FIRST DRSSTC SKM400GB128D OCD SETTING

[Alex_1996]

Hello everyone. After having built two successful DRSSTC's with a full bridge of TO247 FGA60N65SMD IGBT's, I have built a new model with a full bridge of SKM400GB128D. Here I share some data:

•Secondary 36cm high and 10cm wide with 0.2mm diameter enameled copper wire. The whole secondary coated with epoxy resin.

•Toroid 35,5cm x 10,5cm made with aluminum rings. The discharge wire is 47cm long. I think it may be too long.

•Resonant frequency of the secondary with the toroid is about 155kHz.

•7-turn primary and 66nF/12kV MMC bank for a primary frequency of 142kHz at the tuning point, which is a detuning of about 8.38%.

•2 DC bus capacitors connected in voltage doubler with a total of 900V/1650uF (each 450V/3300uF).

•2 snubber capacitors of 1kV DC and 2.2uF each.

•1.5ke400CA bi-directional protection TVS diodes (two in series).

•Feedback CT ratio is 1/625 (25 turns x 25 turns)

Driver 2.7.The phase advance adjustment was performed correctly and successfully.

Here comes my doubt. The discharges, in my opinion, are a bit small. Understand me. I achieve some 80cm discharges when they hit a target on the ground with a voltage input on the full bridge of 362VDC (150VAC from my variac) and the OCD is set to 350A. The tip of the discharge needle is 90cm high (maybe a little high). My question is:

Can I raise the OCD value to 400A or higher? I don't know if this model of IGBT's can support 400A or higher running at a primary frequency of 142kHz. They are placed on a good heat sink. The power consumption with the normal interrupter operating mode (70-100uS / 560 bps) and in MIDI mode is between 600W and 800W, respectively. My goal is that the discharges can reach the ground without having to place an object that shortens the distance to help the lightning to reach the ground more easily.

I had thought about raising the OCD limit to 400A and maybe 400V DC bus power supply in order to achieve higher discharges that can reach the ground without any problem.

I attach photos and if any data is missing do not hesitate to tell me about it. I also attach a video of the operation.

Video: https://youtube.com/shorts/fBSI0B8LMx4?feature=share

Best regards and thanks for your help

[flyingperson23]

Yes, the transistors can handle 400A fine. SKM400 is a bit overkill for a coil that size. Usually drsstcs run their igbts at 3-4x nominal current, so skm400 with ~500A nominal rating would see 1500-2000A, so I doubt you could hurt the IGBTs with too much current even if you tried.

You can also turn the bus voltage way up, I don't see anything stopping you from going up to 800Vbus, although that would probably shorten your on time even more.

The tvs diodes are probably not recommended. Afaik the IGBTs have avalanching capability and external diodes can end up damaging them. You're unlikely to get enough voltage to damage components if you're using <600V bus anyways.

The discharge wire isn't too long, it's a little on the longer end but probably not enough to cause problems. I don't think the length matters a ton, although it can cause problems if it's too short.

Partly unrelated, but your topload looks really good. Do you have the 3d files?

[Alex_1996]

Thank you very much for your quick response.

I was concerned about running the IGBTs at 142kHz with 400A, because the SKM400GB128D are used in larger and therefore lower resonant frequency coils.

I am talking about 400A due to the fact that my beams at 350A are close to the ground, although they lack a little bit to impact fully against the ground, so I have estimated that with about 400A in the OCD they will definitely impact against the ground without any problem. Even, if I dare, to raise the OCD to 450A or more.

I have the 3D files on an old PC. I'm going to look for them

Best regards

[flyingperson23]

Bricks can switch pretty fast https://4hv.org/e107_plugins/forum/forum_viewtopic.php?85229
They seem to handle 350A fine, so a few hundred more amps should also be fine. You might have to turn up the on time to see results.

[Admiral Aaron Ravensdale]

Nice build!
Maybe your Variac is the limiter at the moment, I just figure out that may 6A 320V Variac was not able to deliver the current for my QCW coil which sucks in a high peak 3.5A!
The sparks stay 15-20cm shorter than with my new self build variac who can deliver 20A.
Maybe it was the pulsing of the coil and the variac was to slow to follow the single high peaks of the coil.

