Benjamin's DRSSTC 2 in progress

[Benjamin Lockhart]

Thanks Dave, I'll do that

[thedoc298]

Very nice looking gear, I was wondering on your primary, it looks like your connected to the outer and the inner. How are you moving the tap, as I don't see any.

[Benjamin Lockhart]

You can see the tap point here.


I've mostly finished rebuilding the coil now. Hopefully this aluminum block will be enough heatsinking for the rectifiers. Edit: it seems to be enough.








Making things fit.


I put a few extra CTs on the bridge outputs to check current sharing and it appears to be identical on 2 separate bridge outputs!

[Benjamin Lockhart]

Hi all, Here's an update on this project.

First light was making 7 foot sparks at 100us and 240V in with minimal tuning, but a primary strike and flashover to the bottom secondary turn seems to have killed the FPGA in the UD+.

I suspect that equidrive was a bad idea and the high voltage on the inner primary turn could have caused a spike on the ground line during flashover, but I really don't know.

I've since replaced the driver with a spare for now, reconfigured the MMC for higher primary impedance (now 0.296uF instead of 0.67uF) and no split MMC.
I simply paralleled the bridge outputs.
I also raised the secondary an inch to lower coupling to about .142


Here's the new test run at only half input voltage 340V bus, and 500us on-time. It has hit 7 feet now at only 120V in and trips the 20A breaker quickly. There is 240V 50A available  for full power.
I'd like to build a better toroid with less unwanted breakouts before trying 240V in.

[flyingperson23]

.296uF seems small enough to really limit your current. That also probably makes it sound fuzzier at low bus voltages. I think it'll sound a lot better at full power, especially if you turn up the duty cycle.

[Benjamin Lockhart]

Yes it is, I might change it to 0.44uF with a larger toroid. The initial current ramp up gets to 750A and the steady state current when I turn it up to 2ms is about 500A. This is just at 340V bus.

[Benjamin Lockhart]

I added 2 more primary turns so I can experiment with a high impedance primary and larger toroid.
I'm still building the toroid at the moment.

[Benjamin Lockhart]

I've made some progress on a new 10x48 inch toroid, but if I were to do this over I'd buy aluminum tubing for a multi ring toroid instead.

It's made up of 60 cardboard strips hot glued into plywood and foam disks, and I'm going to wrap the surface with cloth soaked in glue to cover the bumps and gaps kind of like fiberglassing, then sand it down to make a smooth surface for aluminum tape.

[Mads Barnkob]

Now it's getting real, the physical form taking up all the remaining space in your workshop. The size of your monster is slowly being realized. "Where can I even put these parts..." "Where can I even run such a large coil..."

Good work and ingenuity on the topload. I made my topload too hard underneath the aluminium tape, with automobile-filler. A topload is the part most prone to transport and storage damage, so it actually needs to be able to take a bit of abuse. It's no worse, than scratches and ripped tape, is fixed with new tape and flattened out with a plastic card of short.

[Benjamin Lockhart]

Thanks Mads, Do you think I'd be better off using something softer like duct tape, rather than cloth and glue to smooth out the surface before aluminum tape?

Yes it's getting real now! It's quite big. Thankfully I have space to run it right in front of the shop.

[Mads Barnkob]

Thanks Mads, Do you think I'd be better off using something softer like duct tape, rather than cloth and glue to smooth out the surface before aluminum tape?

Yes it's getting real now! It's quite big. Thankfully I have space to run it right in front of the shop.

I am not sure what the best approach is, but cloth and glue sounds semi-soft. Aluminium tape is not going to hide any bumps or imperfections, so it will take on the pattern of the layer underneath. It's more a appearance issue than an electrical issue. As long as you can smooth out any spikes that would cause corona, its all fine.

It's only going to look nice up until the first time you drop it I cracked my DRSSTC1 topload from a drop and bumped my large ring topload, so its not looking completely circular anymore. Noone sees this but me, so it takes a builder to know it.

Everything also looks nicer on pictures than in real life.

[Benjamin Lockhart]

I decided to use the cloth and got it glued on.
I'll coat this with either more glue or varnish to make it stronger and hard enough.

