My 160 (?)mm sstc

[oneKone]

I tried a secondary CT...... It kinda worked but not very well. So I'm just gonna keep the antenna. I added a larger topload and had some internal arcing between the topload and the internal primary, also the internal primary to the secondary.

Luckily nothing got damaged so I'm going to rewind (with different wire) both primaries and add 3mm corflute with giving everything a coat of polyurethane. Hopefully this will fix the problem, but I suppose I could just have too much topload capacitance now.

[oneKone]

I rewound both primaries with 5 turns each and gave less coupling by adding a 3mm air gap. I'm going to have to switch to active cooling as the heatsinks I have are simply not big enough. I'm now getting over 700mm sparks to free air. It's almost deafening loud indoors.

[Mads Barnkob]

Congratulations!

I think this is the best performing SSTC to its size that I have ever seen.

Do you have more detailed pictures of the double primary construction?

[oneKone]

Congratulations!

I think this is the best performing SSTC to its size that I have ever seen.

Do you have more detailed pictures of the double primary construction?

Cheers.
Yep I'll grab some pictures for you tonight.
I'm going to have to use bigger heatsinks because the ones I have a too small. So that gives me the opportunity to change the layout of the inverter and make an MDF case so it's more easily accessible.

[oneKone]

In the name of a better setup I decided to redesign the full bridge. I made 2 half bridge pcb modules, I was hoping this was going to be straight forward but I'm having some problems. Every now and then an igbt would die (even on low voltages) so I'm wondering if there's something wrong with my design.


I'm not sure if there's something I'm missing here but to me it should be working. The only thing that has changed is that the driver is now sitting besides the secondary rather than under it.
Also mads, here are the primaries. Nothing special really....

[oneKone]

In the pursuit of finding the problem I killed about 8 igbts and ended up killing my driver board (the power supply for it decided to deliver full 20v into it). I rewound the gdt triple checked all the phasing, one thing I did note that one of the igbts wasn't getting warm when I tested the driver as a flyback driver. I tested both modules as half bridges then in full bridge with both configurations working perfectly.

At one last attempt I'll isolation mill the inverter section as the same layout as I originally had it on a perf board. Im honestly thinking if I can't get it working I'll just switch it to a drsstc..... The drivers on AliExpress while questionable at first glance but are cheap enough to give a go.

One thing I will ask is I'm driving my gate at 20v peak to peak, so I'm going to say 10v + and -. When people talk about gate voltage are the meaning peak to peak or +/- ?

Cheers

[Mads Barnkob]

It seems you have a few bits and pieces of copper trace/solder between the traces, maybe you have something that gives full voltage into the gate and kills it that way. But that does not explain that you already tested them beforehand and it worked.

If you have wrong phasing or defective switches, it would normally show as high current draw from low voltage and just ever increasing, the variac might also give some dunk, dunk, dunk noises that fit the interrupter signal.

If you provide your drivers with +20VDC, you will be driving the MOSFETs/IGBTs through the GDT with +/- 20VDC

[oneKone]

It seems you have a few bits and pieces of copper trace/solder between the traces, maybe you have something that gives full voltage into the gate and kills it that way. But that does not explain that you already tested them beforehand and it worked.

If you have wrong phasing or defective switches, it would normally show as high current draw from low voltage and just ever increasing, the variac might also give some dunk, dunk, dunk noises that fit the interrupter signal.

If you provide your drivers with +20VDC, you will be driving the MOSFETs/IGBTs through the GDT with +/- 20VDC

Cheers,
The stuff between the traces is just flux from the solder, I wonder if this could have an affect on it.

Also (just to get it through my head!) As you've stated with a drive voltage of 20v the square wave on the oscilloscope will read 40v peak to peak?

Cheers again

[Mads Barnkob]

If you use a GDT then you have implemented bipolar drive and it will drive it positive to the supply voltage and negative to the supply voltage.

I recommend that you take a look at these two sites: http://thedatastream.4hv.org/gdt_index.htm and http://www.richieburnett.co.uk/temp/gdt/gdt2.html

This was the best screenshot that I could find with a short notice

[oneKone]

Cheers Mads, I now have a better understanding. I've realised where some of my confusion came from was inputting a square wave into my driver and looking at the output voltage. It was only when I actually ran the coil with the oscilloscope connected to a spare output on the gdt that I could observe what was going on.

One other problem became apparent. My analogue oscilloscope had a hard time displaying the output of the gdt when the Tesla coil is in operation (when connected to a spare channel on the gdt). When compared to a screenshot like yours mine is (while readable) much more difficult to make out. So I might start looking for a DSO

[Mads Barnkob]

Cheers Mads, I now have a better understanding. I've realised where some of my confusion came from was inputting a square wave into my driver and looking at the output voltage. It was only when I actually ran the coil with the oscilloscope connected to a spare output on the gdt that I could observe what was going on.

