Thick vs. thin wire update
After replacing the 1500 turns of #28 wire on my 42" spark TC,
with 660 turns of #20 wire, and replacing the 22 turn, #12 primary,
with 9 1/2 turns of 1/4" copper tubing, the spark was weaker. It did
reach 42", but only one time, and it needed 590 watts, instead of the
580 watts it used before.
I did more tests. First I re-installed the original .0147uF cap, to
make the comparison more valid, and I added 1 turn to the copper
tubing primary to give me enough tuning range. Now the best tune
point is around 10 1/8 turns. After trying a range of coupling, tuning,
sync-phase, and ballast settings, the best I can get is a 40" spark
at 590 watts. Even then, the spark very rarely reaches the 40" mark.
It hits 40" less often than the old set up hit 42" at this same power
level. I'm still using the 4" by 13" smooth toroid on top of the 3" by 10"
smooth toroid. I even gave the new setup an unfair advantage by
raising the toroids a little more, all to no avail.
I then replaced the #12 tank wiring (which got warm), with # 6 wire
which doesn't get warm. This made absolutely no noticeable
difference to the spark length or appearance. In general, the entire
primary and secondary stays much cooler than the old set up, but
the sparks are shorter now.
Basically two things have changed; the frequency is higher
(200kHz vs. 85kHz before), and the primary surge impedance is
lower now. I have no easy way to determine which factor is having
the most negative effect on the spark length, but in any case the
thick wire is reducing my spark length to a noticeable degree. There
is nothing wrong with thick wire per se, except that it reduces the
inductance for a given coil size. A larger TC of course requires thicker
wire, and this does not cause a problem.
The gaps seem to be brighter now, so it is quite possible that the
lower surge impedance is increasing my gap losses. This may be
causing the problem. The surge impedance issue has been
discussed extensively on this list by Malcolm and others in the past.
I suppose that as the surge impedance is increased, a point is
eventually reached at which further increases give little benefit.
I don't really know how much effect the increase in frequency might
I'm sure that the losses in my primary and secondary are lower than
they were before, yet this cannot fully compensate for the other
negative factors of this new TC configuration.
This is the most careful comparison of this type that I've made so
far. This is a test I'd been wanting to do for a long time. I am not
aware of any flaws in my test. I'm losing about 5% spark length
using the thick wire. Regarding the common prescription to use
"1000 turns of wire" on the secondary, this might help some, but
might not give the full benefit of using 1500 turns. Of course it's
really the inductance that's important, so a wider coil can
successfully use fewer turns, etc.
As the number of turns in both the primary and the secondary is
increased, undoubtedly a point eventually will be reached at which
the losses caused by using thin wire, over-shadow the benefits of
a higher surge impedance and decreased gap losses. It is possible
that a coil with 2000 turns or more would work as well or better than
a coil having 1500 turns, but this in unknown to me.
I also made some tests at higher power levels up to 1000 watts, and
the new arrangement remains weak. I obtained a maximum of 49"
sparks, vs. 52" to 58" with the old setup. The sparks just don't look
as impressive now....it seems like a different (weaker) coil.
Based on this work, it seems likely to me that a number of coilers
are seeing short spark lengths due to their use of a too thick
secondary wire, and too few turns on their primaries. But the spark
length difference is unlikely to be more than 5%, which is certainly
not the end of the world :) If the coil is made physically larger,
then the wire can be thicker, because the inductance will still be
high, and the primary surge impedance will be high, and the gap
In other old experiments in which I doubled the tank capacitor size,
and reduced the number of turns to tune, I did not lose any spark
length. I figure that the surge impedance was still high enough in
those cases to keep the gap losses low, and the sparks at full
It would be nice to determine if frequency has much effect on spark
length. I don't mean super high frequencies, but frequencies more
within the range of common TC's such as 85kHz, vs. 200kHz, etc.