[Alex_1996]

Effectively, as I increase the supply voltage I have to decrease the t-on so that the OCD does not act so fast.

I don't think my 3kW variac is giving problems because it can handle up to 16A.

[Mads Barnkob]

You should use a thicker round rod for break out, to avoid so much corona discharges around it, this is energy lost from building the main spark longer.

You can easily set your OCD to 800A, those bricks can burn your secondary to the ground before IGBTs are sweating

Variacs act as a passive PFC for a DRSSTC, brings the PF from about 0.5 to 0.67

Bricks can switch pretty fast https://4hv.org/e107_plugins/forum/forum_viewtopic.php?85229
They seem to handle 350A fine, so a few hundred more amps should also be fine. You might have to turn up the on time to see results.

I got some scope shots of just how fast a CM600 brick can switch in resonant mode, its all in the IGBT article on DRSSTC design: https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/ this is the same article that you should read if you are in doubt about your maximum resonant frequency, peak current and power dissipation of your bricks. It can be calculated.

[Benjamin Lockhart]

I agree, you can turn the OCD way up.

[Alex_1996]

You should use a thicker round rod for break out, to avoid so much corona discharges around it, this is energy lost from building the main spark longer.

You can easily set your OCD to 800A, those bricks can burn your secondary to the ground before IGBTs are sweating

Variacs act as a passive PFC for a DRSSTC, brings the PF from about 0.5 to 0.67

Bricks can switch pretty fast https://4hv.org/e107_plugins/forum/forum_viewtopic.php?85229
They seem to handle 350A fine, so a few hundred more amps should also be fine. You might have to turn up the on time to see results.

I got some scope shots of just how fast a CM600 brick can switch in resonant mode, its all in the IGBT article on DRSSTC design: https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/ this is the same article that you should read if you are in doubt about your maximum resonant frequency, peak current and power dissipation of your bricks. It can be calculated.

Thank you very much for this information.

I didn't know that the thickness of the discharge needle influenced the final length of the discharges. It is certainly a very important piece of information

Also, I will use the guide you sent in the link to make some precise calculations on the maximum frequency of the IGBT's. I actually used this pair of SKM400GB128D in full bridge configuration because I found a bargain and they were well priced.

I think with a new discharge needle and raising the OCD current value I will get the discharges to the ground easily. I will share here the results when I do the tests.

Best regards

[Alex_1996]

I agree, you can turn the OCD way up.

Thank you for your response. I will try and share the results with you.

[Alex_1996]

Hello everyone!

I finally set the OCD to 350A and made a new discharge wire, this time, 10mm diameter and shorter in length.

¨Magically¨, the discharges started to reach the ground without any problem, so my main problem was solved. Now, the discharges reach 1.10 meters in length measured in a straight line. I only had to change the discharge wire and the problem was fixed.

Now, the coil runs with a supply voltage of 150VAC from my variac, which is rectified and doubled to a total of approximately 362VDC and the OCD set to 350A. These parameters are more than enough for the electrical discharges to reach the ground.

Fun fact. As I move the coil closer to a wall in my room, the discharges become more intense and violent, even increasing in length. I suppose that the wall must influence the resonant frequency of the secondary, causing the frequency of the primary and secondary to be tuned even further.

Thanks for the help. I attach a photo

[Admiral Aaron Ravensdale]

I only want to note that the SKM150 run stable at 400A in this coil!
So 3-4 times of the rating is still save...

https://highvoltageforum.net/index.php?topic=2873

[Mads Barnkob]

Fun fact. As I move the coil closer to a wall in my room, the discharges become more intense and violent, even increasing in length. I suppose that the wall must influence the resonant frequency of the secondary, causing the frequency of the primary and secondary to be tuned even further.

Great looking performance

This suggest that you can tune your primary lower yet, how many % is it detuned compared to the secondary? Somewhere around 5-10% detuning of primary circuit, is normal to account for the spark load of the secondary circuit.

