Ghetto I (450-2000W Class "TVS" SSTC) / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Ghetto I (450-2000W Class "TVS" SSTC)

previous 1 2 3 next

Move Thread

LAN_403

Marko

Sun Nov 25 2007, 01:13AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Nice sparks considering what are you using...

Dago wrote ...

I dont think you are running in class-E. Your schematic is also very confusing because its hard to make out which crossing lines are connected and which ones are not.

Matt wrote ...
PWM and Class-E don't mix...

And I'm worrying that with a very large breakout from the coil, the Q may suffer so much that maintaining class-E operation would be quite impossible... I'm also worrying that at such a low frequency, Class-E is pointless, and a halfbridge would be more rewarding.

Still a fantastic coil, and great pictures!

I don't think coil is running class E either - more in class-TVS with all the troubling power getting burned off as heat.

Mates, have you even tried to scope the C-E of the IGBT before calling the coil class E?

With true class E operation you won't need any TVS anywhere in the circuit.

That cap and diodes on collector are completely useless.

Also, how did you come to figure of 450W of your power consumption? Have you happened to measure cos fi of your system? It might be 450VA but it's terribly misleading to call it watts with huge amount of reactance at the input.

I guess I don't need to say that it would be enormously wiser to find an appropriate power supply for the coil rather than using insane amounts of motor caps as a reactive ballast.

And much safer for your circuit.

If you want to amke good progress go and use a bridge.

Good luck with audio-modulation too - last time I checked building a 400 watt audio amplifier of any type was not a trivial task.

Modulating the driver directly via open loop VCO or PLL might give you few cm of audio modulated sparks, but hard switching becomes really bad after that.

Make sure to use diodes on your bridge if you are going to do this.

Mates

Sun Nov 25 2007, 09:50AM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Marko wrote ...

I don't think coil is running class E either - more in class-TVS with all the troubling power getting burned off as heat.

Mates, have you even tried to scope the C-E of the IGBT before calling the coil class E?

With true class E operation you won't need any TVS anywhere in the circuit.

That cap and diodes on collector are completely useless.

Also, how did you come to figure of 450W of your power consumption? Have you happened to measure cos fi of your system? It might be 450VA but it's terribly misleading to call it watts with huge amount of reactance at the input.

I guess I don't need to say that it would be enormously wiser to find an appropriate power supply for the coil rather than using insane amounts of motor caps as a reactive ballast.

And much safer for your circuit.

If you want to amke good progress go and use a bridge.

Good luck with audio-modulation too - last time I checked building a 400 watt audio amplifier of any type was not a trivial task.

Modulating the driver directly via open loop VCO or PLL might give you few cm of audio modulated sparks, but hard switching becomes really bad after that.

Make sure to use diodes on your bridge if you are going to do this.

I measured the power consumption via of the speed of rotating the disk on my flat power meter. I related it to speed of rotation when my 2KW boiling tea-kettle is running (I do not have AC Amp meter).

The bridge topology:
I need two channels oscilloscope and probably different transistors for this. I was partially successful with the bridge but then suddenly both transistors died

I do not know why. Considering seeing your problems â€“ experienced power electronics users â€“ I rather stick to my concept. More over I think I can beat any of the FET bridge SSTC Iâ€™ve seen until now (regarding the length of sparks)

Class E:
Yeah you probably right. Most likely the coil is not Class E. I started to call it like that because the power part is similar to Steveâ€™s

Class E. However I have no idea what exactly the ClassE means â€“ I just remember from the secondary school that is type of switching the transistor and I thought that the parallel resonance cap is the key. Thus small explanation would be great (Iâ€™m a biologist â€“ sorry my ignorance)

The motor foil caps:
I do not plan to substitute this caps (BTW: By what? 1KW transformer?). I personally think that they are the key for such a relatively good performance of the coil. They are discharged into the primary long time before the transistor is switched of (this I measured by oscilloscope). This means I can use the IGBT and do not need to care about the fall time. The transistor stays much cooler than it was in my bridge topology because of this.

TVS:
The TVS heating is really not too much. I can run the coil for minutes and the TVS are just little warm (but the fan is on!). By increasing the coupling the TVS are cooler and cooler so hopefully I soon reach the stage that they are more or less spikes un-exposed.

cheers

Avi

Sun Nov 25 2007, 12:59PM

Registered Member #580 Joined: Mon Mar 12 2007, 03:17PM
Location: Melbourne, Australia
Posts: 410

That is interesting. Ive never seen a series capacitor work for a single switch setup, but then again ive never tried placing it in parallel to the switching device. Soulds like it would work since it would discharge the capacitor before the next cycle although i think that would be canceling out the idea of LC resonant rise. Correct me if im wrong anyone?

