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

Re: SSTC- feedback vs. none

Original poster: "K. C. Herrick" <kchdlh@xxxxxxx>
Yes, perhaps I wasn't thinking too clearly: it's surely desirable to be able to forget about syncing a free-running oscillator to the Fr of the primary. And I think you're right about the slow IGBTs contributing most of the problem. With only 5 us per half-cycle to work with in my case, slow IGBTs are going to cause switching way up on the current curve. My fancy-dan digital phase-shifter may or may not take care of that; haven't found out yet although it does work, by itself, "as advertised".
As a matter of interest while I'm loafing around, I reworked my simulation of the low-level part of the feedback circuit, even incorporating rather sloppy current transformers (k = 0.8). My "pilot oscillator" sync's beautifully and there's no phase shift at all, from the (simulated primary) current-in to the logic-level IGBT-drive out. So absent any significant shift in my 5-winding toroidal drive transformer, the trouble if any will likely lie with the IGBTs. Unless my figuring is wrong, a 100 ns delay there will cause switching at 6% of the 100 KHz I-max, 200 at 12% and 500 at ~30%--not including other delays, of course.
What I would love to have is "bricks" that will switch on/off within not much more than a hundred ns, or at most 200. Pie in the sky? Looks as if, for the Semikron SKM400GB123Ds that I have, time-on can be between 200 & 400 ns and time-off, 700-900. Not too good without some kind of compensation for it. 

KCH

Tesla list wrote:
Original poster: Steve Ward <mailto:steve.ward@xxxxxxxxx><steve.ward@xxxxxxxxx> 

I say if everything is tuned properly, it will work just fine.  This
is much easier said than done!!  The streamers mess everything up.
Using feedback usually just eliminates 1 tuned part of the scheme.

I think you are in an interesting situation where your IGBTs are slow,
but your secondary Fr is kinda high, so feedback is just difficult to
make work properly.

Good luck!

Steve

On 1/26/06, Tesla list <mailto:tesla@xxxxxxxxxx><tesla@xxxxxxxxxx> wrote:
> Original poster: "K. C. Herrick" <mailto:kchdlh@xxxxxxx><kchdlh@xxxxxxx>
>
> A voice from the past, still around...
>
> My "bucket-primary" SSTC sits there un-fixed but I hope to get to it
> again before I expire.  In the interim it's occurred to me to ask,
> Why bother to employ feedback to sustain oscillation, when circuit
> phase-shift and/or inability to track exactly the secondary's Fr
> requires that excitation cease after the first few cycles, to evade
> that pesky "notch"?  Absent contrary words from the wise, I'm
> inclined to change my setup to a) stop each spark event after a max.
> of just a few cycles and b) merely use the starting-oscillator I've
> already incorporated to provide the drive to the IGBTs.  I currently
> employ a binary counter to cut off the drive, which would yield
> cutoff at only 2, 4, 8, etc. cycles.  If I change that to a
> decimal-output counter, I can cut off at 1, 2, 3,...10 (or maybe 9)
> excitation cycles.
>
> What say you?
>
> Ken Herrick
>
>
>