[Alex_1996]

Hi, sorry for the delay. I've been a bit busy with work.

The detuning is about 9%. I have tried adjusting the primary somewhat better and I get a slight increase in discharge length, however, I think that is the most I will be able to achieve with the OCD at 350A and 362VDC bus power. Attached are some screenshots of the last video I recorded.

I have placed a 1m x 1m wire mesh grounded on the ground. Before, I had placed an aluminum plate of 50cm x 25 cm. What do you think is the best option for the lightning to hit the ground? * Note: I do not have a garden, I connect the aluminum plate as the metal mesh to the ground of the house.

Greetings and thank you very much!

[Benjamin Lockhart]

You should really set the OCD limit way above 350A, I would double it.

[flyingperson23]

As far as I can tell, the primary is insulated wire and the mmc is cheap induction heater caps, neither with significant cooling. The bridge can definitely take 1kA+ but the rest of the coil might not be able to.

[Alex_1996]

Hello!

Effectively, because I don't have much budget I save costs in some parts of the coil assembly.

What capacitors would you use for the MMC bank and what wire would you use for the primary winding in order to build a more robust system to support more amps?

Thank you very much

[Mads Barnkob]

A cheap MMC can be made from WIMA or KEMET MKP/PP/Polypropylene capacitors, avoid polyester film.: https://dk.rs-online.com/web/p/filmkondensatorer/1734195 Some 24'ish of these and you have a good MMC at 0.33uF, I forgot how many in series and parallel, maybe it was 3 strings of 8, or 4 strings of 6.

Primary winding, nothing beats copper tubing in coils, used for water and gas, so its widely used and therefore pretty cheap. 10 mm is standard for water. If you want smaller, there is 4 or 6 mm "brake line" copper tubing, again fairly affordable because of the wide use in cars.

[flyingperson23]

For mmc I'm a fan of dawncaps, they're a little sketchy but seem to work great and cost per amp is really good. I have a coil similarly sized to yours and I use 1 of these:

https://www.aliexpress.us/item/3256805356938958.html?spm=a2g0n.search-store.0.0.1006U8UDU8UDfU&browser_id=da6abc91876a4a029b5fb1f177997a7e&aff_trace_key=0cd93ca4228c45d6aef78be0b2713457-1720896888586-02113-UneMJZVf&aff_platform=msite&m_page_id=zyuhtulhccaridup190e57e42cc199fec31715751a&gclid=&pdp_npi=4%40dis%21USD%2117.77%2117.77%21%21%21128.75%21128.75%21%402101fb1917218371754772657e3894%2112000033476967120%21sea%21US%213080507328%21X&algo_pvid=9dabed1b-ddf4-4753-b7cd-92d455c73c16&_gl=1*y7yjkg*_gcl_au*NjI0NTM4MjkuMTcyMTgzNzE4MQ..*_ga*MTc1MTAwMjA5NzE0MzE0NC4xNzIwMTMwMjcyMjc5*_ga_VED1YSGNC7*MTcyMTgzNzE4MS4xLjEuMTcyMTgzNzQwNi42MC4wLjA.

If you want more current, you can put a few 20-40nF caps in parallel.

Copper tube is definitely the way to go for the primary, in the US you'd want to use either 1/4in or 3/8in tube.

[Alex_1996]

A cheap MMC can be made from WIMA or KEMET MKP/PP/Polypropylene capacitors, avoid polyester film.: https://dk.rs-online.com/web/p/filmkondensatorer/1734195 Some 24'ish of these and you have a good MMC at 0.33uF, I forgot how many in series and parallel, maybe it was 3 strings of 8, or 4 strings of 6.

Primary winding, nothing beats copper tubing in coils, used for water and gas, so its widely used and therefore pretty cheap. 10 mm is standard for water. If you want smaller, there is 4 or 6 mm "brake line" copper tubing, again fairly affordable because of the wide use in cars.

Sorry for the delay

I was traveling and it was impossible for me to reply to the messages.

