Drsstc 3

[Netzpfuscher]

I strongly suggest to use active cooling on the bridge. I calculated the numbers for my SKM400 coil. At 16% dutycycle with 1000A peak I have around 500W power dissipation.
The cooler you start your burst the more headroom you have to the max junction temperature.

That's the reason why my new coil has a big heat sink and a strong cooling fan. 

[Hydron]

16% duty cycle is really high, is this with pulse-skipping and very long on-time?

As for the OP's bridge rectifier going pop, was it due to the streamer hit to the primary?
Commonly accepted practice is to tie one side of the DC bus to earth via a moderate value capacitor (e.g. 100nf), so that any current from a streamer strike has a low-impedance path to earth from the primary. The theory is that if this low-Z path is not present then you may run the risk of internal breakdown of the IGBTs or diode bridge to the earthed heatsink they are mounted to. Not sure how tested it is other than Steve Ward removing his strike ring and encouraging arcs to hit the primary of one of his coils (without any damage), but it does make sense as a theory!

[Netzpfuscher]

Yes 16% is with pulse skipping and for SID files with 4 simultaneous voices. With this numbers I wanted to indicate that there is a fair amount of dissipation in a DRSSTC even at lower duty cycles.

The capacitor solution is a problem with RCDs, the high frequency ripple on the bus drives a current through the capacitor and this trips the RCD. I suggest to use two strings of TVS diodes from the inverter output to earth.  (6x 100V 1,5kW)

[oneKone]

Cheers for the nudge to go active cooling @Netzpfuscher, I'll implement it when I replace the rectifier.

@Hydron the primary strike wasn't what killed the rectifier but I think it could have weakened/damaged it somehow, the failed when a greater input current was used.

[Mads Barnkob]

Even a electrostatic discharge can weaken a silicon-based component, so you could have a internal failure where the insulating layer between the copper base plate of the rectifier and its silicon die was broken down through a micro channel. Just a single small flash through and some burned goo or plastic and its done

[oneKone]

With the toroid on my mind I set out to fix it. I tried bending each segment more to achieve a tighter radius but I still had sharp edges, to fix this I added copper rings to the top and bottom to cover them... And boy did it work! I'm getting almost 2m streamers now. There's still a bit of leakage but I think with a Polish and a coat of polyurethane it will fix the problem. I also think I might have the secondary too close to the primary, most of the time a ground strike is met with an arc from the bottom of the secondary to the primary. The first video is without the rings, the second video is with the rings.

/>

/>

[Mads Barnkob]

Good that you see improvements! I watched the whole MIDI video, good stuff

Your coupling is too high or it is badly tuned when you get racing sparks, it can also be a insufficient grounding of the secondary coil.

I am actually not sure if varnishing the topload is a good idea, will it disrupt the electron charge and the toploads ability to have a surface capacitance, anyone can answer this?

[Uspring]

I am actually not sure if varnishing the topload is a good idea, will it disrupt the electron charge and the toploads ability to have a surface capacitance, anyone can answer this?

The varnish will have two properties. It will be a dielelectric and an insulator. As a dielelectric it will be polarised, causing an electric field opposite to that of the top load. Since it is a volume effect and the varnish probably thin, there will be a negligible increase of top load capacitance.

As an insulator, the varnish won't stop the appearance of an electric field around a charged up top load. Near points of high curvature, ionisation of the air will occur. If the top load is say, at a high positive voltage, the free electrons created will be attracted to and stick to the varnish and generate an electric field opposite to that of the top load. Eventually, this will stifle the ionisation.
But once the top load voltage reverses, the stuck electrons will increase the field and ionisation will start again. This time positive ions will stick to the varnish. So you will have some ionisation on each cycle.

Another way to see this is, that the varnish will act as an insulator of a capacitor, insulating the top load from a more or less conducting volume of air around it. And a capacitor is an AC conductor. The appearance of racing sparks along secondaries and primary secondary flash overs illustrate, that high frequency voltages don't really care about insulators. But they do help somewhat.

The long breakout rod you use is certainly a good idea. The breakout from there reduces the top load voltages. But the best option is to avoid sharp edges.

[oneKone]

Cheers for the info guys. I'll polish the top load and mess around with the tuning, I might even make some discs to lower the coupling if need be.

As it stands time is running out in this coil for the year, as summer approaches the days get longer so I won't be able to run the coil 😓

[oneKone]

small update.

After trying for a while to get the spiral toroid working i finally listened to @Acid Byte. A quck trip to the hardware store and a few cuts on the cnc router later i had a new toroid! the diameter is 700mm with a height of 150mm. I also started to get working on the active cooling, it's simply a 3d printed box to house two hp server psu fans. I'm hoping to get some real runs in and fine tune the setup this upcoming friday... until then some pictures!

[Mads Barnkob]

Nice sparks and also a nice looking coil, I am a total fan of secondary coil diameter and toroid topload minor diameter being the same!

You properly had to de-tune the primary some more? Noticeable enough to have higher peak current and thus getting longer sparks from a lower impedance system?

[oneKone]

Cheers,
Yes the tuning changed quite a bit, I think something like 18khz or more. I still have fine tuning to go. To be honest I'm not sure what it would normally pull before as I had the OCD set to 900a.

One thing I've been thinking about making is essentially a vu metre for current, have the leds in 100a steps 🤔

[Mads Barnkob]

Cheers,
Yes the tuning changed quite a bit, I think something like 18khz or more. I still have fine tuning to go. To be honest I'm not sure what it would normally pull before as I had the OCD set to 900a.

One thing I've been thinking about making is essentially a vu metre for current, have the leds in 100a steps 🤔

I actually made such in 2011 and never finished the project It was based around a cheap DMM IC called ICL7017 and a VU meter IC, guess microcontrollers like Arduino surpassed that project by now. But it was meant to just be a homemade CT and a 0-2000A display + VU meter.

[oneKone]

Cheers,
Yes the tuning changed quite a bit, I think something like 18khz or more. I still have fine tuning to go. To be honest I'm not sure what it would normally pull before as I had the OCD set to 900a.

One thing I've been thinking about making is essentially a vu metre for current, have the leds in 100a steps 🤔

I actually made such in 2011 and never finished the project It was based around a cheap DMM IC called ICL7017 and a VU meter IC, guess microcontrollers like Arduino surpassed that project by now. But it was meant to just be a homemade CT and a 0-2000A display + VU meter.

If I ever get around to it I'll post it up, to me it's an easy way to see what's going on..


I think this might be the last update on this coil 🤷‍♂️.

Final OCD is set to 900a, pulling 15a arms from a 230v outlet @ 100us. I'm getting over 2m streamers at 100us, I'm tempted to push it but I'm also happy with the results. I'm also happy as this is the first drsstc build that I've done with an almost 100% success rate, with the bridge rectifier being the only part to fail!

All in all I'm happy and am keen to get onto my next builds. So far my upcoming builds will be a small 180khz skm100 fullbridge coil, and some gu50 tube projects! Ad well as the big dr project...... That's a lot on the horizon!!

Some pics: