Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
the points describe paths which are a geometrical property of the system,
inflection circle, dealt with in this section, offers a direct, purely geo-
Figure 10-1 shows the fixed and moving polodes, TT/ and wm, which
P and points away from 517 (see also Fig. 3-22); pt is the pole tangent, the
counterclockwise from pn, and the distance PS. In the following, dis-
positive or negative, the positive sense being from P to the moving point.
217
218
moves to O'm, S moves to S', and the point of contact of the polodes moves
to the new position P'. P' is thus the position of the instant velocity
pole after the time interval dr. The elemental displacement P > P' is
denoted by dl.
Rays
P-S=rs ; 0S-S= ps
(a)
FIG. 10-1
cos 6, =
dlfm
dl
(a)
(c)
Since d<pm is the rotation of the plane, and dl the consequent shift of the
velocity pole, dtpm/dl is the same for all points on m and is always positive.
219
dl
(d)
or
and
(1
I r\ j
cos 6, = -T-
cos 08 = -T
_1
Cos , ~ di
(10-1)
motion of a plane are the position of the velocity pole and the angular
Rays
Qf
FIG. 10-2
from a given ray, and the length rf, are not known. The objective of the
The Inflection Circle. Equation (10-1) shows that points such as Si,
82, etc., on a given ray describe paths which differ in shape in the vicinity
that there exists on the ray a point I, which passes, at the given instant,
through an inflection of its path. For such a point, the radius of curva-
ture of the path and the distance Oi. > P are both infinite. Equation
and
cos 6, = -j
di
> /.
i cos 6,
(10-2)
220
which shows that on any given ray there really exists one, and only one,
inflection point and that the locus of all inflection points in a given
Figure 10-2 shows the inflection circle for the kinematic system of
Fig. 10-1.
By Eq. (3-17),
However,
ap, = a>m2c?,ip.
= ,2d,.(cos 0.)i. - (P -
Rays
apm Pn
FIG. 10-3
(10-2)
The result obtained shows that d,-, as defined by Eq. (3-15), is identical
I.
(10-3)
formed as follows.
221
and
Hence
Os
(O&
P = (Oa-*S) - (P
I, = (p _> S) - (I, -
.-*+- l
S)
and
S)(I,-
S (P-
') = (P -
8)-
(10-4)
which is the second form of the Euler-Savary equation. Since (P > <S)2
negative, which means that 0s and / are always on the same side of S.
ToOD
Eaye
'Ray h
FIG. 10-4
For points S inside the inflection circle, I8*S is negative and, con-
Hence such points describe paths which are convex when viewed from P,
while the paths of points outside the inflection circle are concave. For
If the position of the inflection point on a given ray is known, then the
location of the center of curvature of the path of any point on the ray may
222
(line ||). Through P draw a line parallel to line |, intersecting the auxil-
iary line at Z. Through Z draw a line parallel to line ||, and obtain Os at
I.
Oa
FIG. 10-5
of similar triangles.
I.>8 Y->S
OS>S Z>S
Hence
Figure 10-56 shows the same construction for a different relative dis-
if (a) one pair of conjugate points and the corresponding ray angle are
given, or (6) two pairs of conjugate points, on different rays, are specified.
Case a. The most usual example of this case is illustrated in Fig. 10-6.
Given are the conjugate points Om and Of for which the ray angle is zero.
(Of is the center of curvature of the path of Om. See Sec. 3-13.)
223
1,1 1
Of
and
or
(10-5)
(10-5o)
where the plus applies if the two polodes are convex, and the minus if one
polode is convex and the other concave. The result is thus identical with
FIG. 10-6
(Fig. 10-17).
cos
or
cos q =
cos '8
The above equation is solved for 0,, and d,- is then found by Eq. (10-1),
applied to Q.
determining separately the inflection points /, and I,, using the con-
struction of Fig. 10-5 in reverse order, and then erecting normals through
these points to the respective rays. The two normals intersect at /, the
only in that the auxiliary line and point Z are no longer arbitrary. Here
the auxiliary line coincides with the line through the moving points Q and
this particular combination of line and point, the two inflection points Iq
FIG. 10-7
the line connecting the relative instant centers Pis and P24 of the physi-
cally separate links is the collineation axis of driving and driven links; it is
Bobillier's Theorem. This theorem states that the angle between the
pole tangent and one ray is equal to the angle between the other ray and
the collineation axis of the two rays, both angles being measured in the
same sense. The usefulness of the theorem in analysis and synthesis will
The correctness of the theorem may be verified with the aid of Fig. 10-7.
, _ dl _ dl dr
d<p dr dtp
or di = (10-6)
wm
In Eq. (10-6), wm is the angular speed of the plane m, and vp the speed
with which the instant center of rotation changes its position along the
fore, of links which are pivoted to the frame at appropriate points, and
rotate with the rays, and of blocks which are free to slide along these links,
plane if the velocity pole, one pair of conjugate points (S and 0s), and
Solution. The original system is shown in Fig. 10-8o, and the auxiliary
linkage in Fig. 10-86. The latter consists of the bar s, pivoted to the
226
q guided along the pole tangent pt and r free to slide along s. The point
VOm = Um(POm)
Fig. 10-8
from
Vp = Vx. + VPX,
227
Figure 10-8c shows the construction for the particular case of specified
P on 5 and 6)
X2 on 2 \
Xt on 4 \
FIG. 10-9
on bar 2, and block 6 free to slide on bar 4. The points on bars 2 and 4,
228
The input velocities are v^ = w3 X (Pis > A) and VB = wa X (Pis > 5);
and
= 0.5rad/sec
(assumed)
(a)
0.5rad/sec
FIG. 10-10
from the fact that any point on link 3 describes a conchoidal path.)
rotating link 4, pivoted to the frame at 04, and the two blocks 5 and 6,
and block 6 along slider 2. The coincident points are again denoted by
X2 (on 2) and Xi (on 4). Since links 3 and 4 form a sliding pair, t>4 = u3.
For this reason vx, = ujX (04 > Xi). Since link 2 is in rectilinear
Vp = vx, + Vpx,
the mechanism shown in Fig. 10-1 la. The mechanism consists of the
crank 2, the bar 3 hinged to the crank at A, and the stationary surface 1
are the bar 2, pivoted to the frame at 02, the bar 4, pivoted to the frame
which block 5 is free to slide along bar 2, and block 6 along bar 4. The
points coincident with P are X2 (on 2) and X4 (on 4). Since the aux-
Vp = vx, + Vpx,
two stages. The first state, shown in Fig. 10-126, involves the determina-
stage consists of the eccentrically pivoted bars 3 and 4 and the blocks 7
bar 3, and block 8 on bar 4. The points on bars 3 and 4, coincident with
Vr = Vx, + Vfix,
auxiliary linkage for this stage consists of the bar 10, pivoted to the
frame at On, = 02 and passing through R, the bar 6, pivoted to the frame
FIG. 10-11
rad/sec, ccw
P on 5 and 6
Inflection circle
X4 on 4, X2 on 2
231
R on 9
X,0 on 10
FIG. 10-12
P on 11 and 12
Y6 on 6, YIO on 10
= 0.25rad/sec (assumed)
232
with block 11 free to slide on bar 10 and block 12 on bar 6. The point
bars 10 and 6, coincident with P, are shown as YIO and y6, respectively.
