To see a Tesla coil sending
out showers of arcs in all
directions at close hand is
an unforgettable experience.
Malcolm Watts describes
the attributes and design of
one such coil.
inventing devices that ‘would evolu
tionise the use of electricity Possibly his
pest invention was the ac clot motor,
enabling the use ofan
te dlstbution sysiem EB
for powerng industrial TA
machine ood lighing
and heating systems
Of alls tventions
the Tesla coil is the [a
tos fascinating. Apart
ffom the spectacular
discharges — used for
lighting effets in Hollywood the coll isa
very compact high-volage generator. The
allabliyof cheap ebt rectifier, such as the
3BY700 fom Philips series makes output r=
tification of low fo medium power devices @
real possibilty.
J ist over a century ago, Nikola Tesla was
rr
efully the
coil described here is
POC ae ence eats
build or use it unle
Pore eue ney
and even then, follow
afety notes on page 195
Air-cored windings
The Tesla coil is simply a transformer with
primary and secondary windings but that is
‘where the similarities with a ‘normal’ iron-
cored transformer
end. The windings
ate aircored and
‘comparatively losely
‘coupled. This implies
a large amount of
‘magnetic fx leakage
and because of this
the ‘coil’ behaves
more like current
source than a voltage source. In fact, it has a
finite internal impedance determined by phys-
ical parameters that will enable it to deliver
maximum power to a load, in this case the
spark, only when the load matches this
impedance,
you fully
ety
Energy not dissipated in sparks will be dis-
sipated as electromagnetic radiation and heat-
ing of the primary components. Badly tuned
and incorrectly coupled tesla coils have given
the devices a notorious reputation as wideband
emitters of emi. The design procedure pre-
sented here should minimise this problem.
Feeding large amounts of power into the
arrangement, shown in Fig. 1, generates volt-
age gradients in the kilovolts-per-turn range
across the secondary coil ~ hence the single-
layer winding. Peak voltage at the top of the
secondary col is obtained just prior to sparks
breaking out. As soon as sparks appear, a load
is effectively placed across the secondary,
causing the output voltage to drop as output
‘current rises
Spark behaviour is govemed by convection
and loading. In a pulse-driven coil, bursts of
sparks are produced at twice the mains fre-quency and feed the are so formed which rises
by convection. When the arc, behaving a8 a
resistor formed from hot ionised gases,
‘becomes stretched, current and ionisation are
reduced to the point where it extinguishes and
anew one forms, It follows that power con-
sumption and loading of the system is far from
constant and harmonic output will depend on
the degree of variation.
Quantifying the peak output voltage is not a
simple matter of turns ratio as with a standard
transformer. Spark length, unlike a de or low
frequency situation, is not an accurate indica-
tor of voltage since repetitive ionisation of the
discharge channel causes successive sparks to
reach progressively further until primary and
secondary currents are at maximum values.
Spark length is probably a more accurate indi-
calor of system power throughput. For exam
ple, an arc over 150mm long can be generated
by a low voltage high current arc welder.
Experiments show that lengths of 300mm per
joule of energy stored in the primary capacitor
‘ean be reached in a well built device.
How the Tesla coil works
The system shown in Fig. 1 may be recognis-
able to those of you familiar with spark-gap
transmission systems. The major difference is
the lack of an aerial which is, in effect, entire-
ly wound on the secondary,
The top of the secondary coil connects to a
metal terminal possessing some capacitance,
This enhances series-resonant behaviour. See
references 145, pages 45-50, for an excellent
description ofthis. The ratio of inductance to
capacitance in the secondary system also
determines the voltage to current ratio of the
output.
Primary and secondary coils are initially
ver-coupled in order to throughput a decent
amount of power. In order for the system to
approach critical coupling with minimum
bandwidth and maximum power transfer it is
important that the coil is driven hard enough
to produce heavy sparking. Loading of the
secondary causes the Q of the entire system to
drop sharply which causes primary-secondary
coupling to tend to the critical value2?-*6 This
action also minimises radiation since power is
being expended in sparks
‘The primary circuit couples pulses of power
to the low impedance end of the secondary
coil and must resonate at the same frequency
as the secondary for maximum power transfi
While the secondary is tuned by its self-capac-
itance and terminal capacitance, the primary
coil is tuned by the primary capacitor which
also serves as an energy store,
The primary capacitor is charged by the
step-up transformer on every half-cycle of the
mains. When its voltage causes the spark gap
to fire, this energy is dumped into the primary
coil and hence into a rapidly changing mag-
netic field that also ‘cuts’ the secondary coil
‘The speed with which the field builds up is
determined by capacitor voltage and the value
of inductance, including mutual secondary
inductance, that the capacitor is discharged
into, The faster this rate-of-change, the greater
rie ee
breaker aT
T switch — |}
i
ero
1
Fig. 1. Schematic of the Tesla coll generator showing
primary and secondary tuned circuits,
Choke
the induced emf in the secondaty. As an
example of the very high power levels
involved, a capacitor dumping one joule of
‘energy in ten microseconds produces a peak
power of 100kW. The net effect is to transfer
energy from a relatively large primary capac-
itor to a very small secondary capacitance,
Its essential that the resonant frequency of
the primary circuit matches that of the sec-
ondary circuit as closely as possible. The pri-
‘mary capacitor, which virtually dictates the
quality of the system, should have a minimum
intemal inductance and resistance, ie, low est
for good coil performance. Old oil-filled
‘capacitors such as those used for smoothing in
transmitters work, but their current delivery
capability is hampered by unsuitable internal
construction, namely the thin lead wires. Theit
major virtue isthe selfshealing of the dielectric
after breakdown,
Suitable initial primary-secondary coupling
is required for best results. Finally, the pri-
mary chokes, whose job itis to disconnect the
step-up transformer from the primary circuit
when the spark gap fires, should have an
impedance considerably higher than the
impedance of Lp at the resonant frequency.