[Benjamin Lockhart]

I finished the toroid and ran a test of the system!

The cloth and wood glue coating worked well.




I also made a water tank from polycarbonate for the primary cooling loop. You can also see my unfinished tesla guitar interface in the back!



Here's the whole coil standing outside.


And the tests. This is with the 0.296uF MMC and I'm not even hitting the current limit at 940A so I may need to increase capacitance.
Also only 10% primary detuning. I think I should detune the primary way more to take full advantage of this arrangement.

I had some explosions and racing sparks, and my run was cut short by snow blowing in. Also my scope wasn't set right so I don't know the exact primary current.

The biggest issue was the outer primary turn flashing over to the strike rail which caused the power plug to flash over (dumping the tank circuit between ground and the power feed line!) and melt some steel there.

Also, the sharp ends of the strike rail were initiating racing sparks. as can be seen here. Even though coupling is only 0.142
The strike rail has more clearance from the primary than my DRSSTC3 so I thought it would be ok, but I think I need to remake the strike rail with more clearance and no sharp ends.
An extra coat of epoxy on the secondary would also be a good idea even though it survived.

/>

/>
Edit: Here is a picture of corona from the sharp end of the strike rail. I definitely need to do something about this!

[Benjamin Lockhart]

I'm not sure what to do regarding the strike rail.

In any case I'm going to remove the existing one, and either make a new one with a larger diameter, maybe spaced lower down in order to be further away from the secondary.

Also maybe a large acrylic disk over the primary.
Or maybe I could get away with no strike rail (I have a 0.1uF bypass cap from the negative rail to ground.)

Edit: Made a new larger strike rail and all is well now.

[davekni]

Also, the sharp ends of the strike rail were initiating racing sparks. as can be seen here. Even though coupling is only 0.142
The strike rail has more clearance from the primary than my DRSSTC3 so I thought it would be ok, but I think I need to remake the strike rail with more clearance and no sharp ends.

Also maybe a large acrylic disk over the primary.

Or maybe I could get away with no strike rail (I have a 0.1uF bypass cap from the negative rail to ground.)

I had the same issue with strike rail arcs triggering racing sparks when I first built my DRSSTC even though my primary was covered with polycarbonate plates.  Made larger diameter strike rails using larger diameter tubing (19mm).  Solved the issue for several years.  Recently a top-load defect on bottom of toroid caused many arcs to strike rail, ionizing enough air to eventually start racing sparks.  Fixed my top load now.

Electronics would hopefully survive strike to primary coil.  Initial hit from top load is often near peak top load voltage.  Top rapidly discharges causing short high-current pulse.  Electronics survival depends on path of that short high-current pulse and its resulting inductive drop through wiring.

[Benjamin Lockhart]

Edit: I just found a video still of a single shot, 8ms, 13.5 foot streamer


I made a larger strike rail with ball ends and gave the secondary a second coat of epoxy, tuned the primary 22% low and now I've hit 4 meter sparks! just over 13 feet.

Mads I think this beats your current record, and from an 850mm tall secondary. That's 4.7 times the winding length!

The current limit is set to 940A, and peaks just over 1000A. Current ramp up is pretty slow with this 14 ohm surge impedance. My theory is that slower current ramp up is easier on the secondary volts/turn and allows more power to be pushed through the system without failure compared to low impedance coils.

I detuned the primary incrementally at 15% then 19% then 22% low and the sparks kept getting bigger. I still have enough turns to go 30% low.

I also noticed something interesting, there is a notch in the primary current and it moves further out as I tune the primary lower. I think that's the point where the secondary comes into tune and pulls some energy from the primary.

This suggests I could still detune more since it comes back out of tune again? What are your thoughts on this? 22% seems like a lot already.

Also my coupling is 0.142 with the secondary raised above the primary 3/4 inch. I don't know if I dare increase it.

Scope is set 500A/div, 100us/div.

