Re: OLTC update - primary IGBT loss

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Paul Nicholson by way of Terry Fritz <twftesla-at-qwest-dot-net>" <paul-at-abelian.demon.co.uk>

Terry wrote:

> I do note that the Vce is about 90 degrees out of phase with the
> primary voltage.

That's to be expected, since Vce (resistive) is in phase with Ip
and therefore 90 degrees from Vp.

> The internal inductance of the emitter is speced at 13nH.

which translates to 2*pi*F*L = 0.003 ohms reactive.  In conjunction
with your estimated 0.003 resistance, we'd expect a phase angle
between Ip and Vce of 

  arctan( 2*pi*F*L/R) = 45 degrees. 

When I look at your 

 http://hot-streamer-dot-com/temp/OLTC09-02-03.gif

I see the Vce at about 88.5 deg from Vp.  I don't see much sign of
the claimed 13nH.  An Ip trace would help to settle the matter.
Come to think of it, if you had 13nH in the emitter, wouldn't that
be upsetting the drive signal Vbe by a few volts?  Something's not
quite adding up here.

> The three turn primary was nice in that it gave lower Lpri.  But
> from a theoretical and modeling point of view, it sure is a pain.

Yes.  What I'll do is 'fabricate' a cylindrical primary that gives
the same k factor and tuning.  It'll have the same radius and overall
height and will be approximately 1 turn. Then we can get some sort of
a model, and we can try to predict the effect of hacking the secondary
down to a lower L/C ratio.  You'll trade volts for charge and it will
be interesting to see the effect.  Probably end up with a wide but
thin toroid - bicycle tyre shape :).  Hope we can get a lot of data
from the big L sec before you start to unwind.  

>  A messy problem.  You don't need to worry with it unless you really
>  want to.

It would be quite a job to add that in.  I've been contemplating adding
in the stuff to model the capacitively coupled dual resonator, which 
is less daunting than this.  We're some way off being able to model
arbitrary resonators by just bolting together the integral operators
for each distributed reactive component, but that's where it's heading
if you guys insist on doing these ruddy non-standard coils!  In fact,
putting together the integral operators is the easy bit. You end up 
with a big matrix of complex numbers that describes the whole
resonator, and the hard bit is to find the vectors (voltage and current
profiles) which pass through this matrix unchanged.  These are the
resonant modes of the system.  There's no way to solve this computing
problem in a general way for an arbitrary resonator matrix and you
have to apply heuristics with some fore knowledge of the pattern of
solutions for the particular resonator layout involved.

> I wonder if coupling in three spread out rings hurts the waveform
> on the secondary.

No sign of a problem.  The FT of your beat envelope shows only the
normal amount of higher mode ringing.  No sign of HF modes from
the primary either, so no racing arcs from this cause!

>> Can you use a fibre-optic current probe between topload and breakout
>> point?  That would give us the streamer current.  We could at least
>> see where in the RF cycle the streamers are taking current.  

> Sure!  Tonight I'll work on such things.

That would tell us a great deal, qualitatively, I think.

> Perhaps instead of getting a probe to measure 500,000 volts, we need
> a coil that can go down to 60,000 volts...  :-))) 

The solid state coils certainly are appealing - controllable, nice
clean waveforms.  Operating at lower voltages with a smaller coil
does have some appeal.  Probing the coil has a greater relative effect
so proportionally more care is needed to account for these effects.
Also, it is not clear how conclusions drawn from a few cm of streamer
are to be scaled up to normal sized coils.  I'm sure we could tackle
these problems, and it would be better to make some progress by looking
at a low voltage breakout, than not do anything at all.
--
Paul Nicholson
--