Finished my oneTesla TS! :-)

[V1-Hyper]

I used to build classic spark gap TCs back in the 1990s as a physics student.


Last year I restarted my beloved old-days hobby. This is my oneTesla TS, beside a cheap Ebay kit, playing Mission Impossible. I am well impressed by the out-of-the-box TS kit, its power-to-size ratio and pre-tuned condition
My only little gripe is that I get nowhere near the advertised 20" spark length but 11" max. Did I overlook something?

Watch on youtube.com

EDIT: I just measured the resonances of the primary tank and secondary and found a possible explanation for the 11" streamer limit; they are not well in resonance.
f_sec = 215.5 kHz
f_pri = 236.5 kHz
with C_pri = 65.8 nF and L_pri = 6.50 µH (directly measured) or 6.88 µH (deduced from f and C)

I read the Primary resonance should be about 10% below the Secondary. Given that, I'd need an optimum C_pri of about 100 nF in place of the 68 nF that came with the kit.
I have now ordered a 100 nF version and will report back if the new capacitor improves performance or not.

[V1-Hyper]

I have tested the TC with a C_pri = 100 nF and it made performance considerably worse. Maybe my frequency measurements were wrong.
I put back the original 68 nF and restored max streamer of 11.5” (I slightly improved the RF ground by connecting it to an earth rod)

I guess in order to fine-tune resonance one would need a variable HV capacitor.

[E.TexasTesla]

Please let us know if you find a better capacitor value.
Nice work on that conical coil. Did you run it with the breakout that low ?

[V1-Hyper]

Please let us know if you find a better capacitor value.
Nice work on that conical coil. Did you run it with the breakout that low ?

I can't remember munch detail, been decades ago. I think sparks were half a meter long? I possess 1 old photo in action:


Yes, I will post my findings for the TS's C_pri vs max streamer length very soon. Capacitors arriving today via post. This is my simple measurement setup with a sliding grounded counter-electrode and a ruler:

[V1-Hyper]

I have now varied the primary capacitance to find resonance...
To get my conclusion out of the way, before giving some details: Pri-Sec resonance with the original 68nF cap was already close to optimum, and thus, streamer length cannot be improved beyond the original 11.5" or 29 cm.

Here is my original cap (precisely 65.8 nF) and my home-made stack of parallel 10 nF caps of same make and model (Polypropylene series 940C)

I measured max spark length versus total C_pri (max duty cycle at 1000 Hz), then removed one small cap at a time, successively reducing C_pri. Resultant plot:

Interpolating my data points, the optimum C_pri value appears to be 62.5 nF.
The original 65.8 nF is therefore close to perfect. The reason why my curve for the parallel cap-stack is overall a bit lower has probably to do with slightly higher ohmic resistance that comes with my test arrangement.

The streamer-loaded resonance frequency of my secondary is therefore about 242.7 kHz.

I plan, however, to replace my original 68 nF primary cap by two parallel 33 nF caps to double the current pulse and dis/charge rate. I may even beef up the primary tank related PCB tracks, further reducing ESR. Maybe this will boost streamer length a little. For now, C_pri alternatives could be:

[Bayley]

Contact us and we'll send you the 4 oz primary we sell to demonstrators and museums, the usual use is to reduce primary heating during frequent use but the lower ESR should improve streamer performance as well.

[V1-Hyper]

Contact us and we'll send you the 4 oz primary we sell to demonstrators and museums, the usual use is to reduce primary heating during frequent use but the lower ESR should improve streamer performance as well.

Thank you, Bayley, much appreciated

[V1-Hyper]

Today my primary coil 4oz PCB arrived (thanks Marissa).
It's coil track measures 0.09 Ohm instead of 0.16 Ohm for the regular one shipped with the kit.

In order to take best advantage of the reduced ohmic resistance I also beefed up the remaining tracks of the primary tank:

I also swapped my primary cap as discussed in previous post, and replaced it for two parallel 33 nF versions for reduced ESR and greater dV/dt, I_peak and I_rms:


Bit sobering but maximum streamer length didn't really improve, now standing at 12". But I'll be leaving it at that.

I ran about 3 or 4 midi files when suddenly there was a massive BANG and my house fuse box tripped! The TS's fuse did not blow. Hence, there must've been a current flow to earth. The flashy short came from around the bridge rectifier:

I remove the rectifier to get a closer look. I suspect the resistor R5 just behind it overheated but will diagnose the components soon. A quick estimate of R5's and R6's power consumption gives 1.2 W (340²/100k), which they might not withstand for prolonged time. Whatever failed, I'll try to replace it with a stronger one.

[Bayley]

That is an...interesting failure mode. Sadly at this point the coil is modified enough from stock that I don't really have any advice to give, but if you have any interesting findings let us know.

[V1-Hyper]

After checking, I do indeed have some interesting findings:
All components are still fully intact, nothing failed or overheated. However....

Bizarrely, and after minutes of running midis at full power, the 350 VDC bus voltage somehow managed to jump across from one end of R5 (the one that's connected to the bridge rectifier's positive O/P) to the nearby top copper layer that connects with the bridge rectifier's negative O/P. Hence, the rectifier's output was instantly short-outed! The rectifier actually survived, but some copper layer vaporised as seen in the photo

Maybe there was a minute voltage spike in the mains line? I cannot see any other major voltage amplification there. (I checked my mains voltage is 245 VAC)
Would it seem that from a PCB design point of view the track gap right there is a little close to 'just wide enough'?

Best wishes