Subject: Primary Coils
From: richard.quick-at-slug-dot-org (Richard Quick)
Date: Thu, 31 Aug 1995 02:00:00 GMT
>Received: from ns-1.csn-dot-net (root-at-ns-1.csn-dot-net [18.104.22.168]) by uucp-1.csn-dot-net (8.6.12/8.6.12) with SMTP id WAA17994 for <tesla-at-grendel.objinc-dot-com>; Wed, 30 Aug 1995 22:24:36 -0600
* Original msg to: usa-tesla-at-usa-dot-net
Quoting "Lima, David" <dlima-at-analogic-dot-com>:
> Thus far I have not been able to understand the benefits and
> differences between cylindrical, flat ,and saucer shaped primaries,
> Do you have any FAQs or text on this? Or if not maybe suggest a good
Man, until I write mine there is simply no real good books out there.
I have a rather extensive library on the subject and I can find a
section in one book, a chapter in another, but not one single refer-
ence that has any real dirt on primary coil designs and characteristics.
The standard line is "wind a few turns of heavy conductor".
So let's discuss awhile and see if I can put this in fuller perspective.
The primary coil design and construction is critical to a well balanced
and efficient Tesla coil system.
First lets talk about the actual conductor. The Tank circuit pulse
oscillates with very heavy RF current. This energy exhibits "skin
effect" in that the electrical current does not flow through the
conductor as much as over it. Penetration has been calculated as only
being a few mils (.001 inch = 1 mil). The best conductor for this
type of electrical energy is a smooth flat strap with a few mils of
silver plating. The problem with flat strap primaries is that the
high voltage promotes corona loss from the thin edges, and this
loss can be significant. The sharp edge also provides a jumping off
spot for arcs to strike the secondary when the secondaries are
"overcoupled" (more on this later). One way around this is to use
a bar instead of a strap. The bar can have the thin edges rounded
to reduce corona losses and flashover, but it is not an efficient
use of conductor.
Another conductor of choice is thin wall copper tubing. This has
many advantages. The thin wall tubing is an excellent conductor
for tank circuit energy, and it is available in a wide variety of
outside diameters. The smooth rounded surface reduces corona loss.
I use the thin wall soft copper refrigerator tubing.
Heavy Litz wire would work well.
Conductors to avoid would be solid wire, braided strap, stranded
wire that has a "twist" in the conductor (parallel stranded is OK).
The next aspect of primaries is the size and shape. In the past many
designers have advocated small primaries consisting of five or so
total available turns, and frequently tapped in at three turns or
less. This line of thought is derived from radio theory where the
voltage gain was assumed to result from a ratio of turns trans-
formation. We now know that this is a misconception.
The Tesla secondary is in fact an open ended resonator. While it
is true that the ratio of turns transformation is responsible for
some voltage gain in the secondary, the majority of voltage produced
from a Tesla coil is pure resonant rise such as associated with open
ended transmission line VSWR. With this understanding the importance
of having a high ratio of turns between the primary and secondary
is diminished, and uniform excitation of the resonator becomes a
The small five turn primary at the base of a coil does a very poor
job of uniformly exciting the secondary coil. Most of the energy
is forced into the bottom third of the secondary winding on these
systems and these turns are unnecessarily stressed. The solution
was frequently to space wind to unstress the secondary, which in
turn reduced the ratio of turns transformation, which further
crippled the poor design. The small primary, large capacitor, space
wound secondary is a design loser. These systems are highly stressed,
inefficient, and require large and expensive capacitors and trans-
formers that are prone to high failure rates.
The solution of course is to move in the opposite direction. The
primary needs more turns which cover a larger area. This means a
substantially larger primary coil. Instead of a five turn primary
coil tapped in at three turns, the coil needs to have fifteen turns
tapped out at twelve or more turns. This primary design creates
a much larger field flux which completely engulfs the secondary
winding. The resulting excitation is much more uniform.
What happens next is very interesting. Because the secondary coil
is uniformly excited, the secondary coil can be close wound with
magnet wire without breakdowns and excessive corona losses. The
close wound magnet wire secondary has a very high inductance per
unit volume of coil form. When coupled to a large, high inductance
primary, the resulting coil set will achieve the same ratio of turns
transformation as the small primary design discussed previously.
This is a "win-win" situation.
The large primary coil tunes with a much smaller (and cheaper) tank
circuit capacitance which requires a smaller power supply to charge.
