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

Re: DRSSTC-3 VCO drive test (long)

Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi Steve,

At 12:40 PM 2/4/2005, you wrote: 
I did some more tests recently.  I devised a way to shift the
frequency of the VCO during a burst.  This didnt really have any
noticable benefit from what i could tell, simply introduced more
uncertainty from what i could tell.


Cool!


So i realised that my DRSSTC-3 usually operates on the upper pole, and
pspice simulations tell me that in order to run it on the lower pole
successfully, i need to tune the primary quite a bit lower than normal
 Since i didnt have any more primary turns, i simply changed my tank C
to get lower tunings.  This also changes my Zo.  Anyway, i got it
tuned (or at least think i had it tuned) and was happy to see that i
could maintain ZCS throughout the burst, except at the very beginning
where things were just slightly "off".  Figured it would be ok.


Neat!

Got
the coil making 22" sparks at about 180W input, with running the coil
for 80-90uS and about 400-420Apk on the IGBTs.  Interestingly, the
sparks were thinner, more like a SGTC, and they produced MUCH more
ozone!

! Really... Ozone production in Tesla coils is odd. My harmonic free solid state coil continues to produce virtually none at 900W real power into the arc...

Also, the sparks were relatively tiny until about 60-70% input
voltage, where at that point they grew enormously.  This is very
similar behaviour to that of my spark gap tesla coils, of which i
purposly tune the primary low to promote streamer growth.  I had to
reduce my coupling considerably to prevent racing sparks up the side
of the coil.

That is how I am setting up to run too. Primary with lower frequency and low coupling...

It seemed like things were finally starting to come
together... when the IGBTs failed :-(.


Darn!


I cant seem to figure out why the IGBTs failed.  I observed basically
ZCS (very close anyway), and it seemed to be running smoothly.

If the VCO lost it, the current might double, but your IGBTs should take that. There are four things that kill IGBTs...

1. Temperature - unlikely since you were running ~180W.
2. Voltage - Our reverse diodes and rails make that very unlikely.
3. Current - Possibly...
4. Cross conduction - Seems like a real possibility... But only because the others seem much less likely...

"I" would go looking for a false trigger that left an IGBT pair on.


Well, i think its time i get cozy with some more pspice simulations...
simulating feedback situations and tunings.  My next thought is to
look at primary feedback again (since this would ensure ZCS) and use
an overcurrent detection circuit to protect things.  But i want to
work with a Pspice model first and see if i can identify some tricks
for tuning the coil for best results.


I did that today ;-)   And that is how my coil works now >:o)  Check this out:

Cpri    Ip (peek)       Vs (peek kV)    Streamer Power (W)
130     183             233             666.07
135     124             222             488.42
140     129             220             570.26
145     141             226             620.93
150     167             242             650.85
155     191             252             666.07

As I vary the primary inductance, the streamer power and voltage stay pretty constant. But the "current" seems to be what is getting tuned! If it is out of tune, you get real high primary currents which feed just about the same secondary power. When it is in tune, the primary current is minimal for best efficiency. 135uH seems best for my coil which is 4.1% lower Fo than the secondary.

I think primary current limiting in the case of a primary synced coil is super important!! You could play with the coil and the spark will apparently stay constant, but the primary current can really go nuts!! I think it needs to protect itself!! Also, the primary current limit might save things when the uP goes haywire... Of course, if the IGBTs have like 50X current rating... ;-)

Here is my MicroSim model:

http://hot-streamer.com/temp/DRSSTC-10.sch

I could make a coil that is far better "matched" to the DRSSTC driver, but this is the coil I have at the moment. I think it is super cool being able to run just about "any" coil too. I changed to start pulse to 2uS so I may be able to do up to 500kHz now. Maybe not a perfectly matched system, but it can do it ;-)


These DRSSTCs are still evolving i guess.  If you dont have the parts
and patience... might not be for you ;-).  I think if i didnt have so
many problems and challenges arrise, i would not be having as much
"fun" (ok, so its only fun after its working :-)).

It is super nice being able to probe them with not too wild test probes and such. That REALLY helps!! They also behave very nicely to what the models predict and are super repeatable!!! Very nice coils indeed!!! I think we will fish out these odd events that make them blow up pretty quickly now!

Cheers,

        Terry


Steve W.