VP = Vr, + VPY
Vp = Vr10
and
10-5. The Inflection Circle for the Relative Motion of Two Moving Planes
Thus far the discussion centered around the construction of the inflec-
Pn
(m rel. /)
FIG. 10-13
and motion of / relative to m) are polar images of each other, a fact which
Solution. Figure 10-146 shows the auxiliary linkage for the inversion in
which the cam is held stationary. The linkage is basically the same as the
B)
and
Vp = Vjf, +
VP = Vx, +
233
Since the inflection circle provides the means for a purely geometrical
methods of Chap. 6.
w32=-lrad/sec
(assumed)
(a)
of 3rel. 2
B(03) J ; ->
FIG. 10-14
stant speed. Required are the angular acceleration of link 4 and the
acceleration of C.
234
with the aid of the inflection circle. After the path curvature of B has
(a)
(b)
,>., (PC)
v*s
FIG. 10-15
With link 4 curved, as shown in Fig. 10-16a, the problem may be solved
either with the aid of the inflection circle or by the substitution of the
235
FIG. 10-16
236
and 0s of the paths of Q and iS, respectively, are made to traverse the
prescribed loci e0 and es- (The locus of the center of curvature of a given
(c)
FIG. 10-17
Pf
(ap)n
(a)
In order to evaluate Eq. (a), a.p will be computed with the aid of the
given ray, and at right angles to it, were laid down in Sec. 10-1. Since,
in Fig. 10-17, which forms the basis of the subsequent derivations, all
relevant directed lines are positive, the diacritical overbar will be omitted
parts, viz., the tangential component of &x, and the Coriolis component
of P relative to X,:
Ps
wi uw vm \MV <.<i(/
u y. [ra d(Os -> X,) _ (Os > X,)rs dps + Os>X, dri]
(d)
The three derivatives in Eq. (d) may be expressed in terms of the given
dl dl ~ dl
dps = pfada
[dps is positive because (O's ->-S') > (Os -* S).] From Fig. 10-16,
-^ = - sin 6. {/)
[drs is negative because (P' -> S') < (P -> S).] With
|_Ps2ps rs ps
(opjrjc0r = 2upsvpx.
VPX. = WP sin 0,
n O
Ps ps
Equations (g") and (h) may now be substituted into (6), giving
ps Ps rs PS
or
tan 5 =
,, Ks tan 5
= wm2d,--
/ N ,j KQ cos 9, Xs cos
P/ KQcos6, - Kscos6q
where the coefficients KQ and X's are defined by Eqs. (i) and ( j) .
means that Of and the inflection circle are on opposite sides of the pole
tangent.
general case. Consider now the special case where the path curvatures of
from Eq. (e) that, to satisfy this condition, the evolutes er and ez must
pz PY
"
P/ =
-ia
2di (10-8)
cance will be made clear in Sec. 10-10) and its derivatives, in the following
manner.
By definition,
'
*A. S. Hall, Jr., Inflection Circle and Polode Curvature, Trans. Fifth Con/, on
P(
cos2 6y
and
p/
OY > P
cos
Consequently,
d2(R ,. _ , d(R .
(R - 1
(R
241
If, in Eq. (r), cos 6U is expressed in terms of tan 6y, and the latter in
terms of d(R/dij and (R by Eq. (q), then the following expression results:
p/ =
- I)2
- 1)3[(R2((R - 1) - 2(CR')2
where
and
(R" =
P) _,,_d2(0r->P)
r = 0,
r' =
r=
coordinates.
The polodes associated with the motion of the coupler 3 relative to the
frame 1 are ir/ = iri and 7rm = 7r3. It is obvious from Fig. 10-19a that
242
Hall's equation may be used directly for the computation of the curvature
Of T1l
2d,
P1 =
[^ -]a + + [-#-](. -
fc)
(10*10)
(10-11)
O^Ob
Fig. 10-20
Solution. Here P -> = 1.3 in., 04 -> B = -2.3 in., P -> A =1.7 in.,
k = 0.755
1 = 2.45 in.
and / = 1:
Ps(-2.45)
4.22 =
p3 + 2.45
and
p3 = 1.55 in.
The polodes in the vicinity of P13, with the centers of curvature denoted
polode 7T2, associated with the motion of the output link 4 relative to the
input link 2:
sin (03
(10-12)
, , 3 1 ,in 1Q,
where k = - ^ - ^4 (10-13)
It should be noted that, since the discussion is concerned with the motion
of link 4 relative to link 2, the positive sense along bars 1 and 3 is the
this case. Comparison of Figs. 10-18 and 10-19& shows that dtp2 = dy
= giZLg = = 21 (10-14)
_-
Thus, in connection with the relative motion of the output and input
d(R a4
Equation (10-14) relates the position of the relative instant center P-u
input link, the tangent of the angle 61 between the pole normal and the
sion which makes it possible to determine the time rate of change of the
FIG. 10-21
function generation.
angle formed by link 1 and the pole tangent equal the angle formed by the
61 = 104.5
and
03 = 123
(P -> S)* m (Ob -> S)(I. -> S) - (Os -> 5) [(P -> 5) - d, cos 0.]
Hence (P24 -> 04)2 = (02 -> 04)[(P24 -> 04) - d. cos O1]
the pole tangent opposite to the inflection circle.) By Eq. (10-5), with
m = 4 and / = 2,
fit" = 0.72
^P = 5.76 rad/sec8
ar
a4 = 0.)
Proof. Equation (10-16) states that (R' = ((R2 (R) tan 0i. Hence,
for the conditions stipulated, the pole normal coincides with the fixed
in Four-link Mechanisms, Trans. ASME, vol. 78, no. 4, pp. 779-787, 1956.
tions, Trans. ASME, vol. 79, no. 6, pp. 1305-1312, 1957, with discussion by A. S.
Hall, Jr.
246
link, the pole tangent is perpendicular to the fixed link, and by Bobillier's
mechanisms is constructed on the same fixed base line 02 > 04, whose
members have the same instantaneous velocity ratio (R, the same first
derivative (ft' = d(R/d<p2, and the same second derivative (R" = d2&./d<pz",
then the instant centers PIS of the couplers lie on a circle which is centered
on the fixed link and passes through the common pole P24.
