases of Measurement', Sir Alfred Herbert Paper, 1957,
POINTS CN THE SURF.,\CE
CCN'T;':'i~!ED
BY TWO PLANES
SEPA.,(ATED BY 85 UNITS
NOH;: SEL:::CTEO
Pf{OJECTlC~1
SHOWN. OTHERS
CLARITY
1. Prod. E.
- - 'Engineering Dimensional Metrology' (N.P.L. Symposium), Vo15. I and II,
H.M.S.O.
W. A. J., Workshop Technology, Parts II and III, A.rnold.
J. c., 'The Pneumatic GallgingTechnique in its Application to Dimen~ional
Measurement', J. 1. Prod.
36, No.2, 1957.
HABELL, K. J. and Cox, A., Engineering Opric3, Pitman.
CHAPN.lAN,
EVANS,
HEMSLEY,
project again
are widdy
increclscR.
MERRITT,
H.
Gears, Pitman.
NICI(OLS,
PARKINSON,
POLLARD,
228
ll1
1933.
n., Macmillan.
229
il;-ietrology for
Pra:cticd Engineering lvletro!ogy, Pitman .
K. 'N.
.-~- 'Notes on
Organisation and Dimea:sional Control' (Admiralty),
fLM.S.O.
'Gauge [vTaking and Ivlcasming'
.L.), E.NLS.O.
'Gauging and
Screw Threads' (N.P.L.),H.M.S.O.
d.;'.,,:u;>, 'N., A{ccfu:mism and Kin(!malics r~f
Longmans, Green.
C. J.,
and Future Development', J. 1. Prod.
37, ~'.jo. 7, 1953.
T!Mr~i;O. 6, 'l~~~~nt Development 1n Spm and HdicaI Gears', J. I. Prod. E., 39,
H.
~restfng,
and
Odhams,.
PriSln -::.u:d Lells
ACCURACY;
37
of detennin~tion; 1. 7-9
grades of; 32. 33
Addendum; 131
modification; 135-6, 141, 147
Airy, Sir G. E.; 4
Aligmncnt:
principle of; 35, 64
te!escope; 82
tests; 116-23
installation level; ltG-I?
spindlcl; 118-19
straightnes,; ami flatness; 119-23
Allowance; 9S, 96
Ambient conditions; 3
Angle, involute funcUoll of; 129
Angle dckkor; 79-82, 88
Angular measmcmcnt and circular division; 67-92
circlihr division; 838
optical instt:!ments for; 76-83
precision levd; 756
sine bar; 67-70
squarcness; 88-92
taper gaubrcs; 70-5
Assembly, selective; 94
Auto-collimator; 789, 85, 86-7, 91, 119, 120-3,
i.'l4-5, 220
230
CAUllRATlO>i :
of precision polygon; 85
of slip gauges; 42
Carbon steel. for gauges; i [3-14
Centre line aVerage fCL.A.) valu.;; 1357
Circular division; 67, 83~8
Circular pitch; 131, 133
Clearance; 13!
Cleo.ro-nee fit; 95
Clinometer; 76
Collimariorr; 76, 7S, 79, 82, 119
Colours) wavekngths of; 14
. Combir!ation Dale gl1uges; 79, 80-2
Comparators; 42
Comparators, design of; 43.62
elcctdcal; 60-1
fiuid (!isplaccment; 61-2
hjgh-magnlfi(:,o.tlon; 43-6
rnecilanic<!J; 47-50
mcc;lallic'-l1-optical; 50-4
pneumatio; 55-60
Comprcs~ed limit control charts; 2089
Conlicbllce Emits; 8
Con"tant chGd system; 14(3
Constraint; 39-40
Control charts; 196209
average; 198201, 2Co, 107
compressed limi:: 208-9
d:l,;;' sheet; 206
fruction defective; 2045, 208-9
rangc;201,204.206,207
Control chart limits:
modil1ed, ior sample average; 203
for sampic range; 202
for range; 202
231
Index
Flow~ve!ocit:!
DAVID
EfJEN-ROLT
Ciinvar; 113
Emis310n Theory; !3
cr.";)r~; in mcasurc:n..::nt; 1.. 11
'lligmncot; 2
'unbicnt conditi()ns; 3
antlln1elic; 4
(.:~l,:ullnous or cltastrophic; 2
CGm;lOUnd; 6-7
,;osinc= 3
~ia!)tjG deformation; 4
measnring; 6
reading; 2, 5
.sc~L~e: 5
Expansion, coefficient of; 3
gauges; 101-2
tolerances on; 99-100, 104, [05,
Gauge bloc:{s, comparison elf; 41-3
Gauges:
accuracy; 33
materials for; 112-14
smf~tcc finish for; 114
verific:ltl0Tl of; 33-4
GZ.Ugiflg tolerances, standard:>; 93-H4
hole dcpt~l gauges; 108
h~rgc diaxntters; 108-12
limit gauges; 98-103, 104-8
screw thn;ads; 103-5
faylor's theory; 1GO-2, 103
Gtw:;sian distribution; 193
Gear measurement; lJ6-55
addendum modification; 135.6
aHowablc errors; i55
E~n0r(11 tests; 136..49
individual dements; 136,137, 139, 149-55
b'/olm(" [oml; 153-5
pitch; 149-53
rclIers. mCE.surement over; 136, 133-9
roail}:; test~-;; IJG t 137-9
GAP
n::'(l:mreInent of;
F"'rlG UE life; 176
Fiw, sYSt1;IT1S of; 94-'7
Brifish Standard Systcln; 97
hck Gasi::); 96
(..iewall Systcnl; 97
than hosis; 96
F!~lnk angle:
error; 103, 105
ctft:(:t of error; 16~~~f\ 174, 175
!)lCa~u:t;lllcnt; lCO, 162-4
T-<'l;i~ness testing; 1[9~2J
136~55
Indexing devices:
calibrating of; 85-6
and cumulativ:! pitch errors; 151-2
Inertia; 38
Installation level; 116-17
Instruments, design and opera don of; 34-8
accuracy; 37
alignment; 35-6
inertia of moving parts; 33
sensitivity; 38
variance; 38
Interference:
fit; 95
microscope; 187-90
InterfeT(\m~tcrs; 21-30
N.P,L.; 21-4
Pitter-N.P.L; 24-30
Interferometry; 13, 15-18
in flatness testing; 18-21
Invar; 113
Involute curve: 128
Involute form; 153-5
Involute [\joetton; 129
JOHANSSON 'Mikrokator'; 47-8
l(n.;EMATICS; 34, 38-11
constraint; 39-40
freedom, one dcgree of; 40-1
LIlNGTH bars; 33-4, 62-3
Length standards:
angles derived [relm; 67-70
evolution; 12
light waves as; 13-30
Light, natlU'e of; i3-i4
Light waves as length standards; 13-30
application, field of; 13
flatness tes~inf;; 18-21
interferometry; iJ, 15-13
monochromatic rays; 14-17
nature of light; 13-14
Limit ganges; 98.. j()3, 104-8
heat treatment of; 113
hole depth; 10:;
materials for; 112-14
screw threads; !O3-5
taper; 105-8
Taylor's theory; 100-2, 103
Te.bo; 102-3
Limits and fits, systems of: 94-7
Line of action; 129
Linear measurement; 31-66
comparators, design of; 43-62
MAAG
System; 97
l'lewro!1, Sir 1&~'\G; 13
Noise problem. tears and; 127
N.P.L. flatness interferometer; 21-4
N.P.L. projector; 160-2
NEWALL
50
233
Index
P:.~S,s.~ vJty;
Screw threads ~
38, 58
prism; 92
:,~augc; ~C9~lO
Pi.t~;~:,
;Ga,r,
rnea')ureG1cttt ()f;
149-53
1:'l:~',ms:
lr::_~;,::,.:'ure,
theory; 2!Q-.t2
F,,)csS
:,t:"";:ld(~ni
deviation;
19'~w5
Sc;unn~res3, :"11e:.lSUrcmcnt
II1ctItcds;
~06,
'ZD7
coc:troi
(E..P.7.~;
83-~)O
tes~;;
for;
::::D9... t~)
119~23
i.m:Lrumcnt'); 130.. 1
'If:sts;
~(njar,'~
r'~Gu . ;rm\~::.s;
.1.
~Yl-9
78-lIO
i~i.~!~1lJC Lifv;
fo t'
ga~Jgcs; 11'~
nl~.;::'.ning
::L\~,~p:';~~~Cllnspc:ci,{on; 2D9~19
;~;dJk'8h:l::'r,
S-C:r',~" J:
'dG:lr;
Surr;,~cc
Df; 177-30
177
i1n[;':h
n~Ca:)urer(;cnt~
.; 30-90
mel:lod,; 180-3
replica; 190
TALLY
Thury, M.; 83
Timoshcnko, S,; 4
Tolt:ranec; 95, 96, Sec also gauging tolerances
Tomlinson surface meter; 181, 182, 187
Tooth thickness, measurement of; 132, 1;;9, 147
base tangent method; 143-7
constant chord; 141-3
at pitch linc; i 40-l
TraMil!on lit; 95
, 'Turn round method' of sqlw,cness test; 123
VA1UANCE; 378
Waviness; 178-8 I
Wcar; 177
aHowanc~; 98~ 105
'Wringing'; 3], 81