1
Solid State Tesla Coils (SSTC) / 120VAC Mains Ramped Builds
« on: January 21, 2022, 03:00:04 AM »
Got a lot of useful information reading this website so for what it's worth here are some pics and a little info about a ramped build I thought worked out pretty good. ZakW posted a lot more info about a similar build which has gotten great results. I'll probably try to rebuild mine and jack it all up so here is the breakdown.
I saw good results from the Loneoceans RSSTC and one shown on SciTubeHD's youtube channel so I basically tried to wind some coils with similar dimensions and resonant frequency. I also tried some other staccato interrupters already out there but Gao's seemed to work the best for me. I've found it difficult to exceed 12 inch arcs on an SSTC without upping the coil size so the budget QCW mains ramped approach seems like a really good setup and who doesn't like sword sparks.
They are based on Steve Ward's SSTC 5 driver (https://www.stevehv.4hv.org/SSTC5/miniSSTCfnlsch.JPG) with a "staccato interrupter" in place of Steve's 555. I'm running FGA60N65SMD IGBT's on it and still using a 1:1 GDT but using a 15v regulator to power the gate drivers and this just makes things simpler.
Other changes are things like the DC blocking capacitor value at the driver outputs (which should probably be experimental, I think I have like a 2uF ceramic) and the addition of a 1uF DC blocking capacitor in series with the primary (a few .33uF 1200V MKP caps I had).
I had a fairly large T37 core so I used that to wind a single GDT for driving both halves of the bridge. It allowed me to fairly easily wind 5 twisted wires. I have a CT for the feedback with about 30 turns. To add PLL feedback I put a 5v driven CD4046 in between the hex inverter and the gate drivers.
Below are the two builds I've tried with the half bridge on the left and full bridge on the right. They are both using very similar coil dimensions of 36 AWG on 9cm tall by 8cm wide formers (and run about 400khz). A 32 AWG coil of slightly larger dimensions seemed to work well also. The one on the left has the staccato and logic transformer inside the box and the one on the right currently uses an external logic supply and gets the mains sync from a transformer inside my interrupter box. I plan to get a nice toroid for the full bridge and hopefully hit that 20 inches. They are both more or less identical builds but with one using the full bridge and PLL feedback.
This is is a big universal interrupter box I went ham with trying to put everything I could into it over time. The interrupter circuit is isolated from the RCA plug with an optocoupler and 9v battery which I found almost necessary without 'getting lucky' dodging interference. I did the mains sync like that because I figured I'd want to be able to run any other builds I make mains synced without having to make a new circuit each time. I'm using Gao Guangyan's circuit seen here https://www.loneoceans.com/labs/sstc3/schema_sstc3_staccato.jpg but with C4 modified for the full 60hz half wave (quieter thicker arcs due to adding and shutting off after the downward slope).
Bridge and driver board. I believe there is about 2uF worth of film caps across the bridge and I stacked what I had at the time.
There are 15v, 12v, and 5v regulators, the 12v just being for the fan. There is a janky heatsink I used which was just a couple alu plates from somewhere I put together so it has this fan on a 40C N.O. switch. At ~8ms and highest BPS it appears to pull around 5A from the wall with the current tuning. I noticed the arcs were decent enough without the PLL though.
Using the two UCC2732x gate drivers which seem to run the full bridge ok in this operation but it's possible at ~8ms on times at highest BPS they will start overheating eventually. If I run it hard for a while to the point I can feel the heatsink getting more than just warm (with fan blowing) then the output seems to drop a little probably indicating the drivers and switches don't like it. Nothing gets too warm at the lower break rates it typically runs so it all kind of works out. I get them from mouser and Ebay/Amazon or something like that will always give you fake ones so don't do it.
A pretty basic layout that hasn't punished me for it but hey I'm not recommending a slack build or anything. This is the 2nd time I made a full bridge "test layout" like that I ended up just keeping in a final build, the other being a DRSSTC which uses the same kind of bridge layout, but to be fair might be getting saved by the snubber. I usually test with an isolated DC supply first, then low voltage AC, then full mains. So far with various coil testing these IGBT's have held up so they are pretty trusty.
A few shots of some ~18 inch arcs. I have a piece of foil above set to 20 inches from the breakout point which I could not get it to strike no matter how clever my insults were. When I had it set to 18 inches it would strike it every now and then but who knows what slight variations could result given environmental changes etc. I was originally struggling to get it to hit 16 inches before just messing with the primary a little bit and randomly testing various phase angle adjustments.
Here's a 16 inch strike from the half bridge catching mainly the tail end of the arc. I was pretty liberal with the size of the boards so in reality the spark to coil ratio could look much better after condensing the box down and everything looks smaller. I tuned that half bridge, got it to strike 16in, then was too excited about it to mess with it any further lol.
When it comes to the longest arcs from such a build it seems like it becomes a battle of finding a running frequency which is not too high or low, as well as trying to get it to form arc channels which don't branch on their way out as actual swords. That is to say I've seen tuning which seemed better in phase and probably had more power in the arcs but due to a tendency to branch more, a different tuning which pulled less power might yield longer discharges. I've kept fairly high coupling with these something around normal SSTC range and tried to keep the primary just out of streaking. There seems to be a certain science to 'encouraging' discharges to form just right as that voltage in the secondary starts to skyrocket. Slight changes in the breakout point length can for example make a big difference. A stuttering output with popping and clicking may indicate this energy trying to go somewhere and ending up in primary discharges you might not see in the light, where it may otherwise find its way out a longer breakout point, or one which is oriented to the side for example. I think using mosfets and an even smaller coil maybe closer to 600khz could get either the same arcs or longer. That's about all the insight I can personally offer into it so if you made it this far reading thank you for your sacrifice.
