Steve Ward SSTC5 build

[Michelle_]

Gotta make doors for the case. Debating trying the other secondary. Overall seems rock solid at full power, it doesn't like a bigger top load though. Seems able to make maximum 10" (250mm) sparks.





Here is a video of it running with the interrupter slowed down:

[Benjamin Lockhart]

Nice work! The primary being in such close proximity to the secondary is asking for trouble at higher power (full bridge) though.

[Michelle_]

Nice work! The primary being in such close proximity to the secondary is asking for trouble at higher power (full bridge) though.

Thanks for the warning I'll try to insulate it with thicker material.

[Benjamin Lockhart]

Insulating materials other than an air gap can actually increase the dielectric stress in that area and eventually cause a failure. The best thing would probably be to increase the primary diameter and wind it on a larger piece of pipe or some other form, but that reduces the coupling so you might need to play around with the height.

[davekni]

I added terminals for the primary and the input power but I don't know if it was a good idea.

That's fine. Bit of added inductance for those two places is not important.  Terminals on FET or IGBT connections would be more problematic.

Thanks for the feedback on the diodes. I wanted to change them and the caps to top mount anyway so now there's a good reason!

With good plane layout as you have, diode location is less important.  Fine in your new layout too.

The only issue I had was the TVS diodes...

Are you using bi-directional TVS diodes?  If uni-directional, that would explain your problem.

Pretty interesting looking

Yes, looks great!  You may want to consider a larger diameter for most of your breakout point, with a small diameter only at the tip.  That will minimize corona along breakout wire, leaving a bit more power for actual arc.  Of course, wire corona looks interesting itself.  You could leave breakout as is (thin wire) if that corona is a desired visual effect.  Playing with interrupter pulse width and frequency will give vary results.  See what looks best to you.

Nice work! The primary being in such close proximity to the secondary is asking for trouble at higher power (full bridge) though.

Thanks for the warning I'll try to insulate it with thicker material.

Insulating materials other than an air gap can actually increase the dielectric stress in that area and eventually cause a failure. The best thing would probably be to increase the primary diameter and wind it on a larger piece of pipe or some other form, but that reduces the coupling so you might need to play around with the height.

Insulation can be fine if there is no air gap adjacent secondary.  This has been done with electrical tape or filling gap with epoxy.  If there is an air gap at all, then make it at least several mm.  Otherwise corona discharge may occur within a small air gap and slowly degrade secondary insulation until failure occurs.

[Benjamin Lockhart]

Oh, I stand corrected about the insulation.

[Michelle_]

I thought they were bi directional TVS diodes but maybe they aren't, I'll check them out. The way they were hooked up should have worked right?

Also I definitely need to do something with the primary because it was sparking when I ran it last time. I started printing a tube that will fit over the secondary but now I'm wondering if I needed to make the gap bigger. I kind of want to avoid filling it with epoxy because I'm allergic to epoxy, will search for ideas. I wonder if I can use electrical tape then the tube fitting tightly over that.

Something else interesting is I tried the blue secondary again and at full power it was pretty close to the green one except it blew the GFCI again. Now that I think about it it's only been the blue one that's blown the GFCI and never the green one. I have another higher current EMI filter, should I try it, or use both even?

[Michelle_]

So I tried some wider primary winding situations, one of which was around a tube with an airgap between it and the secondary. Everything I Was doing was making things worse;  sparking between primary and secondary, blowing the GFCI, bad/skipping output. With the low break rate it's pretty easy to tell when the sparks happen because it skips a pulse and you can generally see or hear them happen. That or the coupling was bad and the variac would make weird noises.

What seems to have fixed the problems was going back to the thin insulation with the primary wrapped as tightly as possible then taped over. This not only provides the biggest output but doesn't seem to trigger the GFCI, or spark at all from primary to secondary. What seems to have initiated the sparking was disturbing the winding when I was swapping secondaries (from its originally tight state).

This also might explain why the blue one blows the GFCI, maybe the coupling is too low and I need to raise up the primary or add more turns.

This seems counter intuitive to me and yet here is steve wards:

https://www.stevehv.4hv.org/SSTC5/base7.JPG

and Kaizer's which seems to have a small amount of solid insulation with tight wrapping over it: 

https://kaizerpowerelectronics.dk/tesla-coils/sstc-design-guide/

[Benjamin Lockhart]

Here is mine by the way. I've never had primary to secondary sparks unless I run it without a breakout point. This is a half bridge at 340V DC.

[davekni]

I thought they were bi directional TVS diodes but maybe they aren't, I'll check them out. The way they were hooked up should have worked right?