FIG. 10-22
By inspection,
Hence
= AP24Pi304 -
(P24
and
D . D (P24-^04)(PZ4^B) sin(0
3) sin
sn
respect to link 2 and with k expressed in terms of the angles t?i and #3,
reads
'=
0Q Jf, . sin
02^04
(d)
Substitution of Eqs. (c) and (d) into (6) and rearrangement of terms yields
D 3(P24 > 04)rf, sin t?i cos #3 sin i?3 sin (?3 t?i)
\ P2/
sn #, cos t?i sn
If now the right-hand side of Eq. (e) is substituted for P24 > B in Eq. (a),
\ P2/
By Eq. (10-14),
= (1 -(R)(P,4->00 (0)
With the aid of Eqs. (<?) and (A), Eq. (/) is transformed to
P , P 3(R(02 -> 04
L \ P2/ J
Equations (10-14) to (10-16) show that mechanisms which have the same
instantaneous values of (R, Si', and (R" have, ipso facto, the same inflection
circle and the same polode curvature. Hence, for such a family of
mechanisms,
COS #3
construction:
COS ui I dc
With these values, Eq. (10-21) yields (R" = 0.72, which agrees with the
Let Y denote a point on the circling-point curve, and apply the Euler-
(As in Sec. 10-7, the overbar indicating directed line elements will be
dpy
dl
-# = sm 0, (c)
d0 _ dr\ d
dl dl dl
d| = J.
dZ p/
(d)
(e)
, d; ryd<pm ry 1 ,,,.
dZ py ^ py d,-
_j. d0 ry 1 1 , >
Hence -^ = T (#)
at py a,- p/
Finally, elimination of py between Eqs. (h) and (a) yields the equation of
sin 20
K2-t>
sin 20
M sin 6y + JV cos 0B
(10-22)
inflection circle with the phase of the moving body; that is, d(d,)/dl. The
computation becomes feasible if, as shown in Fig. 10-17, the path curva-
250
displacement velocity:
tan 5
Kq cos 0 Ks cos 0,
Ks sin 0, Kq sin 0,
(ap)t =
(ap)n
tan 5
um2di
Ks sin 0, Kq sin 0.
However,
d(d<)
dl m dl
d'l
0)
(*)
(0
(m)
(10-23)
Fig. 10-23
6- = arctan
B)
(10-24)
The asymptote itself does not pass through the instant center. Its offset
n is determined as follows:
M sin Oy + N cos 6y
value is found to be
2 cos 6y .
Y Sln 0" =
instant center P
stituted into Eq. (10-22), two independent equations are obtained, from
Solution. The positions of the pole tangent and pole normal are found
mined by measurement:
With these constants, points on the curve are located as shown in detail
sin 250
rc =
= 5.44 in.
Figure 10-24 shows clearly the degree to which the actual path of a
the phase considered, by the osculation circle. For the point C, the two
Osculation
circle at C
Fia. 10-24
mechanism as shown, the output link 6 would remain sensibly at rest dur-
The location of the hinge between links 5 and 6 which coincides with,
inflection circle:
(P1. -> Ay = (0, -> A)(h -> A) = (0, -> A)[(PU -> A) - di cos .]
0O = 145.5
radius of curvature:
0c = 125
CHAPTER 11
INTRODUCTION TO SYNTHESIS;
GRAPHICAL METHODS
function generation
analytical, for the solution of these problems. The accent will be on the
graphical approach, because it is the more direct of the two and yields
Since in most cases the solution will be based on the four-bar mecha-
The general symbols used in the following discussion are listed below
and, with the exception of the rotocenters, are defined by Fig. 11-1.
254
255
four-bar mechanism:
(a)
FIG. 11-1
range. It should be noted that the forward and return strokes of s do not
oscillate.
All three types can be obtained as inversions of the same linkage, pro-
vided the sum of the lengths of the shortest and longest links is less than
256
their full range. If a four-bar linkage does not satisfy Grashof's criterion,
FIG. 11-2
257
who first suggested the use of this parameter as a quality index for the
FIG. 11-3
258
Line r3
Line rt
Line r3
Line rj
Fig. 11-4
Figure 11-3 shows that n reaches its extreme values in a motion cycle
when the driving link q is collinear with the fixed link p. The corre-
259
synthesis of such a linkage for prescribed values of Mmin and ^13 is shown in
Fig. 11-4.
s = O,B, is drawn in its phases Si and s3, and the lines TI and r3 are drawn
angle. Varying coupler lengths rx = BiAix = B$A3X are then tried until
aline A3xj4ix = A 3Ai is obtained which passes through O,. The required
FIG. 11-5
in this manner, with the line drawn through B3 at 45 to the ray r3, clearly
driven link. It will be observed that the crank angle <24, corresponding
to the forward stroke of the rocker, is larger than the return angle <t>n ',
that is, <24 > 180. Thus the crank-rocker mechanism is inherently a
quick-return motion.
Hence
and
0gB2 = r + q
and 05Z?4 = r q
+ OgBi = 2r
- OqB4 = 2q
(a)
(6)
1 A. S. Hall, Mechanism Proportions, Machine Design, vol. 20, no. 2, pp. 111-115,
1948.
260
given degree of quick-return action, i.e., for prescribed values of <24 and
^24-
The driven link s is drawn in its extreme positions $2 and s4. The
(6)
(c)
FIG. 11-6
Since the position of Oq on the "pivot circle" is arbitrary, and thus the
1. The mechanism should have the largest possible value of /umin. The
3. A prescribed link ratio q/s. Since, by Eq. (6), OqB2 OqBt = 2q,
4. A prescribed link ratio r/s. Since, by Eq. (a), OqB2 + OqBt = 2r,
3 O,MB2 = 2fa
If (024 - 180) < fai, O. is inside pivot circle and fa < (90 - ^ J
If (024 - 180) > ^24, O, is outside pivot circle and fa > (90 - ^ J
As shown in Fig. 11-7, the pivot circle passes through O and any other
point on it is a suitable location for the pivot Oq. Since, in this case, the
(^-180) = *2
262
Fig. ll-8a.
and
Hence
and
q + e = p cos <2
q e = p cos (360
2e = p(cos fa cos
= p cos
(c)
Figure 11-86 shows that the special conditions ^24 = <24 180 and
Figure 11-9 shows a body r, defined by the line AB, in two distinct,
263
where on the normal bisectors [n^]i2 and [na]u of AiAz and BiB2, respec-
FIG. 11-9
point, denoted by [Rr]u, is the rotocenter of the body, or plane, r for the
bility of confusion, the suffix r will be dropped, and the rotocenter will
the triangles AiRuBi and A2R^B2 are congruent, the amount of rotation
about RII is equal to the actual angular displacement [tpr]u of the body.