Thank you very much for the information about the improvement in the quality of construction materials. No doubt, my next DRSSTC will be built with such materials to ensure maximum reliability and robustness in its operation.

Best regards and see you next time

[Alex_1996]

For mmc I'm a fan of dawncaps, they're a little sketchy but seem to work great and cost per amp is really good. I have a coil similarly sized to yours and I use 1 of these:

https://www.aliexpress.us/item/3256805356938958.html?spm=a2g0n.search-store.0.0.1006U8UDU8UDfU&browser_id=da6abc91876a4a029b5fb1f177997a7e&aff_trace_key=0cd93ca4228c45d6aef78be0b2713457-1720896888586-02113-UneMJZVf&aff_platform=msite&m_page_id=zyuhtulhccaridup190e57e42cc199fec31715751a&gclid=&pdp_npi=4%40dis%21USD%2117.77%2117.77%21%21%21128.75%21128.75%21%402101fb1917218371754772657e3894%2112000033476967120%21sea%21US%213080507328%21X&algo_pvid=9dabed1b-ddf4-4753-b7cd-92d455c73c16&_gl=1*y7yjkg*_gcl_au*NjI0NTM4MjkuMTcyMTgzNzE4MQ..*_ga*MTc1MTAwMjA5NzE0MzE0NC4xNzIwMTMwMjcyMjc5*_ga_VED1YSGNC7*MTcyMTgzNzE4MS4xLjEuMTcyMTgzNzQwNi42MC4wLjA.

If you want more current, you can put a few 20-40nF caps in parallel.

Copper tube is definitely the way to go for the primary, in the US you'd want to use either 1/4in or 3/8in tube.

Thank you very much

I keep this information for the construction of my next DRSSTC.

Best regards

[radbloke]

Hello everyone. After having built two successful DRSSTC's with a full bridge of TO247 FGA60N65SMD IGBT's, I have built a new model with a full bridge of SKM400GB128D. Here I share some data:

•Secondary 36cm high and 10cm wide with 0.2mm diameter enameled copper wire. The whole secondary coated with epoxy resin.

•Toroid 35,5cm x 10,5cm made with aluminum rings. The discharge wire is 47cm long. I think it may be too long.

•Resonant frequency of the secondary with the toroid is about 155kHz.

•7-turn primary and 66nF/12kV MMC bank for a primary frequency of 142kHz at the tuning point, which is a detuning of about 8.38%.

•2 DC bus capacitors connected in voltage doubler with a total of 900V/1650uF (each 450V/3300uF).

•2 snubber capacitors of 1kV DC and 2.2uF each.

•1.5ke400CA bi-directional protection TVS diodes (two in series).

•Feedback CT ratio is 1/625 (25 turns x 25 turns)

Driver 2.7.The phase advance adjustment was performed correctly and successfully.

Here comes my doubt. The discharges, in my opinion, are a bit small. Understand me. I achieve some 80cm discharges when they hit a target on the ground with a voltage input on the full bridge of 362VDC (150VAC from my variac) and the OCD is set to 350A. The tip of the discharge needle is 90cm high (maybe a little high). My question is:

Can I raise the OCD value to 400A or higher? I don't know if this model of IGBT's can support 400A or higher running at a primary frequency of 142kHz. They are placed on a good heat sink. The power consumption with the normal interrupter operating mode (70-100uS / 560 bps) and in MIDI mode is between 600W and 800W, respectively. My goal is that the discharges can reach the ground without having to place an object that shortens the distance to help the lightning to reach the ground more easily.

I had thought about raising the OCD limit to 400A and maybe 400V DC bus power supply in order to achieve higher discharges that can reach the ground without any problem.

I attach photos and if any data is missing do not hesitate to tell me about it. I also attach a video of the operation.

Video: https://youtube.com/shorts/fBSI0B8LMx4?feature=share

Best regards and thanks for your help

Here is the one that I built a few years back. It’s got exactly the same switches as yours i.e. the SKM 400 switches in a full bridge configuration. The OCD is set at 1000Apk.  I have dawncaps 375 nF bank - these get a little warm after long runs but work great. Using freewheeling with UD+
Fres is 52kHz with 31 inch x 6.5 inch secondary.