Connection is correct.  Though a bit better to have both TVS diode and GDT secondary solder to source lead close to FET body.
Looking closely at your picture, I can see a cathode indication band around one end of TVS diode.  That shows it is uni-directional.  Bi-directional TVS diodes don't have a cathode demarcation.

Thanks Dave, the 3 turn wire connected to the signal generator is floating and not grounded. I'll try scoping it the way you're saying though.

BTW, I'm surprised you got any feedback without connection to other end of 3-turn winding.  It is a current transformer (CT), so you need to drive it with current.  Driving with voltage as you did, only current is due to stray capacitance.  Doesn't matter now, however, since you have actual feedback working.

Something else interesting is I tried the blue secondary again and at full power it was pretty close to the green one except it blew the GFCI again. Now that I think about it it's only been the blue one that's blown the GFCI and never the green one. I have another higher current EMI filter, should I try it, or use both even?

Yes, I'd try both.  Difference may be just slight difference in operating frequency.  Line power wiring back to breaker may be resonating close to blue secondary frequency.  Also, adding a counterpoise under coil connected to ground (to secondary bottom after CT) will likely help.

This also might explain why the blue one blows the GFCI, maybe the coupling is too low and I need to raise up the primary or add more turns.

Yes, GFI tripping may be due to sparks from secondary to primary.  Perhaps more likely than line wiring resonance as in my above paragraph guess.

What seems to have fixed the problems was going back to the thin insulation with the primary wrapped as tightly as possible then taped over.

Even if not sealed, tight may be OK.  If tight enough, air gap is so small, capacitance is high, so voltage across gap may be below corona threshold voltage.  I've seen tape as first layer over secondary as a way to get tight spacing, with tape adhesive filling in gaps.  But you may be fine even without adhesive against secondary given good tight construction.  Corona issues are most common with DRSSTCs using vertical primary windings, and with air gaps of a few mm.  SSTCs have a lower peak primary voltage, so less issue with corona there.
And, yes, SSTCs perform best with high coupling.  Larger primary diameter lowers coupling.

[Michelle_]

Here is mine by the way. I've never had primary to secondary sparks unless I run it without a breakout point. This is a half bridge at 340V DC.

Interesting. I believe you by the way, I've seen plenty of pictures of ones like yours. Not sure what would make it work or not work right though.

[Michelle_]

I thought they were bi directional TVS diodes but maybe they aren't, I'll check them out. The way they were hooked up should have worked right?

Connection is correct.  Though a bit better to have both TVS diode and GDT secondary solder to source lead close to FET body.
Looking closely at your picture, I can see a cathode indication band around one end of TVS diode.  That shows it is uni-directional.  Bi-directional TVS diodes don't have a cathode demarcation.

Thanks Dave, the 3 turn wire connected to the signal generator is floating and not grounded. I'll try scoping it the way you're saying though.

BTW, I'm surprised you got any feedback without connection to other end of 3-turn winding.  It is a current transformer (CT), so you need to drive it with current.  Driving with voltage as you did, only current is due to stray capacitance.  Doesn't matter now, however, since you have actual feedback working.

Something else interesting is I tried the blue secondary again and at full power it was pretty close to the green one except it blew the GFCI again. Now that I think about it it's only been the blue one that's blown the GFCI and never the green one. I have another higher current EMI filter, should I try it, or use both even?

Yes, I'd try both.  Difference may be just slight difference in operating frequency.  Line power wiring back to breaker may be resonating close to blue secondary frequency.  Also, adding a counterpoise under coil connected to ground (to secondary bottom after CT) will likely help.

This also might explain why the blue one blows the GFCI, maybe the coupling is too low and I need to raise up the primary or add more turns.

Yes, GFI tripping may be due to sparks from secondary to primary.  Perhaps more likely than line wiring resonance as in my above paragraph guess.

What seems to have fixed the problems was going back to the thin insulation with the primary wrapped as tightly as possible then taped over.

Even if not sealed, tight may be OK.  If tight enough, air gap is so small, capacitance is high, so voltage across gap may be below corona threshold voltage.  I've seen tape as first layer over secondary as a way to get tight spacing, with tape adhesive filling in gaps.  But you may be fine even without adhesive against secondary given good tight construction.  Corona issues are most common with DRSSTCs using vertical primary windings, and with air gaps of a few mm.  SSTCs have a lower peak primary voltage, so less issue with corona there.
And, yes, SSTCs perform best with high coupling.  Larger primary diameter lowers coupling.

Thanks Dave for sharing your thoughts. My bad about the diode, a mistake was made lol. That makes sense that it shouldn't have the demarcation.

I'm going to try the CT for real (I've been using the antenna) so we'll see how it goes.

