DRSSTC Questions

[alan sailer]

I am playing around with making another small DRSSTC. I have two reasons for this. First, I enjoy the mechanical
challenges of making a coil. I don't have a fancy machine shop or 3D printer and I like figuring out how to make
parts with a table saw, belt sander and drill press.

Second, like many coil makers, I would love to have a two inch high coil that shoots out twenty foot sparks.
My last coil is about three feet high (including electronics) and has about two foot sparks. Not my dream coil.

A while back I ran across Arc Attacks commercial "Thundermouse". It's about 18 inches high and puts out 24 inch sparks.
That's more like it.

The most interesting aspect of the coil to me is the curved primary.



I have  few questions about this design. First I assume the curved primary allows higher coupling while keeping
the chances of a secondary flash-over under control. Is this true?

The second question is the coupling of this design. According to the web-page it's 0.308 which I understand
is very high for a DRSSTC. I tried modeling the coil in JAVATC using a helical winding (ie bottom radius smaller
than the upper radius) and got 0.311 coupling. I also tried modeling the primary as a cylindrical coil
with the radius the average of the curved primary and again got a similar coupling factor.

So I guess the question is, is a DRRSTC with such high coupling realistic?

Cheers.

[davekni]

So I guess the question is, is a DRRSTC with such high coupling realistic?

Short answer:  Yes, I think so.
AFAIK, DRSSTCs are better with as high coupling as possible without creating racing sparks along lower part of secondary.  Low secondary aspect ratios (height/diameter) allow higher coupling without racing sparks.  Low aspect ratios are common for short coils such as this.  I think I've seen several such small DRSSTC examples with similar coupling on the forum, but don't recall any specific ones to mention.

[alan sailer]

David,

Thanks for the input. I did some searching to find examples but I don't have the best talent for picking
search terms that give me the results I want. Since you believe examples of short, high coupling DRSSTC are in here
somewhere I'll try again later.

Cheers.

[flyingperson23]

Another physically similar coil is kaizer's drsstc 4. Looking at the thundermouse's page, it lists on time up to 20ms. This is more likely why its max arc length is so good. It also claims it can run as a QCW coil, which needs very high coupling compared to standard drsstc. I'm not suggesting against a short high coupling coil, but if you're under the assumption that the coupling/toroidal primary are the only reason for its arc length, you may be disappointed. Florian delmas and frederick hamilton both have long pulse drsstcs on youtube with similar arc length:secondary ratios with more standard drsstc coupling and flat primaries.

[alan sailer]

Flyingperson,

Thanks for the inputs. I am willing to experience mild disappointment. I'm not totally fixated on arc length which is a good thing.
All of my coils are based on other peoples designs and usually get about 70-80% of their performance.  I think this is because I
am a poor tuner, too afraid of killing FETs.

I'll check out those two builders.

David,

I tried another search term last night and ran across OneKone (a designer I have copied before). He has a short coil with high
coupling that I plan to study and compare to the ThunderMouse.

Cheers.

[davekni]

Looking at the thundermouse's page, it lists on time up to 20ms. This is more likely why its max arc length is so good.

You may not need 20ms, but long on times match well with high coupling.  The high coupling allows primary-to-secondary energy transfer for longer (over a wider frequency range as arc loading reduces secondary frequency).  That's the key advantage of high coupling.

[TMaxElectronics]

You might be interested in the small coil project I've been working at: https://highvoltageforum.net/index.php?topic=2394.0
I'm in the process of open sourcing it too, including the mechanical design.

It uses a standard flat primary without a high coupling coefficient but still gets 60" sparks from a 12" secondary (total coil height is around 24") at 1ms ontime (as you can see here: https://www.youtube.com/shorts/LmHMwYoh9tM). The biggest issue that I've been having with the current setup is primary strikes (no doubt due to the flat primary ) which are so energetic that they ionize the air in the space between the strikerail and the outer most winding and start an arc there too. I'm currently working on reducing the footprint of the secondary as well as adding a current ramp to the UD3 even in normal drsstc mode which apparently helps with starting the arc at the breakout instead of at some random location around the topload (both things that the at least the thundermouse already has).

[alan sailer]

TMax,

That is awesome. I'll definitely check it out. I'm sure to learn something even if I don't try to clone it.

Cheers.

[flyingperson23]

60" sparks from a 12" secondary

Can it sustain that? Does it not cook the coil running at 20kW?

[TMaxElectronics]

Can it sustain that? Does it not cook the coil running at 20kW?

The issue is the power input. That can't really handle 20kw constantly. Running the coil at a slightly lower power level (10kw or so) is perfectly fine in terms of thermals though. I've also run the coil at low power (2-3kw) without the fan on for a few minutes without the temperature going crazy. So it seems to be decently efficient.

[alan sailer]

Another question I've never had a good answer for. I have seen over the years many DRSSTC/QCW set-ups
 in what looks like the development/tuning stage. The primary leads going to the tank capacitor and the driver
can be quite long and the videos/photo of the set-up shows a great spark length.

Then a later picture shows the set-up all buttoned up with shorter leads still working great. I have always
assumed that the unit needs to be re-tuned after the primary is re-packaged. Or maybe people just roll up
the excess wire and call it good.

My goal has always been to get the coil working, longer leads and all, but not tuned for maximum output.
Then I package it nicely with a short as possible connections and do the final tuning.

It there a better way?

Cheers.

[Edi]

Hi guys
I need some help with GDT. Im using FGY75N60SMD with min treshold voltage of 3.5 V and max 6.5V. GE volage is rated fo +-20V. How should I wind my GDT if my UD2.7C driver has an output of +-24V for GDT.

[Hydron]

Just run the driver at a voltage lower than 24V, I'd suggest something in the 15-18V range. Then you can use a 1:1 CT for easiest and best results.

[Edi]

Then I need to adjust UVLO for lower value right? And I also calculated 15:6:6:6:6 ratio. Would that work? It would droop the voltage from 24 to 9.6V

[Hydron]

Best to keep the GDT ratio to 1:1:1:1:1, allows to wind using twisted pairs, with one wire of each pair a secondard, and all the other wires in parallel as a single primary, with very good coupling (low leakage inductance).

Then just run the UD2.7 off a lower voltage (change the regulator or just supply the correct voltage) and adjust the UVLO to suit (a few volts below the nominal drive voltage).

Unless you're pushing very large currents well beyond the IGBT rating (probably not a great idea with TO-247 parts) you won't need more than 15-18V. I've run the same parts up to at least their Icm rating of 225A at 300+kHz (though you need to be careful about hard-switching at those currents!)

[Unrealeous]

Hi guys
I need some help with GDT. Im using FGY75N60SMD with min treshold voltage of 3.5 V and max 6.5V. GE volage is rated fo +-20V. How should I wind my GDT if my UD2.7C driver has an output of +-24V for GDT.

Page 1 of the datasheet https://www.onsemi.jp/download/data-sheet/pdf/fgy75n60smd-d.pdf shows Vges +/- 20 with Vges transient +- 30

This looks like a pretty normal IGBT when it comes to GE voltage.   I think you might be getting confused with the Vge (threashold) voltage of 6.5 max on page 2.  The threshold voltage is the minimum voltage to turn the IGBT on, but will not give you it's maximum current.  You want to drive at Vges at +/-20V.



As Hydron says, I would use a one:one ratio and using a 30v zenner between gate and emitter, like most designs out there (see loneoceans, kaiserpower etc).    You can read about why everyone drives at slightly over the rated voltage (24V instead of 20V) at https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/

[flyingperson23]

Many IGBTs are rated for 20v Vge max. DRSSTCs usually run IGBTs past their rated current limits, so it helps to run them past their rated Vge limit as well. According to https://www.infineon.com/dgdl/Infineon-IGBT_Characteristics-AN-v01_00-EN.pdf?fileId=5546d462533600a40153559f8d921224, dielectric rupture happens at ~80v, so anything <30v is certainly fine. Steve ward ran his IGBTs at 30v. Yours will be fine on 24v, but if you're worried, many laptop charger bricks are 19v. Use that or something similar for input to the UD and adjust the UVLO accordingly. Often cat5 is used to wind GDTs as it's readily available and has twisted pairs for minimal leakage inductance.

[alan sailer]

So I have gradually gotten the parts together and have run the coil a few times. I'm having a problem
with flash-overs from the bottom of the top-load to the strike rail. One of the features that is puzzling
to me is that at higher rep-rates (~300Hz) there is no flashover but at low pulse rates (~40Hz) I'm
getting flash-overs.

Other than raising the top-load I'm not sure what to do about this.

Cheers.

[flyingperson23]

Try a different breakout point, maybe just a copper wire?
Other than that, I can't think of anything besides smoothing any rough edges or changing the coil geometry. On one of my coils, when the coupling was too high I'd get arcing over at <10hz that went away with higher bps.

[alan sailer]

Thank you for the reply. It sounds like your coil had the same issue.

The coupling on this coil is very high, almost QCW levels. I can't model it directly using JVATC but using two methods
the coupling is about 0.3.

I'll try another breakout. I am assuming sharper is best.

Cheers.