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

OLTC update - Curiouser and curiouser!



Original poster: "Terry Fritz" <twftesla-at-qwest-dot-net>

Hi All,

I was running numbers and models today when I noticed that this picture of base
current (blue) sure has a long ringdown time (like 2mS!!!):O

http://hot-streamer-dot-com/temp/OLTC10-05-03.gif

There was no breakout in these low power tests but a 2mS ring down time was
easy to see as "odd" compared to the models!  So I fiddled around and finally
took the following data:

Pinger

PK  FREQ kHz (Error +/-)    Q FACTOR (Error +/-)   LEVEL
 1   37.229 (0.01%,3Hz)       84.19 (0.09%, 0.1)  -0.4dB
 2  113.864 (0.01%,8Hz)       46.26 (0.70%, 0.3)  -12.3dB
 3  172.448 (0.01%,12Hz)      34.37 (2.79%, 1.0)  -17.0dB
 4  223.862 (0.01%,16Hz)      29.69 (11.17%, 3.3) -20.3dB
 5  273.388 (0.01%,19Hz)      25.94 (18.78%, 4.9) -24.1dB
 6  273.391 (0.01%,19Hz)      26.01 (18.78%, 4.9) -24.0dB


Running Coil

PK  FREQ kHz (Error +/-)    Q FACTOR (Error +/-)   LEVEL
 1   37.148 (0.01%,2Hz)       83.73 (0.49%, 0.4)  0.0dB
 2  114.014 (0.01%,7Hz)       33.80 (22.33%, 7.5) -26.5dB
 3  172.208 (0.01%,11Hz)      35.87 (7.90%, 2.8)  -28.1dB
 4  226.978 (0.01%,15Hz)      37.49 (18.78%, 7.0) -29.4dB
 5  120.239 (0.01%,8Hz)        5.20 (100.00%, 5.2) -21.3dB
 6  119.635 (0.01%,8Hz)        7.28 (18.78%, 1.4) -23.3dB

The Q seems to have almost exactly doubled to 84 now!!  Of course, I am not
complaining ;-)) but I am thinking the SonoTube picked up humidity originally
and it is slowly drying out over time.  The Q of SonoTube is wildly affected by
humidity.  I really can't recommend it for TC use anymore even if it is fast
and cheap.

TCMA did not find PK5 in the normal running coil test above and PK2 was odd.  I
think that may be due to the gap still conducting in one direction with the
IGBTS?  That should affect the harmonics.

My models are very accurate these days:

http://hot-streamer-dot-com/temp/OLTC10-07-01.gif

I can now completely model the 240BPS case and actually understand the "why"
behind it all.  Having the IGBTs conduct during the negative swing does allow
for odd sweet spots in coupling and timing.  I need to play with that more.  
Also the models predict some noise spikes that are not really there.  The new
P5202 probe is helping a lot finding the "real" waveforms ;-)  I have to tune
the models there,  Probably stray capacitance somewhere is eating those spikes
that the "theoretical" case is missing.

Last night I switched to new transorbs (two 650 volt strings) in preparation
for 240VAC operation.  Voltage spikes are not a problem if it is tuned right
but if something goes wrong or I am playing with the knobs, they protect the
IGBTs.  I also put in 7 amps fuses but that was not rocket science ;-)

I think I am going to have to make a new IGBT driver for 27 volt drive.  My
present box just is not right.  I will also add a third (maybe forth) stage for
15 IGBTs in the new design.  I may try my present one first to see if maybe
going to 20 IGBTs would be a good thing.  Since IGBT destruction is not a
problem, I can go for the full setup there and lower the astronomical currents
they see now.  I may also go to two small 12 Volt gel cells for drive.  9v
batteries tend to drain vary fast.  I have a nice dual charger from my robot
stuff that could easily charge them so it would not be a big deal to do that
way.  I am probably hitting the gate at 360BPS right now and it seems to be too
much power for the little 9 volt batteries to take for long.  If they
discharge, they could cause low gate drive which is "bad"...

240VAC may need lower coupling since primary to secondary arcing is right on
edge now at 120VAC.  That would change timing and all a bit.  Hard to say
there...

I think more IGBTs would really help primary Q and I am getting very convinced
a plastic secondary would really help secondary Q (although, at 600 ohms, is
"good enough" now).  If I can successfully get it to 240VAC, that is 4X the
power ;-))  I certainly don't see any reason why it can't do that other than it
is "getting close" on primary current for the present system and primary to
secondary arcing is a problem.

I still don't know about gate drive resistance.  Lower resistance would help
some things but it would also increase transients as switching speed
increases.  Need more thought there.


Cheers,

        Terry