That makes me feel better there's a reason the tight wrapping can work. I really couldn't understand why lower coupling made it worse. Maybe at a certain voltage level it's better to have it farther away?

[NyaaX_X]

Hello~

I think, what about dry cardboard? Is that a good isolate material?

And the waveform to drive MOSFET should be measured at G-S, but not through the gate resistor.

The uni-TVS could be place back to back to make a bi-TVS.

[Michelle_]

Hello~

I think, what about dry cardboard? Is that a good isolate material?

And the waveform to drive MOSFET should be measured at G-S, but not through the gate resistor.

The uni-TVS could be place back to back to make a bi-TVS.

I would think cardboard could insulate especially if it's not corrugated. Good to know I was measuring it wrong haha.

I don't really know what to make of the tight versus loose primary. I've seen both ways, loneoceans for example has the loose primary. Maybe the loose primary also needs more turns and that's why it didn't work for me. I am willing to try different things to see if it works better.

I feel like 10" sparks is pretty good for a 110v half bridge SSTC? Just based on what I've seen from other people. So I'm not sure what if anything else I can do to improve it besides going to a full bridge or higher input voltage. I could try running it off our single phase 208 at work but I don't know if it's a good idea.

For now I guess I want to finish the enclosure, clean up the wiring, and test the CT - and see if it remains reliable.

[davekni]

I really couldn't understand why lower coupling made it worse.

Lower coupling requires more primary current to transfer same power to secondary.

I don't really know what to make of the tight versus loose primary. I've seen both ways, loneoceans for example has the loose primary. Maybe the loose primary also needs more turns and that's why it didn't work for me.

I'd guess you need fewer turns (and associated much higher current) or higher bridge voltage (208 or full bridge) resulting in somewhat higher current.

I feel like 10" sparks is pretty good for a 110v half bridge SSTC?

Yes, seems good to me.  Pushing this design may be fun, but also may result in dead FETs etc.

[Michelle_]

I really couldn't understand why lower coupling made it worse.

Lower coupling requires more primary current to transfer same power to secondary.

I don't really know what to make of the tight versus loose primary. I've seen both ways, loneoceans for example has the loose primary. Maybe the loose primary also needs more turns and that's why it didn't work for me.

I'd guess you need fewer turns (and associated much higher current) or higher bridge voltage (208 or full bridge) resulting in somewhat higher current.

I feel like 10" sparks is pretty good for a 110v half bridge SSTC?

Yes, seems good to me.  Pushing this design may be fun, but also may result in dead FETs etc.

Yeah it seems to work well so I'm not sure how much I should mess with it. I was going to build a RSSTC next so I might just build a full bridge for that and leave this one alone.

[Michelle_]

Tried the secondary CT for feedback and it works great as a plug and play replacement for the antenna. I did have to reverse the wires since it was wrong on the first try. This does seem to be a superior alternative although the antenna did work fine besides looking ugly and there being corona on it.

[Benjamin Lockhart]

With secondary base current feedback, one thing to be aware of is that it doesn't always pick up the correct frequency like primary current feedback does with a DRSSTC.

I've killed 2 small SSTCs by running with the secondary shorted.
Once was just to discharge the bus caps. I didn't have it grounded and just connected the base wire to the topload. Next time I ran it, the bridge was dead.

The other one was just from pulling arcs too close to the toroid! (this was a different full bridge SSTC)

A PLL (phase locked loop) can solve this problem for SSTCs but these simple designs work fine if you don't short the secondary during operation.

[Michelle_]

With secondary base current feedback, one thing to be aware of is that it doesn't always pick up the correct frequency like primary current feedback does with a DRSSTC.

I've killed 2 small SSTCs by running with the secondary shorted.
Once was just to discharge the bus caps. I didn't have it grounded and just connected the base wire to the topload. Next time I ran it, the bridge was dead.

The other one was just from pulling arcs too close to the toroid! (this was a different full bridge SSTC)

A PLL (phase locked loop) can solve this problem for SSTCs but these simple designs work fine if you don't short the secondary during operation.

Interesting, I guess it makes sense though I'll try to make sure not to short the secondary. I actually haven't tried drawing arcs from this one yet.

[Michelle_]

I finished the case. The fans are internal and medium sized, just running off the 5V rail which was only powering the CT module before. I am considering running shielded wire to the interrupter but I'm not sure if there's a problem with how it is.





I decided to leave the EMI filters external and keep separate power inputs so that I can use the variac. At some point I will want to put the EMI filters in an enclosure and use cord grips for strain relief. It hasn't been blowing the GFCI with this setup but that could be more to do with the coil being in tune.