two given positions, subtend equal angles at the rotocenter [Rr]u = R^;
the couplers r of all such mechanisms also subtend equal angles at Ri2:
265
links
All such problems are solved with the aid of one or the other of the
the positions of the line AB, the locations of Oq and O,, and the rotations
[cpji2 and [<p,]u of the guiding links are specified. Required are the loca-
with [nB]i2. The normal bisectors [nc]i2 and lnD]u are obtained by
intersection of the line gi and the ray ci, drawn, respectively, through Oq
266
rotocenter.)
since the hinge D, and with it link m, would remain virtually at rest
However, since points other than A and B may be used equally well
Center Cl
FIG. 11-12
Center DI
kept stationary in one of the three phases, for instance, AiBi, and the
located at the centers of the circles drawn, respectively, through Oq, 011,
The Angles of the Rotocenter Triangle. In Fig. 11-136, the body has
coincides with R^. Thus the extended line RuRi3 coincides with [nc]i2
and [ttc]is, and the extended line RuR^ coincides with [nc]is and [rio^s,
and it becomes obvious that the angles of the rotocenter triangle are
tp'13 = 280, about the same rotocenter. In the rotocenter triangle <pn/2
is the rotation of side 1 into side 3. Obviously, there are two possibilities:
1 -> 3 of 140.
the point A. Given the rotocenter triangle and the position of the point
in one phase, the correlated positions in the other two phases may be
268
Side 3
Fig. 11-13
269
side of the rotocenter triangle, for example, AI about side 1, yielding the
cardinal point Ac. The reflections of Ac about the other two sides of
FIG. 11-14
the point H , whose cardinal point Hc coincides with the orthocenter (the
circumscribed circle.
e+X+
= 90
< YRuRu = n
Consequently,
< RuRisHi = n
< Fflisff23 = X
(o)
Image circle 3
Image circle 1
FIG. 11-15
Consequently,
=X
RuH3Z = 90 - X =
Consequently,
271
Thus, since the line segment RuH^ subtends equal angles p, at H2,
RM, H3, and Ru, it follows that all six points lie on one and the same circle.
That the point H is unique in this respect may be seen from Fig. 11-156.
the circumscribed circle is the image circle 1, i.e., the reflection of the
Image circle 3
Circumcircle
FIG. 11-16
and image circle 3 is the locus of all cardinal points Pc whose correlated
image circles have only one point in common, viz., the orthocenter Hc.
correlated positions of a point are collinear, they are also located on the
corresponding image circles, and the line containing them passes through
He.
shows that
<P23
to R^RM,
Thus, since R^Ru subtends equal angles at Ai and Hc, it follows that
AI lies on the same circle as Ri2, Ru, and Hc, that is, image circle 1.
A 3 on image circle 3.
At Ru,
Since Hc, Ru, Ru, and AI lie on the same circle, R^Ai subtends equal
AsHcR^ = v
Hence, at Hc,
= r,
2* + + a = 180
and ^3 pass, respectively, through HI, H2, and H3, and S is located on the
[WB]^, etc., which have been omitted so as not to clutter up the drawing
The lines 613 and ?>23 bisect, respectively, the angle liSls = <pu and the
13 -"23
FIG. 11-17
"~
which shows that S lies on the same circle as Ru, Ru, and #23.
It has been shown previously that HI, H2, and H3 are correlated posi-
tions of a particular point. Therefore the directions HiS, HzS, and H3S
with Zi, I2, and l3, respectively, follows from the fact that, as shown in
Fig. 11-176, the rotocenter triangle and a given point S on the circum-
tion of Zi, Z2, and Z8 would proceed as follows. The rotocenter triangle
yields the magnitude and sense of the rotation ^13. Since the ray Ri3S
ing to the three crankpin positions A i, A2, and A3, lie in a horizontal line.
(As a check on the accuracy of the working, Hc must coincide with the
Hc intersects the image circles at Ci, Ct, and C3, respectively. (A further
check on the accuracy of the result is the condition that the distances
Figure 11-18 shows also the possibility of utilizing the motion of C for
the design of a dwell mechanism. During the rotation of the crank from
three phases of the mechanism, denned by Ai, A2, and A3, M occupies
angle, and the cam surface will be made parallel to a line whose correlated
Fig. 11-18
/Image circle 3
Circumscribed circle
276
usual manner, with the aid of the correlated positions of A and of a suit-
image circles. Points Hi, H2, and H3 are located, respectively, at the
circle, and the cam surface in phase 1 is made parallel to HiM. (An
excellent check on the accuracy of the working is the condition that O,,
line AB, must lie on the circumcircle and that the extended lines AiBi,
AzBz, and A3B3 must also pass through Hi, HI, and H3, respectively.)
tions by means of bar linkages is rather complex. Its solution entails the
lie on circular arcs. Such points (of which there is an infinite number)
may then be used as moving hinges, and the centers of the corresponding
circles as fixed pivots. The locus of the fixed centers is known as Bur-
defined by AB, in four distinct positions, together with the six associated
rotocenters ^12, RM, RU, RM, Rzt, and Rat, which are located at the
construction lines for R13 and R-u have not been shown because they
bisectors [HA] pass through the point OA. The radius 0^A moves from
and
2 22
Figure 11-20 shows that the angle included by the bisectors [n^]i4 and
278
and that the angle included by the bisectors [uaJu and [nA]23,
jr p n p - 5u + 5m -
* K12UAK23 = e
Thus the rotocenter distances #14^34 and R12Rn subtend equal angles
at the point Oa- It may be shown, in a similar manner, that other pairs
Burmester's
curve
512 + ^23
2 &l
[a]u /-
[a]23
5l4 - Sm _
Burmester's
curve
Fig. 11-20
as bases. The simple construction shown in Fig. 11-21 yields two points
which obviously satisfy these conditions. The radii of the two circles
and their center distances may be determined either graphically or, alter-
natively, by adjusting the slide rule to the ratio of the base distances
yields Burmester's curve. The curve passes through the six rotocenters
279
Rotocenter distances
Common
digits
Subtended
angle
Point of
intersection
1 and 2
812
Su
1 and 3
Sn
Sis
1 and 4
5l4
S14
2 and 3
823
23
2 and 4
624
(824
3 and 4
34
S,4
FIG. 11-21
280
R, the reason for which will become clear later, and through the six points
tion for a fixed pivot O,, the corresponding moving point is located by
11-22 depicts the same situation as Fig. 11-20. O, has been placed arbi-
AzBzO,, A3BzO,, and AtBtO, to the base line AiBi. The coupler hinge
C, in its position Ci, is located at the center of the circle through O,,
O}1, O}11, and O}v. (In the present example, because O. has been selected
FIG. 11-22
When faced with the problem of having to design a linkage for the
oped in more detail, and other locations of the fixed pivots investigated.