Here is a link to the coil running with 3/8”  copper tube breakout and 31 inch diam topload at 1000A OCD for the freewheel. The video files were too large to upload directly to this forum.

So you have plenty of room to increase your OCD to well above 350 A!!!

[Alex_1996]

Hello everyone. After having built two successful DRSSTC's with a full bridge of TO247 FGA60N65SMD IGBT's, I have built a new model with a full bridge of SKM400GB128D. Here I share some data:

•Secondary 36cm high and 10cm wide with 0.2mm diameter enameled copper wire. The whole secondary coated with epoxy resin.

•Toroid 35,5cm x 10,5cm made with aluminum rings. The discharge wire is 47cm long. I think it may be too long.

•Resonant frequency of the secondary with the toroid is about 155kHz.

•7-turn primary and 66nF/12kV MMC bank for a primary frequency of 142kHz at the tuning point, which is a detuning of about 8.38%.

•2 DC bus capacitors connected in voltage doubler with a total of 900V/1650uF (each 450V/3300uF).

•2 snubber capacitors of 1kV DC and 2.2uF each.

•1.5ke400CA bi-directional protection TVS diodes (two in series).

•Feedback CT ratio is 1/625 (25 turns x 25 turns)

Driver 2.7.The phase advance adjustment was performed correctly and successfully.

Here comes my doubt. The discharges, in my opinion, are a bit small. Understand me. I achieve some 80cm discharges when they hit a target on the ground with a voltage input on the full bridge of 362VDC (150VAC from my variac) and the OCD is set to 350A. The tip of the discharge needle is 90cm high (maybe a little high). My question is:

Can I raise the OCD value to 400A or higher? I don't know if this model of IGBT's can support 400A or higher running at a primary frequency of 142kHz. They are placed on a good heat sink. The power consumption with the normal interrupter operating mode (70-100uS / 560 bps) and in MIDI mode is between 600W and 800W, respectively. My goal is that the discharges can reach the ground without having to place an object that shortens the distance to help the lightning to reach the ground more easily.

I had thought about raising the OCD limit to 400A and maybe 400V DC bus power supply in order to achieve higher discharges that can reach the ground without any problem.

I attach photos and if any data is missing do not hesitate to tell me about it. I also attach a video of the operation.

Video: https://youtube.com/shorts/fBSI0B8LMx4?feature=share

Best regards and thanks for your help

Here is the one that I built a few years back. It’s got exactly the same switches as yours i.e. the SKM 400 switches in a full bridge configuration. The OCD is set at 1000Apk.  I have dawncaps 375 nF bank - these get a little warm after long runs but work great. Using freewheeling with UD+
Fres is 52kHz with 31 inch x 6.5 inch secondary.

Here is a link to the coil running with 3/8”  copper tube breakout and 31 inch diam topload at 1000A OCD for the freewheel. The video files were too large to upload directly to this forum.

So you have plenty of room to increase your OCD to well above 350 A!!!

/>

This DRSSTC presents an incredible result. The electrical discharges are impressive.

After what has been exposed in this forum thread, I am sure I can set the OCD above 350A. The only drawback I see is the resonant frequency of the primary, which is set at 142kHz. I don't know if that's too high a frequency for the SKM400's.

Also, maybe my small secondary won't handle the excess voltage from the increased current and could break.

However, for future models, I will build coils of a larger size to lower the resonance frequency and I will use higher quality MMC bank capacitors and copper tubing for the primary winding.

[Jj]

Very nice build!

you have already been given the good advice here,  but you could measure the switching time occurrence to check the performance. If you are staying close to zero current all is good. Switching losses are much less of an issue for low duty cycle operation.

Not sure how much more you can get from that secondary coil. Problems usually start when spark length is at 3x the secondary length for my DRSSTCs. Improvements can often be made, though it takes more effort and the returns get smaller.