Benjamin's DRSSTC 2 in progress

[davekni]

Very interesting. Makes me wonder what would happen if I set the startup oscillator to force upper pole operation from the start? I think that's possible right?

Might be possible given your high impedance primary, though perhaps difficult with 22% detuning.  Upper pole frequency startup is typically used for QCW and line-voltage ramped coils.  Slow starting ramp makes it easier for self-oscillation to control startup.  Higher coupling separates the two pole frequencies further apart, which also makes it easier to favor one over the other.  A PLL based driver would have a better chance at upper pole startup/operation.
Not sure there's a benefit, but perhaps.  QCW coils (and line voltage ramped coils) usually perform better at upper pole frequency.  There's been discussion as to why.  If I'm recalling correctly, benefit of upper pole may have more to do with smooth slow arc growth as bus voltage increases rather than any more general advantage.  Simulation may provide some clue as to benefit or not.  I need to think about that possibility more.

[Benjamin Lockhart]

I'm wondering what's the exact reason that a fatter secondary coil allows higher coupling without racing sparks.
This is something that I've heard and repeated to others without knowing exactly why.
It it even true?

Does it have something to do with the number of turns vs inductance?

It seems to me that for long pulse DRSSTCs specifically, there is no reason not to use the highest coupling possible, within the secondary's limits. So I'm thinking about trying a much fatter secondary (I have some 19 inch diameter fiberglass pipe) which would allow me to get coupling to 25% with a flat primary at the bottom turn.

Here's a javatc design with the same secondary height and frequency as my current secondary, but larger diameter and less turns.
Would there be much benefit to such a colossal secondary coil? (it would be very expensive)

[davekni]

I'm wondering what's the exact reason that a fatter secondary coil allows higher coupling without racing sparks.
This is something that I've heard and repeated to others without knowing exactly why.
It it even true?

In general I think it's true, but with some caveats.  I have not personally experimented with two similar coils differing in secondary diameter.  Everything below is based on theory, not personal experience.  (I think there is perhaps some experimental data, but only be comparing many different builds on this forum, which differ in many parameters.  My DRSSTC has a relatively high aspect ratio, 110cm long and 16cm diameter.  Been problematic to get high coupling without racing sparks.  However, it's also low impedance for both primary and secondary.  Hard to draw any general conclusions.)

I know little about the transmission line effects (traveling waves down secondary) caused by sudden top load discharge.  Not sure if those contribute to racing sparks.  Rest of below ignores any such traveling waves.  I'll also ignore higher modes of secondary, such as when center is at high voltage and top and bottom are both at low voltage.  I don't think there's much energy in higher modes, especially not with high impedance primary and resulting slower current ramp rate.

Racing sparks are presumably caused by high local electric field, high volts/cm up secondary.  Anything that reduces breakdown voltage will contribute to racing sparks, such as higher humidity or locally ionized air due to adjacent ground-rail strikes.  (UV radiation from adjacent ground strikes might also be significant.  I know UV is important in triggering Marx generator spark gaps close to simultaneously.)  Most of my personal experience with racing sparks are associated with adjacent ground rail strikes.

With good wire insulation and varnish or epoxy coating, racing sparks should be along surface of coating.  However, one possible risk of using large diameter (low AWG) wire is that the same volts/cm becomes more volts/turn.  If wire insulation thickness doesn't increase with wire diameter, perhaps wire insulation could fail directly turn-to-turn underneath varnish/epoxy secondary coating rather than more normal surface racing sparks.  I don't know if that is a specific concern for your proposed 20 AWG secondary wire or not.

Secondary volts/cm comes from top load resonant voltage plus voltage induced by primary.  Resonant voltage (secondary current times secondary impedance) should be roughly evenly distributed along secondary height.  Primary-induced voltage is concentrated adjacent primary.  With lower pole frequency operation, those two voltages are roughly in phase, so add.  At upper pole frequency, voltages are roughly out of phase, so subtract.  Upper pole operation reduces volts/cm at bottom.  Though for a given top load voltage, volts/cm for upper part of coil is higher.  My experience with racing sparks have always been on lower portion of secondary, so presumably during initial part of enable before secondary frequency drops below primary frequency.

If diameters are all increased proportionately (primary and secondary and primary tube diameter and primary turn pitch), coupling increases without increasing volts/cm.  Primary induces secondary voltage farther up secondary (more coupling) without changing volts/cm (without changing coupling to each cm of secondary).  If JavaTC is accurate enough for odd shapes, you could check coupling for short secondaries (say 5cm) in sections from starting at primary to 5cm above, 10cm above, etc.  If coupling to bottom 5cm is higher than for your existing coil, then I'd be concerned about racing sparks.  (Higher induced volts/cm presuming same primary and secondary impedances.)  That may be an issue with your proposed design since primary diameters didn't increase in proportion to secondary.

Hope above makes some sense.  Everyone is of course welcome to question my thinking.  There could be some flaw in my logic and/or some critical factor I'm missing.