281
through the collinear positions Bi, B2, B3, and B^ A second guiding
A2, A3, and At are easily located and the four rotocenters Ru, Ru, RM,
Burmester's curve
O,
FIG. 11-23
there are four distinct relative positions, O} = O,, O]1, O}11, and O}v.
between the displacements of the driving and driven links which is gov-
lent wear resistance of pivots and sliders, and absence of abrupt changes
factors appear to tell against their use as function generators, viz., their
function with mathematical exactitude over a finite range and the extreme
Thus even the simple four-bar linkage and its slider derivatives possess
ical limits of the input and output displacement ranges, only a limited
the two functions diverge, but the error is kept within permissible limits.
is closer than with precision points, but within a much shorter range.
V = /(*)
283
- fa (11-1)
Bar function generators may be of the open or the crossed type. The
Its characteristic feature is the fact that the input and output displace-
k$ always positive
equation is the expression linking the positions of the output and input
links.
22 = p2 + g2 2pq cos
and $ = 180 - a ft
Q sin <t>
---
T> j
But a = arcsin
,,
and 8 = arccos
Hence
s2 + 22 - r2
2sz
iono
= 180 - arcsin
arccos
cos <)*
2 + </2 + s2 r2
--
2pq cos
0,
2s(p2
cos
,,,_,.
(11-2)
v'
Two zero positions, <t>o and ^0 (or positions <t>i and fa, corresponding to
close points, the second derivative to three, etc.) However, since the
be assumed: both scale factors, one scale factor and one initial position,
one scale factor and one side ratio.) Should the resultant mechanism
and the designer must either be guided by past experience or simply rely
on good luck.
The following rules of thumb may serve as a rough guide for the start:
3. Maximum link ratio <3. (This limit ensures that the injurious
and input links may be reduced to the simpler problem of guiding a body
stationary, and the relative motion of the links is considered. The first
step consists in locating the positions, in the reference phase, of the roto-
centers associated with the relative motion. (In the following, since
distinct phases. The output link s has rotated through an angle ^12,
means of the rotation <f>u about Oq, as shown in Fig. 11-256. Because
It will be noted that the bisector of O,O}1 passes through Oq, and is
285
Bisector [5,5"]
Bisector [A 1A11]
Bisector [OqO"]
Fig. 11-25
at the intersection of the normal bisectors of 0,OJ1 and A1Au, and it will
286
in Fig. 11-25a.
Fig. 11-26
centers OqO, through the appropriate pivot points defines the relative
rotocenter for the rotation 2-3 in phase 2oi the mechanism. Conse-
Ri>(1) = #23, is then obtained by rotating the ray OqR23(2) through 4>u-
287
(o)
Note: R = R0I
-o A
Fig. 11-27
288
venient ones to choose are the two scale factors and the initial position
and #23#34. In Fig. 11-28, these centers have been referred to qt.
Burmester's curve
s relative to q
Fig. 11-28
for the hinge A in its phase A i, the choice of the initial position is arbi-
trary. With A1, Ai, A3, and Ai located on the crank circle, the construc-
follows. The slider is assumed to be stationary in its phase s1, and the
289
As> A111 = Zis, and At> AIV = ln; Bi is the center of the
B1 is center of circle
through A 1i A", A
FIG. 11-29
The most convenient choice is that of the two scale factors k^, and k+, of
the initial position fa, and of the link ratio p/q. (The actual dimensions
of the links will not affect the solution, except for the over-all size of the
mechanism.) The assumptions lead to the values of fa2, fa3, ^12, and
290
phase 1. B1 is the center of the circle through the three relative posi-
The basis of the method of reduction of point positions is the fact that
three.
Hence
-(^r+V1)
sin
0n.
= cos
0n.
sin
0n,n4
n+1
n+1
tan
0,
n+l
(11-3)
i2 = fc,x12 = 10(4 - 1) = 30 i3 = 60 0i 4 = 90
and
The design position is so chosen that one of the input angles, here 0i2,
is bisected by the line OqO and the corresponding link ratio p/q is
291
= cos 15 +
sin 15'
tan 184'
-, = 1.762
B! is center of circle
through A1,Aa,Am,ATf
FIG. 11-31
are to be obtained, then <t>i and the distribution of the precision points
must be so chosen that two input displacements <f>m,n are bisected by the
line OqO,. Thus the choice of the precision positions is no longer com-
pletely free. For this reason, only one design parameter, viz., fc$ may
be assumed at will.
involves the determination of a link ratio p/q such that the output-
(U-4)
ft,
Asymptote
denned by
and (R^ =
number of link ratios p/q between 1 and 3 is tried graphically, and the
corresponding angles A^, and A^0 are measured. Thus the approxi-
reiterative calculation.
Figure 11-326 shows a typical (R* versus p/q curve. It has the fol-
lowing characteristics:
3. Between these values, Gfy is slightly smaller than (R^; outside this
range it is larger.
A ?/ At Aw , \ll
(R* > (R* means that > -r-1 and consequently -~ > -r2-
that is, the function must be such that the slope of the "inner chord" is
mechanism.
solution. For this reason, a design is chosen in which the angles 4>n and
024 are bisected by the line of pivots OqOs, as shown in Fig. 11-33. The
is computed as follows:
^ . yu _ yJLZJH = 0.707
<pls = 6330'
*24 = 4454'
294
Hence
and
2/15
* = 918' i
3145' +u = 5412'
With 0, located and the output rotations known, the relative positions
Ob.O,
(0,0,) =p = 1.85g
Au and AIV form the second coincident pair, and A1u occupies the
Input
Desired output
Generated output
Error
<t>l,n
(.Hn)t
(*..)
W'l.rOd (^!,n)d
029'
105'
+36'
10
155'
239'
+44'
15
415'
449'
+34'
2230'
918'
918'
30
1553'
1526'
-27'
35
2053'
2024'
-29'
40
2614'
2555'
-19'
45
3145'
3145'
295
The exact and generated functions are compared in Fig. 11-34, and
Exact function
- - - - Generated function
10
'a
-i
60
40 50
4j>, degrees
(a)
40
90
10 20 ^--^
JL ^^"^ SO 60 ^
<t>, degrees
>^^_ 80^x^90
(b)
FIG. 11-34
error (about 36', or 1 per cent of the output range) be considered exces-
296
whole range. Either three or two design parameters are chosen arbi-
trarily, and the remaining ones are varied in a systematic manner until a
O.
(b) Overlay
FIG. 11-35
and the corresponding angular increments <f>m,n and \f/m,n are computed.
positions of the input crank, and with the assumed length of the coupler
297
4. The overlay is placed on the first layout and moved across it until
obtain the best fit, different overlay circles will be scanned in turn. It
FIG. 11-36
the frame and of the output link and the initial position of the input
reversing the sense of the overlay. It should be noted that the same
298
overlay for a given output function. However, if the range Aj/ of the
(z > n) which increase in the same ratio as the function intervals, then
sponding increments of the output range A^. Thus this type of overlay
has general validity and may be used in the solution of any problem.
