MMC configuration advice

[zizzer]

Hi everyone,

I'm currently building my first DRSSTC having previously built a medium sized spark-gap coil. I've nearly managed to get hold of all the parts and have recently acquired 60 x 0.47uF 1kV MKP C4AS caps (edit: 20A Irms, 632A Ipeak, 4.1 ESR, 1344 dV/dT) for my MMC. I'm looking for some advice on selecting a sensible configuration.

As seems to be becoming the norm, I'm going for a Kaizer1/Loneoceans DRSSTC 3 sized coil:
Bus caps: 2x Nippon-Chemicon 6000uF 500V (in series, voltage doubled 240VAC in)
IGBT Full bridge: 2x CM300DU-24H
IGBT Snubbers: 2x ICEL 2uF 1000V
Secondary (not yet wound): 160mm x 792mm winding of 0.4mm wire (based on Mads' guide), Fres ~82kHz according to JavaTC

I've made a table of values (edit2: based on 750A primary peak, to give me room to play) to help me choose and based on other coils I've seen of a similar size I was thinking of going for an overall value of around 0.235uF in either an 8sx4p, 10sx5p or 10sx6p configuration. This is a pretty arbitrary selection though, so if anyone could give me some advice or feedback on my plans I'd be very grateful!

[Hydron]

This coil is fairly similar to my 160mm coil as well, so I can confirm that you're on the right track (all the specs and parts sound like good choices).

If you work backwards from your proposed range of capacitance (10s6p is 282nF, and also mentioned 235nF), the frequency (82kHz, though I'd maybe do the maths for something a little lower, e.g. 65kHz for some margin when streamers change the tuning) and assume a peak current of 1000A (probably about as high as you'd want to go for a coil like this - mine runs nice at 700-800Aish) you can work out the required MMC voltage.

Have a look at Mad's DRSSTC design guide and MMC calculators for info on how to work this out (simply the impedance of the capacitor multiplied by the peak current), but a quick check gave me numbers in the 7-11kV range depending on what frequency and capacitance chosen. I would suggest going for a minimum of 10s6p (can be adjusted down by reducing the number of parallel strings) or 12s5p for a bit more voltage headroom but slightly lower capacitance than your target.

You could always build the MMC in a temporary configuration to do some testing (ramping up the current limit carefully to not be caught out) to see how the coil runs at different tuning points and what the actual frequency and primary currents are before locking in the configuration and going for the final build. If this is difficult then personally I'd probably go for 10s6p with the option to disconnect a string or two for tweaking the capacitance if needed.

[zizzer]

Thanks for your reply and advice. It's no coincidence that my specs look similar to your 160mm coil - it's one of the one's I've been using for inspiration/reference  . If I could get performance anywhere near yours I'd be very happy.

The 10s6p with option to drop parallel strings seems like a sound plan, I'll build with that in mind I think.

I thought the 82kHz for my secondary might be a touch too high. I was mostly basing that on the trade off between frequency and getting the impedance to around 50kΩ as Mads suggested would be optimum. I did notice that most people building 160mm coils seem to have gone for more turns of thinner wire for a lower Fres at the expense of that 50kΩ figure. I wasn't sure how much weight to give each factor though.

Going back to JavaTC and using the same winding length and other params and plugging in the 282nF from the 10s6p I get an Fres of 75kHz with 0.35mm wire (60kΩ) or 65kHz with 0.3mm wire (70kΩ). I can also get 0.315mm wire which I assume would sit between the two. The primary inductance is ~15uH@75kHz and 20uH@65kHz.

I'm not exactly sure what all the trade-offs are between these choices, but I know the IGBTs will be happier switching at a lower frequency. My gut feeling would be to try and hit the 70-75kHz range based on all the other coils I've seen though. Does that sound alright, or is 65kHz still going to be a better target?

[Hydron]

I used 0.315mm wire (mostly due to availability) and didn't even know to worry about the 50kohm thing. Note that the wire diameter is not the turn-to-turn distance that you'll actually achieve - I think the insulation is in addition to this number (from memory I got more like 0.38mm between turns).

The coil seems to give good performance without being fussy about tuning - I have run it on both upper and lower pole tuning and various different coupling (prefer upper, plays a little nicer with MIDI but streamers aren't quite as "crazy" as with lower pole).

For reference, I measured secondary resonance of my coil as ~75kHz, dropping to 65kHz with a "fake" streamer (a bit of wire) attached, which matches up with what I can see from the secondary current data I got. I'd say that anything from 65-85kHz would be a reasonably range to go for - wouldn't sweat the choice too much.

[Mads Barnkob]

You have a overall sound plan here, parts in all in proportion with each-other and your worries about Fres and secondary impedance are maybe a little too much in detail, it is not small numbers like that which will make a big difference in the end. I think it was Steve Conner that once said that all Tesla coil work with all our assumptions is around +/- 3dB on all figures

I did fear that the secondary coil example list would be taken as a result list, should I ease up on the wording and highlight that it is merely a safe/proven range but that other configurations will work/be experimental?

0.3, 0.315 or 0.35 mm wire will not change much in performance, use what you got at hand.

I would aim at 0.3-0.5 uF MMC capacitance and then build it according to needed voltage/current specifications, that is how I made the MMC calculator to be used.

Good luck on project and we look forward to see some sparks in the future!

[zizzer]

Thanks both for your replies, that's put my mind to rest. You're right that I've got bogged down in the theoretical details too much. I'll get on and actually build the thing, then I expect I'll come back with some more questions when the real world settles all the finer points for me .