I'd like to know what to try to push this thing harder, more coupling? higher primary current? bigger MMC?
I'm pretty happy about hitting 4 meters though.
Thanks, Benjamin

/>

/>

/>

/>



Primary 14% low


Primary 19% low


Primary 22% low and lower pulse rate

[davekni]

That's interesting that the conical primary can have higher coupling without racing sparks.

Wow, very impressive coil!  48" top load is a big asset for making long arcs.  Increases arc length for a given amount of secondary frequency change due to arc capacitance.  Much better than my 34" top.  Freewheeling looks to be helping a lot too, though I have no personal experience with such.  (My IGBT diodes wouldn't handle that much average current.)

My theory is that slower current ramp up is easier on the secondary volts/turn and allows more power to be pushed through the system without failure compared to low impedance coils.

Expect your theory is good, especially for freewheeling control.

I also noticed something interesting, there is a notch in the primary current and it moves further out as I tune the primary lower. I think that's the point where the secondary comes into tune and pulls some energy from the primary.

Yes, appears so.  The small amplitude modulation due to energy moving primary-to-secondary and back show up separately as about one cycle per division.  Modulation amplitude is limited there due to your high impedance primary, with current ramp-up time similar to modulation period.
If you were to measure secondary bottom current, should show almost 180 degree phase shift relative to primary current, from before to after dip.  Operation switches from lower pole frequency to upper pole frequency as arc load moves secondary frequency below primary frequency.  (Frequency shift isn't obvious since pole frequencies are dropping during switch due to arc capacitance increasing.)

This suggests I could still detune more since it comes back out of tune again? What are your thoughts on this? 22% seems like a lot already.

Try more to see.  By 30% I'd guess dip won't occur, extended beyond your maximum enable pulse width.  With more detuning, initial top voltage is lower.  Eventually it will be too low to grow arc to the point where it is in tune.

Also my coupling is 0.142 with the secondary raised above the primary 3/4 inch. I don't know if I dare increase it.

I'd guess you could go at least a bit higher given your high impedance primary, at least to having secondary start at primary height.  Biggest risk to higher than that may be voltage between primary and secondary.  Of course racing sparks can quickly lead to destructive arcs if coil operation is continued for long.  Always hard to decide how much risk to take in pushing performance.  Given my recent experience with humidity (and Mads' previous comments about humidity causing racing sparks), perhaps you will want to leave some margin for different operating conditions.

[Benjamin Lockhart]

Yes, appears so.  The small amplitude modulation due to energy moving primary-to-secondary and back show up separately as about one cycle per division.  Modulation amplitude is limited there due to your high impedance primary, with current ramp-up time similar to modulation period.
If you were to measure secondary bottom current, should show almost 180 degree phase shift relative to primary current, from before to after dip.  Operation switches from lower pole frequency to upper pole frequency as arc load moves secondary frequency below primary frequency.  (Frequency shift isn't obvious since pole frequencies are dropping during switch due to arc capacitance increasing.)

That's actually the freewheeling driver skipping cycles. Here's a video with just a dummy load. Is it actually switching between upper and lower pole frequency?

[davekni]

That's actually the freewheeling driver skipping cycles.

Oh, yes, makes perfect sense.  Did look a bit fast for energy cycling primary to secondary.  Thank you for clarifying.

Is it actually switching between upper and lower pole frequency?

Very likely.  Switches across the primary current notch.  I've measured that on both my normal DRSSTC and my recent ZVS driven one, and seen it in a couple other coils based on scope plots posted on this forum.  I've also replicated that behavior in LTSpice simulation.  Haven't seen any counter-examples yet.  As I'd mentioned, primary frequency doesn't actually jump up much if at all, since pole frequencies are dropping across the switch.  Best way to see it is by scoping secondary bottom current (or top electrostatic field from probe "antenna" on opposite side from breakout point) along with primary current.  Measure relative phase before and after notch.  Your high-impedance primary should make that easy to see.  With low impedance primary and short enable times, much of the energy is oscillating between primary and secondary.  (In other words, both pole frequencies are present simultaneously.)  With your coil, upper pole frequency component has dropped to insignificant before notch.  Should make switch easy to see.

[Benjamin Lockhart]

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?