Yet the system is able to process power more efficiently because of
the large field flux and uniform excitation of the high inductance
secondary. The overall RF power processing efficiency is so great
that coilers who follow this design method blow away other designs
by producing bigger sparks with much less input power. As an added
boon, the large primary coil and high inductance secondary form a
"slow-wave" system. The tank circuit operates at a lower frequency
with less peak power which in turn reduces the stress on the gaps,
caps, and power supplies. The gaps quench easier and I have found
that the tank circuit capacitors have a much longer life. Run times
are longer with a higher duty cycle. When something finally does
fail, it is cheaper to replace.
So now that I hopefully have convinced everyone that large primaries
and close wound high inductance secondaries are the way to go, what
shape primary coil is best?
The basic primary shapes are the vertically rising helix; the flat
"pancake" Archimedes spiral; and the "hybrid" inverse conical section
or "saucer" coil. I do not like the coupling characteristics and
the shape of the field flux produced by a vertically rising helix
primary. I do not recommend using them unless you are building Tesla
Magnifier systems. For small to medium 1/4 wave coils I would recommend
the "saucer" primary. As I mentioned earlier, I like about 15 primary
turns available, and the conical section should have an angle of slope
of about 30 degrees starting with the lower inside turn and rising to
the outside turn. On larger coils running with higher voltage inputs
and greater peak powers the best primary is going to be a flat pancake.
The reason I like these specific shapes is for their coupling
characteristics, which are derived from the size and shape of the
magnetic field flux they produce. A good primary coil will end up
being about as wide as the secondary coil is tall. When the primary
is tapped out over nine turns or so, the field flux that is generated
couples in the entire secondary winding all the way to the top turn
on the coil. I know this for a fact because in Tesla coil systems
that are overdriven the field flux contains so much energy that it
ionizes the air. Clearly visible corona maps out the field size, shape,
and relative intensity.
Next we need to briefly discuss coupling and over-coupling.
Coupling quite loosely is the inductive interaction between
two coils. In Tesla coil systems the pulse oscillation through
the primary creates a powerful magnetic field. When the gap is
quenched the magnetic field collapses and the energy that was
contained in the system is forced into and trapped in the
secondary. If too much energy is forced into the secondary coil
the resonator frequency will "split". Splitting is the formation
of parasitic harmonic resonances of higher frequency in the
secondary winding. Splitting is observed when the 1/4 wave voltage
peaks of these parasitic harmonics break out as sparks from the
middle of the winding. Sparks will jump up and down the sides of
the coil. Excessive corona resulting in a visible primary/secondary
field flux is another indication that the secondary coil is being
overdriven. Splitting and excess corona indicate that the system
is over-coupled. The cure for over-coupling is to raise the
secondary coil some, which reduces the inductive interaction.
If the primary coil is not properly shaped, like a flat pancake
coupled to a tall skinny secondary, then the coupling will be too
loose and insufficient energy will be transferred to the secondary.
There is no easy cure for this. If the primary under-couples or
couples too "loosely" you have to rework the coils. This is one
reason I suggest that the intended primary be designed and built
with tight or "close" coupling in mind. If the primary over-couples
you can then raise the secondary until the system is "critically
The one thing that did not just fall into place in the text above
is the turn to turn spacing of the primary. The space between turns
needs to be kept small. The largest spacing that I have required
is 1/2 inch between the edges of a copper tube primary where the
operating voltage was 20KVAC RMS into the tank circuit. Most primaries
need no more than 1/4 to 3/8 of an inch inter-turn airspace.
> Tesla built his coil in Colorado with the secondary and primary
> having the same diameter, and with a bottom fed 'extra' coil. Have
> you ever tried building a basement sized coil to these geometries?
There is no need. Tesla developed and built large diameter coils in
order to distribute the charge densities over a very large area. Tesla
was working in an era of "stone knives and bear skins" (Spock), his
coils were wound on open wooden forms and his best wire insulation was
natural rubber gutta-percha (loaded with carbon soot from the curing
process) with a cotton wrapping. This is not like what we use today.
Today I build hermetically sealed all plastic coil forms. Modern
plastics, sealers, insulations and adhesives can contain much
greater charge densities in a much smaller volume with less loss.
There is no legitimate reason to wind large diameter coils when a
properly designed and constructed coil of much smaller dimensions
will exceed the performance of the former.
... If all else fails... Throw another megavolt across it!
___ Blue Wave/QWK v2.12