I have some 19 inch diameter fiberglass pipe

I think fiberglass is likely to have higher dielectric loss than most other pipe materials.  Dielectric loss may not be a significant factor, however.

You already have a very impressive coil!  Further limit pushing is always fun.  But you do need to be prepared for the possibility of an expensive failure.  (I spent well over 1000 hours, perhaps closer to 2000, with my failed attempt to add MMC capacitance as arc grew in order to keep primary in tune with secondary.  Probably close to $1000 of parts in addition to the time.)

[Benjamin Lockhart]

If diameters are all increased proportionately (primary and secondary and primary tube diameter and primary turn pitch), coupling increases without increasing volts/cm.  Primary induces secondary voltage farther up secondary (more coupling) without changing volts/cm (without changing coupling to each cm of secondary).  If JavaTC is accurate enough for odd shapes, you could check coupling for short secondaries (say 5cm) in sections from starting at primary to 5cm above, 10cm above, etc.  If coupling to bottom 5cm is higher than for your existing coil, then I'd be concerned about racing sparks.  (Higher induced volts/cm presuming same primary and secondary impedances.)  That may be an issue with your proposed design since primary diameters didn't increase in proportion to secondary.

Yes what you're saying makes sense, thanks.

I looked at this in javatc and you're right, the coupling to the lowest 5cm section is higher with this design because the primary is not proportionately wider as well. Making the primary wider to match results in about 0.22k instead of 0.25k.
That still a lot more but I don't think the expense and hassle of setting up such a large secondary would be worth it just for better energy transfer probably without very much spark length improvement.
Freewheeling control and further primary detuning accomplishes a similar thing. The major downside of this is that midi suffers because shorter on-times make almost no spark.

I guess the main benefit to higher coupling with a fatter coil would be reduced V/turn stress when heavily detuned because the spark load brings the top voltage down sooner.

My experience so far is that specifically for good music playback, big toploads and high coupling are always better. Both for low and high impedance primaries.
A very nice side effect of the shielding provided by the large toroid is almost no strike rail hits.

[Anders Mikkelsen]

I have some 19 inch diameter fiberglass pipe

I think fiberglass is likely to have higher dielectric loss than most other pipe materials.  Dielectric loss may not be a significant factor, however.

It depends on the resin used in the FG pipe and what it is compared to. I think epoxy resin is worse than polyester for example, but good data is hard to find. Both of them are worse than PP and PE, but PVC is not amazing in terms of dielectric loss tangent either. The clear advantage of FG pipe is the temperature rating and lower thermal expansion coefficient, but this is more of a factor for QCW type coils and higher duty cycle opetation.

It would be interesting to do a deeper investigation into how the secondary voltage distribution is affected by transmission line effects and higher order modes, these should be two sides of the same coin, one induced by the harmonic content of the drive waveform and the other one induced by the rapid change of topload voltage during breakout and ground arcing. A good starting point could be to gather info on the conditions under which it happens and where the breakdown happens. I've personally witnessed and confirmed primary-secondary flashovers from higher order mode excitation in QCW type coils, and recently had this confirmed by another experimenter who observed the effect I documented. For coils with lower coupling, transmission line whiplash is likely a more dominant effect.

[Mads Barnkob]

I went through a lot of plastic/varnish manufactures datasheets and different testing papers to gather the tangent of loss information for plastic types and varnish / epoxy. I did unfortunately not write down any sources at that time, so it is what it is, but can all be found here: https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/secondary-coil/ scroll down to "secondary coil form material" and "secondary coil with varnish"

[Benjamin Lockhart]

I ran the coil again. Sorry for the bad video, I was doing a show and more focused on running it. https://www.youtube.com/shorts/-y9MhDSRa7k

I lowered the secondary half an inch nearly to the top of the primary tubing and took it up to 9ms at 30Hz and now I'm getting ground strikes nearly all the time.
Unfortunately at 10Hz I started to see racing sparks (secondary is still okay) so I'm at the voltage limit of this secondary coil.

I believe I can still get better performance with higher coupling and further primary detuning, but since the secondary is at it's limit I'll just have to wind a bigger one.

Also, I ran the coil hard enough to trip the 50A breaker several times, and at the end of the last song a water cooling hose exploded and shot water all into the bridge! The breaker tripped a split second earlier and everything still works thankfully.

I think my small pump doesn't pump enough water to cool the primary as fast as it heats and I started to get steam pressure in the line. I've never built water cooling for anything else so I didn't allow for an over pressure situation!

This is a note to other builders to think about what happens if the water starts to boil.

[davekni]

Unfortunately at 10Hz I started to see racing sparks (secondary is still okay) so I'm at the voltage limit of this secondary coil.

Secondary isn't only factor in racing sparks.  Larger diameter secondary would be best way to improve performance.  However, you could push existing secondary a bit by changing primary to conical and a bit larger inside diameter.  That allows increasing coupling some without racing sparks.  Larger diameter primary reduces secondary induced volts/turn.  Conical allows larger diameter while maintaining coupling.