(In order to obtain a best fit, different sectors of the overlay will be
tried in turn.)
Line
Angle
Line
Angle
Line
Angle
100'
13
2512'
25
10436'
206'
14
2849'
26
11629'
319'
15
3249'
27
12936'
440'
16
3714'
28
144 06'
609'
17
4207'
29
16006'
748'
18
4730'
30
17745'
936'
19
5327'
31
19715'
1136'
20
6000'
32
21847'
1349'
21
299
Table 11-3 contains full data for the construction of a general overlay,
based on the arbitrarily selected interval ratio (R,- = -\/2, which doubles
so that the smallest interval is at the lower end of the range, and the
60^
to
go a*
2J
(a)
jjj 1ncreas1ng
(b)
- decreasing
Fig. 11-38
total range, for subdivisions into 4, 6, and 8 spaces. The values are
the limits 1 < x < 10, with the range divided into six intervals.
Solution. The input range and the coupler-crank ratio are assumed.
Table 11-4 yields the following values for the output intervals (Ay = 1,
dy/dx decreasing):
2/1 = 0
2/2 = 0.2105
2/3 = 0.4011
2/4 = 0.5738
2/6 = 0.7302
X1
x3
Xi
Xi
1.000
1.624
2.519
3.748
5.373
xu = 0.624
X13 = 1.519
Z14 = 2.748
xu = 4.373
xi = 6.444
*i 2 = 410'
0i, = 1008'
*i< = 1820'
*u = 2910'
*u = 4258'
0,r = 6000'
dy.
Number of intervals, n
-4
increasing
Vu
>/u
Vu
dx
Per cent of Ay
21.42
45.06
71.17
100.00
decreasing
dx
Vm
1/62
1/61
dy .
Number of intervals,
n=6
increasing
2/u
2/u
yu
2/u
2/17
dx
Per cent of Ay
12.82
26.98
42.62
59.89
78.95
100.00
dy
decreasing
2/75
2/74
2/73
2/72
2/71
dx
Number of intervals,
n=8
increasing
dx
2/12
2/u
2/
2/15
2/1
2/17
2/18
2/19
Per cent of Ay
8.62
18.13
28.63
40.23
53.03
67.17
82.77
100.00
301
11-40, which shows its location in crossed and open mechanisms, makes it
Fio. 11-39
1. The scale factors fc^ and fc^ and the link ratio q/p are assumed.
(R
(H 1
where
d\f/ Jcj, dy
d<t> k$ dx
(11-6)
(11-7)
(Because < AfOqA+ = A<, the choice of Af also determines the position
302
of A+.) The coupler positions AfPi and A+Pn are drawn in, and the
coupler length A+B+ is adjusted until Bf and B+ lie on the same circular
(a)
ToP
value. If, as is usually the case at the first attempt, the two angles
unsuitable.
with some practice, takes only 10 to 15 min, this approach leads to the re-
sult more quickly than the overlay method with a "free" output scale.
303
the limits 1 < x < 10, with precision points at x = 1 and 10 and precision
and
Fig. 11-42
The pole distances are calculated by Eqs. (11-6) and (11-7), where
-*- -13.5^429
/C4, OX
Z4 = 0.456p, h = 0.369p
The best possible solution (but not a very satisfactory one) obtainable
with the chosen scale factors is shown in Fig. 11-42. The particulars
2 = 0.710 - = 1.028
304
has been suggested previously in Figs. 11-11 and 11-18. Other uses
Hrones's and Nelson's Motion Atlas contains some 7,000 coupler curves,
pattern.
the Atlas are scanned until a family of coupler curves is found whose
each coupler curve, a separate overlay of the given curve is then con-
structed, so that the major axes of both are equal, and their shapes are
error margin.
The Atlas is also useful as an aid in the design of dwell linkages. For
motion pattern of the rocker u, was developed in this manner. Its basic
the Atlas, aimed at finding a coupler curve which had two appropriately
spaced segments of finite length with the same curvature. The linkage
has the proportions q/p = and r/p = s/p = %, and as far as can be
essential requirements.
by keeping the coupler stationary in its phase TI and moving the frame.
positions, the arbitrary design parameters are so chosen that some rela-
Hannover, 1952.
306
point positions, the fixed pivot O, of the driven link is made to coincide
with one of the rotocenters of the coupler, for example, [Rr]u, as shown
(Because A is a point on q, A
and
and
! is center of circle
FIG. 11-44
length CA. The position of Oq on the base line is also arbitrary, but
once assumed, fixes the crank length OqA. The intermediate positions
of A(A2 and A3) are marked on the crank circle [(C2A2 = CiAi), etc.],
of the coincidence of O, with [Rr]u, 0,IV coincides with O} = O,, and the
307
[nc]u
In Fig. 11-46, 0B has been made to coincide with [Rr]n and [Rr]3t-
! is center of circle
through O.,0?, Of
FIG. 11-45
length CA, and 0, is found at the intersection of the base line with the
normal bisector [nA]n. A2 is marked on the crank circle, and the relative
location of O,, O? coincides with O] = O,, O}v coincides with 0J11, and
(o2), namely, direction of the base line (1 d.o.f.) and distance between
C and A (1 d.o.f.).
308
B! is center of circle
The only arbitrary choice is that of the direction of the base line.
After the rays oi, 02, and a3 have been drawn, a suitable length CA+ is
assumed. With this length, the points Af, A$, and A^ are located on
structed. If, as is normally the case with the first random attempt, 0+
does not fall on the chosen base line, the construction is reiterated with
of this locus with the base line yields the true position of Oq. Next, the
309
0.VI coincides with O} = O,, 0,v with 0.11, 0JV with 0.m, and Bi is located
! is center of circle
through O.,Olf,O
FIG. 11-47
line which would permit drawing a circle through four distinct relative
positions of O,.
310
For instance, it does not lend itself to the synthesis of coupler curves
variant of the method in which the moving hinge B, rather than the
FIG. 11-48
one fail to produce the desired result, the other could be successful.
Four-point Synthesis. Figure 11-48 shows four positions, Ci, C2, C3,
of the fixed hinge Oq. Because of the coincidence of fi1|4 with [Rr]u
the coupler positions BiAi and BzAz are symmetrical with respect to
the line Bi,tOq and are inclined to it at an angle Xu/2, where XH is the
A 2 on the crank circle, and the position of B2 is found with A2B2 = AiBi,t
and C2JB2 = CiBiA. The fixed pivot O, is located at the center of the
Although not used in the solution of the problem, the following kine-
which the coupler is kept stationary in the phase ri, the four relative
BnOq = Bi^O] and Bi,40Jv are symmetrical in relation to the line Bi,tAi
11-49, Bi,6 coincides with [.RJis, and Bs,4 with [/2r]s4. The point pairs
must be chosen so that, in the sequence in which the point positions are
The distance CB and the direction of the line Bi,6Oq are arbitrary, but
the position of Oq on this line is no longer so, as will be seen from the
following.