‘They isolate spark-gap transients from the
Terminal
Capacitor
iS
\
Sel.
Primary + | capacitance
coll
Tesla secondaryCoax rom
signal generator
ins and should be ‘lossy’ at high frequen-
ies to dissipate unwanted rf as heat.
Tesla coil design
‘The type of coil detailed in this article is a
pulse-driven two col system, Other configu-
Tations may be driven by a power oscillator
running atthe resonant frequency. You ean
seen here with custom-built capacitor, and
Coil former containing both primary and
secondary coils.
Signal ead
Fig, 2. Resonance of the primary
capacitor and coil s found by closing the
circuit and adjusting the signal generator
for maximum amplitude on the
oscilloscope.
tend-feed a secondary coil alone from a very
ow impedance oscillator for example, Other
designs include additional windings.
Pulse-riven coils develop by far the highest
‘outputs for a given input due to the trick of
storing energy over a relatively long duration
Of Sms and releasing it in a short period of
10us. The recommended starting point is to
choose a secondary coil height based on
required spark distance. Power feeds to short
high-power coils need special protection. The
philosophy taken here is that simplicity is best
and will eave fewer things to go wrong.
With a suitable height for the secondary
selected, a former with a diameter that is
somewhere between 0.2 to 0.5 times this is
recommended. Secondary formers may be
anything from pve pipe to dried and vamished
cardboard tubes. Rigid plastic tubing yields
the lowest losses next to air except pve which
hha the highest dielectric losses amongst plas-
tics in high frequency power situations!, Some
vvamishes may also exhibit losses.
‘The self capacitance of the coil is calculated
using Medhurst’s formula, Csc-HDpF,
‘where D is the coil diameter in centimetres
and is @ value depending on the height t0
diameter ratio of the coil. Values of H range
from 0.5 for heighvidiameter=2 to 0.81 for
heightdiameter=5. This formula holds if one
end of the coil is connected to ground or low
‘impedance and in close proximity to. ground-
ced surface such as ferro-concrete. This is
hhandy since the self capacitance is indepen-
dent of the type of winding as long as itis sin-
ale layer.
‘The next decision is whether to close-wind
for space wind the coil. Close wound coils
have the highest inductance but do not gener-
ally exhibit higher Q's than spaced windings
because of the proximity effect” and extra
wire resistance. The biggest plus to close-
‘wound coils is that lower resonant frequency
allows higher primary L/C ratios so that the
Q's of the two coils can approach equality
The benefit of spaced winding is that less
heavy vamishing is required. For close-wound
coils, glazed-enamel magnet wire must be
ee
Terminal capacitance for a sphere
é
750
where d is the sphere diameter in mm anc
‘he sphers ik moked about 2/25 above t
secondary coil. For a toroidal terminal,
‘capacitance is,
44, op
SHE eRe sd dikineses of hes
dy is the diameter of the tube and the
terminal is mounted above the coil, All.
emia mound
Sines eee pees
Wheeler's formula for inductance
er
pao
ar+100
‘where n is number of turns, r is coil radi
in inches, and h i coil height in inches,
‘The recommended minimum wire
diameter:
‘which is three times the skin depth at
frequency f; where fis,
1
3 Hz
era
Lis secondary inductance in henries and
is the sum of coil self-capacitance and th
terminal capacitance.
Since C can be calculated first, Z and f
values are flexible.
Primary coil
Minimum height per turn should be,
007K, + Pex rat
where Ve is peak capacitor voltage in th
primary components to minimise connec
wiring as this will add about 0.5uH per fi
and contribute to power losses, mostly in
fom ofem radiation. Hence it may be th
nomrintgral namber of tims is required.
Calculating peak output voltage
B= a
2ev?