I just tried a similar thing in a simulation and it sure seems to kill the resonance.

file>import
$ 1 5.0E-6 10.391409633455755 50 5.0 50
l 304 160 304 208 0 1.0 -0.007635546547644788
c 304 208 304 256 0 1.0E-5 13.506643453460542
r 304 160 304 128 0 10.0
v 352 288 352 128 0 0 40.0 5.0 0.0
w 352 128 304 128 0
f 256 240 288 240 1
w 288 256 304 256 0
w 288 224 288 208 0
w 288 208 304 208 0
w 288 256 288 288 0
w 288 288 352 288 0
v 256 240 256 288 0 2 50.33 12.0 0.0
w 256 288 288 288 0
g 352 288 352 304 0
o 1 64 0 3 20.0 3.2 0

ragnar

Sun Nov 25 2007, 03:41PM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Try with a flywheel network in there:

(Can't tweak the values to get anything right -- anyone else want to try?)

$ 1 5.0E-6 35.02441195350251 50 5.0 50
l 416 256 416 304 0 0.1 0.037622444697678906
r 416 256 416 208 0 5.0
c 416 208 368 208 0 9.999999999999999E-5 -15.121625627588331
l 368 208 320 208 0 0.01 -0.254275816866146
w 320 208 272 208 0
w 416 304 272 304 0
c 272 208 272 304 0 1.8E-6 17.245362889466115
l 272 208 272 160 0 1.0 -0.25987014318196217
w 272 208 224 208 0
w 272 304 224 304 0
f 192 256 224 256 0
w 224 240 224 208 0
w 224 272 224 304 0
R 192 256 160 256 0 1 57.04 12.0 0.0
R 272 160 272 112 0 0 40.0 15.0 0.0
g 272 304 272 336 0
r 464 208 464 304 0 10.0
w 416 208 464 208 0
w 416 304 464 304 0
o 6 64 0 3 160.0 0.1 0
o 0 64 0 2 40.0 0.1 1

Avi

Mon Nov 26 2007, 10:46AM

Registered Member #580 Joined: Mon Mar 12 2007, 03:17PM
Location: Melbourne, Australia
Posts: 410

so what is the capacitor there for even?
to stretch out the 'flyback' pulse?

Mates

Mon Nov 26 2007, 11:59AM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Avi wrote ...

so what is the capacitor there for even?
to stretch out the 'flyback' pulse?

I'm not sure which cap you mean. The one which is in parallel to the diode - forgot about it that one is doing nothing.

On the other hand the 70nF/5KV which is in parallel to the transistor that one is really important. The coil works also without it but the efficiency is very poor and TVS gets really hot. My explanation is that it works mainly as an absorber of the spikes when the transistor gets open and because it is together with primary in resonance with the secondary the energy is used. It was actually quite difficult to tune the whole thing up - I was using 30V transformer for that trying different cap sizes and number of turns on the primary and measured the response on an incandescent tube (how strong it shines and what distance). But I can be completely wrong in my explanation...

Mates

Wed Nov 28 2007, 11:00PM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Hi everybody,
Today I increased the power to approx. 1500W, but unfortunately the increase in the length of the sparks is far from being linear to the energy investment. It looks I'll have to forget about 1m (40") sparks with this setting which was my aim

.

When the fan is on, the transistor heat sink keeps RT after one minute run.(Actually there are two BUP314 in parallel). But, the TVS are getting hot, however I can keep the hand on them after the minute run (it means below 60 C).

Regarding your comments:
I'm afraid of removing the protective diode (made from the bridge and parallel cap) from the circuit - are you guys sure it has no protective function? Ritchie Burnett also used diodes in his circuit which are in series with the FETs

Here is latest picture of sparks (18") and a short movie which was taken by my small digital camera...

Cheers

Tom540

Wed Nov 28 2007, 11:22PM

Banned on 3/17/2009.
Registered Member #487 Joined: Sun Jul 09 2006, 01:22AM
Location:
Posts: 617

I used them on my first few coils then I ditched them. The diodes were the only thing that failed.

Mates

Mon Dec 03 2007, 12:24AM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Not much progress this time - but I have to deal with flashes into primary. It is pretty annoying; because the coupling is so great in this position...I'll try to put more isolation around the secondary. I'm lucky that the set of transils can handle this

Here is the picture of my problem

and one picture taken just before the flashes started

Crow187

Mon Dec 03 2007, 07:53AM

Registered Member #795 Joined: Mon May 21 2007, 06:04PM
Location: Bocholt, Germany
Posts: 35

Good that your electronic is still alive ;)

Where did this spark come from? I can not see where it comes out of the secondary...

Perhaps your insulation is not that good? I have played with insulations between primary and secondary but my result was that the distance with air is nearly as good as the same distance between prim/sec made of plastic or something like that... At me it seemd that the sparks are just going through a lot materials without taking notice of the insulators.

I think less coupling is the only way that will prevent you for flashovers.

What do the others think?

Lukas

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob