SSTC weird resonant mode / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

SSTC weird resonant mode

1 2 next

Move Thread

LAN_403

Frosty90

Fri Oct 02 2009, 09:15AM

Registered Member #1617 Joined: Fri Aug 01 2008, 07:31AM
Location: Adelaide, South Australia
Posts: 139

Hi all,

After working for about two days to get my TC going, I finally got the feedback right and got some sparks (though small, at 20v supply), "brilliant" I thought. Then today I built a nice interuptor, with two oscillators, variable on/off times etc, and got that going on the first go! "Super brilliant!" I thought. Then I decided to see what the actual resonant frequency of my coil was. JavaTC calculated around 260Khz with the topload, and I was running with an extra long 'antenna' breakout point (it had to be far from the toroid when running at low supply voltage to get breakout), so I expected the frequency to be abit lower, around 230khz or so. BUT, with the coil running and the scope probe in the air as an antenna, measurements on the scope seem to give around 460khz!! Whats going on?! So I connected the scope to look at the gate drive waveform, which was sure enough 460khz, and a weird distored shape (its normally almost a perfect 15v square at 100-200khz, but this is to be expected at this frequency trying to drive IGBTs I suppose.). So whats the deal? why am I getting such weirdness like this? The feedback scheme is a 1:60 CT around the secondary base wire, coupled via a 0.22uf cap, and loaded with back-to-back zeners (5v) and shotkeys, exactly like Steve Ward's feedback circuitry, and then that goes to a CMOS inverter etc. There is also a ~100kHz square wave signal from an aditional oscillator, feeding into the input of the CMOS inverter, via a 1k pot and 2k resistor, I adjusted the pot until the 'start-up' oscillator only just triggered the inverter, and when any other signal (from the CT), comes along, it takes over. It works on the bench, with signal generator etc, and it appears to work in the tesla coil at low power, until I look at the actual signals Im getting. So Whats happening?!

Cheers,
Jesse

Also, I'll put some pics in the project thread soon.

teravolt

Fri Oct 02 2009, 07:43PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

1: are you useing Steve Wards design?
2: try adding some resistance in parallel with your feed back coil, I think that the amplitude is to large going into your feed back.
3:have you tried actualy finding the secondary frequency with toride and the primary circuit frequency with a signal generator and scope

I had the same problem with Steve Wards "driver with ocd" and primary feed back so I added some resistance in parallel with the CT and 10k in series with the pcb. hope this helps

Frosty90

Sat Oct 03 2009, 04:06AM

Registered Member #1617 Joined: Fri Aug 01 2008, 07:31AM
Location: Adelaide, South Australia
Posts: 139

: try adding some resistance in parallel with your feed back coil, I think that the amplitude is to large going into your feed back.

Hmm, I'm unsure as to why a large amplitude would cause this problem, and anyway it is clamped at +/- 5v, but I tried a 1k in series with the feedback, and it worked, Im now getting around 260kHz, as expected. I thought at first that perhaps I had misread the horizontal scale on my scope, but I took the R out again and sure enough, back upto about 460Khz. I think what may have happened, is that the start-up oscillator may have been at a certain frequency that it 'pinged' a different harmonic or resonant mode and then the driver started going at that freqency. But the 1k resistor, perhaps, is slightly inductive enough (it was one of the cheap, highly inductive ones!), to block that higher frequency just enough, to let the lower one take over, but Im just guessing here, does anyone else have any suggestions?

1: are you useing Steve Wards design?

No, only the feedback scheme is the same, except for the 'startup' oscillator.

3:have you tried actualy finding the secondary frequency with toride and the primary circuit frequency with a signal generator and scope

I have, and it is about 260Khz, just as JavaTC calculated. The primary is untuned, it is not a double resonant system.

Cheers,
Jesse

teravolt

Sat Oct 03 2009, 04:33PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

I was just as perplexed but I had to do something. I finaly had 11k in series with the pcb and 680 ohms in parallel with the feed back transformer. I am useing Steve Wards new driver with some modifications of my own. I am using primary feed back and the CT can throw small sparks accross the pads on my proto board at higher powers. something else to fix.

Wolfram

Sun Oct 04 2009, 01:06PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

is this the feedback scheme you used? It is made for a DRSSTC with primary current feedback so I'm not sure if it's right for your application. The feedback scheme here

is for secondary base current feedback, and should be more suitable here.

Anders M.

Frosty90

Mon Oct 05 2009, 01:41AM

Registered Member #1617 Joined: Fri Aug 01 2008, 07:31AM
Location: Adelaide, South Australia
Posts: 139

Anders,

I'm using a design similar to the first one:

I use just a single CT and load it with zeners, rather than resistors, and I left out the clamping diodes at the input to the first gate, as I figured that the zeners clamp it to 5v anyway.

The trouble is, that with out the kickstart oscillator, the thing won't begin oscillating by it's self, I've found. And I'm at a loss as to why Steve's circuit actually works with out some sort of start up signal! I think it may be that the ucc gate drivers with the enable pin, have some sort of interference or glitch when the enable pin changes state, and that produces a pulse on the output to start the oscillations. I'm using the TC4429 driver ICs, and the interuptor is ANDed with the drive signal before the drivers. The two inverter gates in the schematic form a crude oscillator to produce a signal of about 100khz, then I adjusted the pot until it just barely triggered the first inverter gate. The capacitor is necessary because otherwise the CT loads the startup oscillator to heavily. Its probably not the best design, but it does work well:

However, it needed some experimentation to get good values for R3 and R1 for the feedback to work reliably. Originally, i didnt have R3, and thats when the thing started to oscillate at around 460Khz, instead of 260Khz. Then I added 1k for R3, which fixed the problem, now it works quite reliably! (except for primary to secondary flashovers!) What Im not really sure of is why adding R3 should make the difference, I'm thinking that It is slightly inductive enough to attenuate the higer frequency modes, and allow the 250khz mode through. But I really dont know.

Cheers,
Jesse

Wolfram

Mon Oct 05 2009, 08:10AM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

In your feedback circuit, the clamping zeners will act as a very low impedance, making the series capacitor represent most of the CT loading if I'm not mistaken. Loading a CT with a capacitor is not really a good idea in this application. I haven't seen anyone using the circuit you use in an SSTC before. Try replacing that part of the circuit with the one from this schematic

.

When you enable the UCCs, one of the UCC outputs will go high, essentially kickstarting the oscillation. It would seem to me that your logic implementation of the enble function should do the same thing though. Maybe your coil is harder to start ue to the strange feedback circuit you're using.

A.M.

Dr. Dark Current

Mon Oct 05 2009, 09:27AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Anders M. wrote ...

What is the purpose of the 10k resistor in series with the CT output?

teravolt

Mon Oct 05 2009, 04:37PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

the 10k probibly helps limmit the current into the diodes

Dr. Dark Current

Mon Oct 05 2009, 05:52PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

teravolt wrote ...

the 10k probibly helps limmit the current into the diodes

well it is a current transformer, so current out always equals current in * turns ratio, of course until you saturate it. So the 10k resistor IMO just wastes power

1 2 next

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