[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Primary- vs. secondary-current feedback



Original poster: "Steve Ward" <steve.ward@xxxxxxxxx>

Hi Ken,

I found that performance-wise, either type of feedback can be made to
work equally, given proper tuning and tweaking of the system.  I
personally prefer primary feedback for 2 reasons.  Firstly, as you
noted, the secondary base current isnt necessarily in phase with the
primary current, this could lead to a serious hard-switching scenario.
Using primary feedback ensures soft-switching provided your circuit
does not have large delay time.  Secondly, i noticed that when using
secondary base feedback, the resonant frequency of the secondary would
jump to a harmonic during ground arc conditions (i have actually
captured waveforms of this happening).  This of course would lead to
very serious hard-switched conditions, which i believe took out many
of my IGBTs.  The big killer here seems to be the voltage transient
produced in these events (the heating should be small, since the duty
cycle is so low).

Ive been using primary feedback exclusively now.  I still do have
failures here and there, but i usually find the source, and its never
feedback related.  My most recent failure (at the midwest 2006
teslathon) was due to a spark jumping far out of its way to hit my DC
supply power cord.  The spark had to find ground through my IGBTs to
the heatsink.  I didnt know what the cause of failure was at the time
of the teslathon, but later Jeff Larson sent me a video where it was
pretty clear to see what went wrong.  Those who attended the RATCB
thon 2 weeks prior could attest to my coil's reliability (no rogue
streamers!).  Long and intense run times at power levels nearing 7-8kW
(real power, not kVA).

Steve

On 11/20/06, Tesla list <tesla@xxxxxxxxxx> wrote:
Original poster: "K. C. Herrick" <kchdlh@xxxxxxx>

Whiling away the time awaiting a replacement 'scope (a used Tek 2465
for my recently-defunct--and also "pre-owned"--7904), I've taken a
further look at a couple of simulations.  Each one consists of a
simple feedback-oscillating circuit driving a 3:500, k = 0.2, 4.5 uH
tuned-primary, ~84 KHz simulated Tesla coil.  In one, I take feedback
from a 1:100 current transformer in series with the primary.  In the
other, it's from the secondary's return-current.  I also include a
spark-simulation: switching in, at 90 us, an additional load of 100K
in series with 10 pF.

Here's what I notice that's interesting to me:

1.  With primary-current feedback, the current appears in-phase with
the secondary voltage for the first few cycles, then the voltage
starts to lead the current.  But after a dozen more cycles, the
secondary voltage ends up lagging the primary current by almost 90
degrees, staying that way indefinitely.  There are two significant
"notches", at ~9-10 and ~16-17 cycles; and after that, no more
notches.  Note that the "spark" occurs at 90 us while the first notch
occurs at ~120 us.

2.  With secondary-current feedback, the secondary voltage uniformly
lags the primary current -- by around the same almost-90 degrees; and
with a very minimal notch at ~150 us.

3.  In the primary-feedback case, the current-transformer-derived
(low) voltage is dead-on in phase with the primary current as would
be expected.  But in the secondary-feedback case, there's a
significant shift: with 600A peak primary current, the derived
voltage goes through zero at ~150-180 A.

4.  In both cases the voltage applied to the primary circuit is
constant, throughout; the primary current stabilizes at ~700A peak;
the secondary voltage stabilizes at ~300 KV; and the rise to maximum
secondary voltage occurs over the same quantity of cycles -- about 7.

So if this more or less mirrors the real s.s.t.c. world, which is the
better option?  On the one hand, it might be nice to avoid the
notches.  But on the other hand, a convenient low voltage that is
fully in phase with the primary current might afford a convenient
reference for the cutting off of primary current at zero-crossing.

And in both cases, since the notches disappear after a dozen cycles
or so, perhaps (as did happen with my sadly-defunct untuned-primary
s.s.t.c.) very long sparking times will give entertaining results.

In my current re-build, I'm employing primary-current-derived feedback.

Comments on this?...

Ken Herrick