Bi,6Oq. O\ = Oq, 01,11, 0JV, and Ovq are the positions occupied by Oq in
the inversion in which the coupler is kept stationary in its phase rI.
Normally, with Oq chosen at random, these points will not satisfy the
essential requirement that they lie on one and the same circle. To
ensure this condition, the rays Si,60J and Bi,sO must be mirror images,
with respect to the line Bi.iAi, of the rays Bi,6O\ and Bi,6O\11; that is,
and
Nvt-u
Fig. 11-49
Thus, since the rays JBi,60g and B3itOq make equal angles with the corre-
which proves that the points Oq, B1i6 B3,4, and #15.34 (the point of
intersection of the normal bisectors [nc]is and [nc]n) lie on the periphery
of one circle. This circle, denoted by k ] 5.34, is the locus of all possible
O\ = Oq, 0J 07 and 0y
The crank length q and the distance AC are thus determined. Next,
by locating O, at the center of the circle through J5i,6 B2, and #3,4.
(CzAz = CiAi marks the position of AI on the crank circle; A 2,82 = AiBi
the arbitrary choice of the length CB. The coupler hinge B is made to
occupy the three double positions Bi,e = [Rr]u, #2,5 = [Rrlu, and
314
time requirements) with the positions qiq3 and qtq^ (which define the
["cl
FIG. 11-50
The theorem states that the curve generated by a given coupler point
either of which may prove to be more favorable than the original design
1 S. Roberts, On Three-bar Motion in Plane Space, Proc. London Math. Soc., no. 7,
pp. 14-23, 1876; R. S. Hartenberg and J. Denavit, The Fecund Four-bar, Trans.
Fifth Con/, on Mechanisms, pp. 194-206, 1958, Purdue University, Lafayette, Ind.
315
as follows.
On the left, links OgD and DCE, such that OqD and DC form a parallelo-
On the right, links O,F and CFG, such that O,F and FC form a parallelo-
Length
(C)
FIG. 11-51
The resultant 10-link mechanism has 13 lower pairs (two each at Oq, O,,
this is so, then the 10-link mechanism may be split up into three separate
four-bar mechanisms, as shown in Fig. 11-52, each of which will give the
The coordinates x and y of the point 0< = H, in Fig. 11-51, are cal-
culated as follows:
316
ccc
ccc
and similarly,
To ^2 at intersection
FIG. 11-52
and
Hence
Consequently,
and
6,6
x = - p cos a
tan
and
y = - p sin
0,0,0, = = tan a
(a)
(6)
(c)
317
Equations (6) and (c) not only prove that 0< is a fixed point, but also
Figure 11-53 shows the construction of the cognate mechanisms for the
special case in which the tracing point C is located on the line AB.
Except that the four triangles degenerate into the similar line segments
ABC, DCE, CFG, and OqO.O,, the construction is identical with that of
1. The instantaneous velocity poles Pr, Pri, and Pr, of the three
couplers (Fig. 11-52) and the tracing point C are collinear in the com-
tl = o>rj = w.
and
found that one of the cognate mechanisms vanishes. Thus there exist
only two slider-crank mechanisms which will produce the same coupler
value.
CHAPTER 12
INTRODUCTION TO SYNTHESIS;
ANALYTICAL METHODS
mental stage, and the result, obtained after lengthy and laborious
Squaring and addition of Eqs. (a) and (b) yields the relation
q s 2qs
reduces to <R( (R, cos <t> + 61, cos \p = cos (<t> 4d (12-2)
319
320
the first method, the generated and specified functions agree in a number
precision points and the nature of the function. (The lowest maximum
error is obtained with a spacing in which e reaches the same value between
each pair of precision points and between the limits of the range and the
adjacent precision points.) In the second method, the generated and ideal
FIG. 12-1
and
or
or
8<t> = k$ 8x
d$ = A> 8y
(12-3)
and y = 0
means of a four-bar mechanism, viz., the three side ratios, the two scale
factors, and the angles <0 and ^0 (or the initial positions fa and ^,).
between the generated and specified functions is seven, e.g., seven pre-
two precision positions, one with four and the other with three deriva-
five.
fa, fa n = 1, 2, 3, etc.
while the limits of the range will be denoted by the new symbols
<t>i, fa fa, fa
The initial point fa, fa may, but need not, coincide with the first precision
Three Precision Points. The chosen parameters are ft*, ft*, fa, and
follows:
Substitution of these values into Eq. (12-2) yields three equations from
Four Precision Points. The parameters assumed are ft*, ft*, and fa,
(R, (R, cos fa + 91, cos (fa + fa,1) = cos [(fa fa,1) - fa]
(R, (R, cos fa + (R9 cos (fa + fa,2) = cos [(fa fa.i) - fa]
(R( (R, cos 03 + (R9 cos (& + fa,3) = cos [(fa fa,3) fa]
(R, - (R, cos fa + (R, cos (fa + faA) = cos [(fa - fa,i) - fa]
Gradual elimination of the side ratios yields a cubic equation in tan fa,
of the form
original paper.) With fa and fa known, the side ratios are found as in the
three-point approximation.
(R, - (R, cos (fa + fa,l) + (R, cos (fa + faA) = cos [(fa,1 - faA) + fa - fa]
(Rt - (R, cos (fa + fa,i) + (R, cos (fa + fa,i) = cos [(fa,i - fa,i) + fa- fa]
(Rt - (R cos (fa + fa,3) + (R, cos (fa + fa,3) = cos [(fa,3 - fa,3) + fa- fa]
(R( - (R. cos (fa + <fc,) + (R, cos (fa + ^,-,4) = cos [(faA - faA) + fa - fa]
(R, - (R, cos (fa + fa,t) + (R, cos (fa + fa,s) = cos [(fa,s - fa.t) + fa- fa]
Gradual elimination of the side ratios yields, after extremely tedious and
solved simultaneously:
where F1, F2, and F3 are cubic expressions in tan ^t-, and F[, F2, and F'3
are similar expressions in tan #,-. (For details of these expressions, the
reader is referred to Freudenstein's original paper.) After 0,- and \pf have
+ =/(*):
, _ d\p _ d\p dy dx
d<t> dy dx d<t>
Further $" =
Freudenstein's equation:
(R( (Rs cos <t> + & cos \p = cos (<t> iA) (12-2)
First derivative:
*'
fC,f,
d<t>
d-ty' dy1 dx
dy' dx d<t>
v2'
V"
kfy
^""
Second derivative:
Third derivative:
Fourth derivative:
R, cos < +
scale factors k$ and k+ and the link positions <f>p and ^p which define the
chosen precision point. The known values, that is, <t>p, lf/p, \l/'p, and \l/'p ,
are substituted into Eqs. (12-2), (12-8), and (12-9), yielding three equa-
and <t>p. Substitution of the known values into Eqs. (12-2), (12-8),
(12-9), and (12-10) and progressive elimination of the side ratios leads
After \l/p has been found, the side ratios are computed as in the third-
order approximation.
The known values are substituted into Eqs. (12-2) and (12-8) to (12-11).
Elimination of the side ratios and of the angle $p finally yields the
equation
After $p has been found, <t>p and the side ratios are determined from the
Means of a Programmed Digital Computer, Trans. ASME, vol. 81B, no. 2, pp.
generating Mechanisms by Means of the IBM 650 Computer, IBM 650 Program
FIG. 12-2
"quality index," and then computes the coordinates of the tracing point
Q = ^A*
tracing point)
Configuration diagram: p + s = q + r
ponents, ASME J. Appl. Mechanics, vol. 24, no. 4, pp. 22-24, 1957.
325
qx + rx = sz + p
Qu t~ fy ^ 0
+ TzWr =
+ ryar =
+ 7yor2 rxotr =
sya,
(12-12)
(12-13)
(12-14)
(12-15)
(12-16)
(12-17)
The projections qx and qy of the driving link, its angular velocity W3,
O, p O, x
FIG. 12-3
The projections rx and ry of the coupler, its angular velocity <ar, and
angular acceleration ar
The projections sz and sy of the driven link, its angular velocity a>,, and
angular acceleration a,
Eq. (12-13) with Eq. (12-14), the following equations are obtained in
which qx, qy, rx, and ry are expressed in terms of the unknown angular
qz = (a)
CO, - C0r
qy = "'- u' ^
rx = - ? (c)
= sy(W, C0,_
COg C0r
Next, Eqs. (12-16) and (12-17) are combined to eliminate ar, giving
Substitution of Eqs. (a) to (d) into Eq. (e) yields a quadratic equation
in of the form
One root of this equation is equal to uq, giving infinitely long links q
and r, while the other root, which represents the actual angular velocity
-,-!-.. m
particulars:
a. = 10 rad/sec2
104,259 ,.
8,765
sx
the data for the driving and driven links were interchanged.)
quantities would finally yield an equation of the fourth degree in qx, thus
ips sv/sx are calculated, using the general procedure outlined in Illus-
requirements:
a, = 0
Table 12-1 gives the results of the preliminary approach. It can be seen
that the required value of tan <p, = sy/sx = -| will be given by two values
of qx, one between 0 and +0.5, the other between 1.0 and 1.5.
requirements of p, q, coq, aq, w a and tan <p but only one of these gives
The details of the two mechanisms, which are shown in Fig. 12-4, are:
328
PLANE MECHANISMS
KINEMATICS AND
OF
DYNAMICS
Mechanism 2:
Q* =
-1.02
in.
ft
=-
-1.72 in.
= 2.00
in
rx =
+3.57
in.
-1.55 in.
= 3.89
in
sx =
-2.45
in.
*.
-3.27 in.
= 4.09
in
?*
t.
Sy/Sx
+2.0
+2.85
+1.5
1.323
+2.61
0.41
+1.0
1.732
+2.31
0.69
+0.5
1.936
+1.90
1.15
2.00
+1.38
2.35
-0.5
1.936
+0.625
+ 11.6
-1.0
1.732
-0.525
+ 1.38
-1.5
1.323
-2.50
TO.53
First, qy is eliminated by combining (b) with (c), and (6) with (/):
The system has been reduced to five equations, Eqs. (a), (d), (e), (g),
and (h), containing the five unknown quantities qx, rx, ry, sx, and sy.
Next, sy is eliminated by combining (g) with (e), and (g) with (h):
The field has been reduced to four equations, Eqs. (a), (d), (i), and (j),
(j) yields
The term sx is eliminated next by combining (k) with (a), and (k) with (d) :
giving qx = 0.268p
PPP
PP
^ = 1.195 -
PPP
rotate with a constant angular velocity w,. Since the driven link s
sense.
For a mechanism in which o>g = const, Eqs. (12-16) and (12-17) reduce
to
~ "" "
/\9/\oioW/
Values of the Angular Velocity of the Driven Link, ASME J. Appl. Mechanics,
331
(12-19)
(12-20)
r2 _ rx2 + r2 _ w,2w,2
(w
(9)
(w9 -
ar22 + ^, Wr2)2wr22
(h)
(i)
w,l2 W22
7 V2 2 a'l = fa"2 _
Wr
Eqs. (A) and (i) and combining the resultant equation with Eq. (J)] gives
The term T on the right-hand side of Eq. (k) is similar to the left-hand
side, except that the suffixes 1 and 2 are interchanged. Equation (k)
is a cubic equation in &>rl and wr2. Since it contains two unknown quanti-
solutions.
either wri or wr2, within the limits set by the following criteria:
d2u,
of the angular velocity of the driven link at the instant when the latter
With a suitable ur1 substituted into Eq. (A;), the cubic equation is
solved for wr2. Of the three roots, the one which satisfies condition
(12-21) and yields positive values for ari2 and ari2 has physical meaning.
fied considerably.
w.
while in the case of a drag-link mechanism, wi and w,2 have the same
following specification:
Solution. In this case uT1 may be taken as zero since this value
The first root does not satisfy condition (12-21) and is discarded. The
second root agrees with this condition and yields positive values for
333
ing conditions:
Crank-rocker mechanism
Fig. 12-5
the range 11 to 18 rad/sec, since values outside this range give negative
ari2 and ar22. With a selected value of wri = 15 rad/sec, Eq. (k) reduces
to
The equation has only one real root, wr2 = 12.5 rad/sec, which satisfies
qx = 30.00 in.
s* = 20.00 in.
r. = 0.00 in.
qv = 0.00 in.
sy = 15.40 in.
ry = 15.40 in.
q = 30.00 in.
s = 25.24 in.
r = 15.40 in.
qx = -23.32 in.
sx = -24.42 in.
8.90 in.
rx =
qy = 18.85 in.
sy = 6.30 in.
r = -12.25 in.
A2
Position 1
Drag-link mechanism
Fig. 12-6
Since the terms qxrx + qyry and rxsx + rysy are recognized, respectively,
as the scalar products q - r and r - s, and the term (r^ r^x) as the
w,2 qr cos (<pq <pr) + wr2r2 = w,2rs cos (<p. <pr) + asrs sin (<p <pr)
(m)
In the next step, arr is expressed in terms of the other quantities, thus:
or asq2rq sin (<pq <?r) arr2 = us2rs sin (<ps <pr) (p)
from which aTr = uq2q sin (<pq <pr) w,2s sin (<p, <pr) (g)
Substitution of Eq. (q) into (n) yields the following expression for a,:
o.2(3a
.)
(r)
335
(12-22)