I saw good results from the Loneoceans RSSTC and one shown on SciTubeHD's youtube channel so I basically tried to wind some coils with similar dimensions and resonant frequency. I also tried some other staccato interrupters already out there but Gao's seemed to work the best for me. I've found it difficult to exceed 12 inch arcs on an SSTC without upping the coil size so the budget QCW mains ramped approach seems like a really good setup and who doesn't like sword sparks.
They are based on Steve Ward's SSTC 5 driver (https://www.stevehv.4hv.org/SSTC5/miniSSTCfnlsch.JPG) with a "staccato interrupter" in place of Steve's 555. I'm running FGA60N65SMD IGBT's on it and still using a 1:1 GDT but using a 15v regulator to power the gate drivers and this just makes things simpler.
Other changes are things like the DC blocking capacitor value at the driver outputs (which should probably be experimental, I think I have like a 2uF ceramic) and the addition of a 1uF DC blocking capacitor in series with the primary (a few .33uF 1200V MKP caps I had).
I had a fairly large T37 core so I used that to wind a single GDT for driving both halves of the bridge. It allowed me to fairly easily wind 5 twisted wires. I have a CT for the feedback with about 30 turns. To add PLL feedback I put a 5v driven CD4046 in between the hex inverter and the gate drivers.
Below are the two builds I've tried with the half bridge on the left and full bridge on the right. They are both using very similar coil dimensions of 36 AWG on 9cm tall by 8cm wide formers (and run about 400khz). A 32 AWG coil of slightly larger dimensions seemed to work well also. The one on the left has the staccato and logic transformer inside the box and the one on the right currently uses an external logic supply and gets the mains sync from a transformer inside my interrupter box. I plan to get a nice toroid for the full bridge and hopefully hit that 20 inches. They are both more or less identical builds but with one using the full bridge and PLL feedback.
This is is a big universal interrupter box I went ham with trying to put everything I could into it over time. The interrupter circuit is isolated from the RCA plug with an optocoupler and 9v battery which I found almost necessary without 'getting lucky' dodging interference. I did the mains sync like that because I figured I'd want to be able to run any other builds I make mains synced without having to make a new circuit each time. I'm using Gao Guangyan's circuit seen here https://www.loneoceans.com/labs/sstc3/schema_sstc3_staccato.jpg but with C4 modified for the full 60hz half wave (quieter thicker arcs due to adding and shutting off after the downward slope).
Bridge and driver board. I believe there is about 2uF worth of film caps across the bridge and I stacked what I had at the time.
There are 15v, 12v, and 5v regulators, the 12v just being for the fan. There is a janky heatsink I used which was just a couple alu plates from somewhere I put together so it has this fan on a 40C N.O. switch. At ~8ms and highest BPS it appears to pull around 5A from the wall with the current tuning. I noticed the arcs were decent enough without the PLL though.
Using the two UCC2732x gate drivers which seem to run the full bridge ok in this operation but it's possible at ~8ms on times at highest BPS they will start overheating eventually. If I run it hard for a while to the point I can feel the heatsink getting more than just warm (with fan blowing) then the output seems to drop a little probably indicating the drivers and switches don't like it. Nothing gets too warm at the lower break rates it typically runs so it all kind of works out. I get them from mouser and Ebay/Amazon or something like that will always give you fake ones so don't do it.
A pretty basic layout that hasn't punished me for it but hey I'm not recommending a slack build or anything. This is the 2nd time I made a full bridge "test layout" like that I ended up just keeping in a final build, the other being a DRSSTC which uses the same kind of bridge layout, but to be fair might be getting saved by the snubber. I usually test with an isolated DC supply first, then low voltage AC, then full mains. So far with various coil testing these IGBT's have held up so they are pretty trusty.
A few shots of some ~18 inch arcs. I have a piece of foil above set to 20 inches from the breakout point which I could not get it to strike no matter how clever my insults were. When I had it set to 18 inches it would strike it every now and then but who knows what slight variations could result given environmental changes etc. I was originally struggling to get it to hit 16 inches before just messing with the primary a little bit and randomly testing various phase angle adjustments.
Here's a 16 inch strike from the half bridge catching mainly the tail end of the arc. I was pretty liberal with the size of the boards so in reality the spark to coil ratio could look much better after condensing the box down and everything looks smaller. I tuned that half bridge, got it to strike 16in, then was too excited about it to mess with it any further lol.
When it comes to the longest arcs from such a build it seems like it becomes a battle of finding a running frequency which is not too high or low, as well as trying to get it to form arc channels which don't branch on their way out as actual swords. That is to say I've seen tuning which seemed better in phase and probably had more power in the arcs but due to a tendency to branch more, a different tuning which pulled less power might yield longer discharges. I've kept fairly high coupling with these something around normal SSTC range and tried to keep the primary just out of streaking. There seems to be a certain science to 'encouraging' discharges to form just right as that voltage in the secondary starts to skyrocket. Slight changes in the breakout point length can for example make a big difference. A stuttering output with popping and clicking may indicate this energy trying to go somewhere and ending up in primary discharges you might not see in the light, where it may otherwise find its way out a longer breakout point, or one which is oriented to the side for example. I think using mosfets and an even smaller coil maybe closer to 600khz could get either the same arcs or longer. That's about all the insight I can personally offer into it so if you made it this far reading thank you for your sacrifice.