One other problem became apparent. My analogue oscilloscope had a hard time displaying the output of the gdt when the Tesla coil is in operation (when connected to a spare channel on the gdt). When compared to a screenshot like yours mine is (while readable) much more difficult to make out. So I might start looking for a DSO

A DSO and differential probes certainly makes life easier, but it does also take an investment to have good tools

I can recommend the Tektronix P5200 if you can find a used to a good price. For oscilloscope the DS1054Z is still one of my favorites when it comes to features/price ratio. A good entry level scope, that I still use myself.

[oneKone]

Cheers for the recommendations. I was looking at something like a Hantek DSO5102P. Naturally I haven't looked at the features yet. Although the hantek comes in at half the price I'll take your recommendation for the rigol.

Luckily I have overtime at the moment!

[Mads Barnkob]

Cheers for the recommendations. I was looking at something like a Hantek DSO5102P. Naturally I haven't looked at the features yet. Although the hantek comes in at half the price I'll take your recommendation for the rigol.

Luckily I have overtime at the moment!

I think this covers most of the points when considering those two scopes against each other: https://www.eevblog.com/forum/testgear/hantek-dso5102-vs-rigol-ds1102-vs-rigol-ds1054/

[Hydron]

Just had another look through this thread - great results for a SSTC!

Regarding the scope question, the cheap options all have their pros/cons, but you have a much better choice than a few years ago, when the only affordable option was used (and used stuff is generally overpriced outside of the USA). I had a DS1054Z until I sold it to a colleague yesterday, it had various annoying aspects but was amazing value for money, and the only reason it's gone is that I bought a much more expensive Rhode&Schwarz unit and found I wasn't using the Rigol. If I were to buy another scope in this class I'd check out the Siglent options too, also maybe Micsig if portability is desirable.

You also mentioned using Chinese 60N60 IGBTs (I assume FGH60N60SMDs) - how were your experiences with these? I've always assumed that the likelihood of getting fakes is very high, but the prices is very tempting (my new coil needs 16 devices!)

[oneKone]

Just had another look through this thread - great results for a SSTC!

Regarding the scope question, the cheap options all have their pros/cons, but you have a much better choice than a few years ago, when the only affordable option was used (and used stuff is generally overpriced outside of the USA). I had a DS1054Z until I sold it to a colleague yesterday, it had various annoying aspects but was amazing value for money, and the only reason it's gone is that I bought a much more expensive Rhode&Schwarz unit and found I wasn't using the Rigol. If I were to buy another scope in this class I'd check out the Siglent options too, also maybe Micsig if portability is desirable.

You also mentioned using Chinese 60N60 IGBTs (I assume FGH60N60SMDs) - how were your experiences with these? I've always assumed that the likelihood of getting fakes is very high, but the prices is very tempting (my new coil needs 16 devices!)

cheers for the recommendations, it will give me more to think about.  as to what IGBTs i was using it was the FGH60N60SMDs. they worked well for what i needed them to, i did have some problems but this was to my own error. I've noticed on coilsmiths (facebook page) that a few people have been using IGBTs from aliexpress with nothing bad to say. for the price they are sold for its worth a shot.

[oneKone]

Ok..... So I've been lurking lately but finally got some work on the coil done. It now has a fullbridge of ixgn60n60. But as always I'm having little niggles... The first problem is that I'm having interference problems with the smps I have onboard. With the last case I had the power supply external so this wasn't a problem. The case of the smps is grounded but when the coil is wound up the fan can be heard boging down in rpm (although as I think about it maybe it's just the fan?)

The second problem is I'm getting heaps of racing sparks. It has 3 coats of polyurethane but I don't think it's the best. I remember reading on 4hv.org someone saying they used floor varnish (still water based polyurethane). Hopefully I can get some recommendations as what polyurethane people are using here.

[Mads Barnkob]

I think that your sphere topload might be a part of the problem with racing sparks, do you have a similar toroid shape that you could try with?

The electromagnetic field around a sphere is quite different from the one around a toroid and the result could be racing sparks. A sphere has its closest field lines going down through the middle of your secondary, whereas the correctly sized toroid would have them going down the sides of the secondary coil.

Another possibility is too high coupling between primary and secondary coils. That is well known from DRSSTC territory

[oneKone]

I think I'm done..... Killed the ixgn60n60 fullbridge on 100v. Had tvs/gate diodes on each igbt and still died.

I'm totally stunned how something can work almost flawlessly then then next minute never be the same.

[Mads Barnkob]

The frustration is real when dealing with loosely controlled power electronics, not much has to be wrong before something dies

I do however want to acknowledge, once again, that you had some really good results with long sparks and are one of few to have made a coil with both external and internal primary coils.

Does this mean that you are moving on full time to DRSSTC and the Chinese drivers?

[oneKone]

The frustration is real when dealing with loosely controlled power electronics, not much has to be wrong before something dies

I do however want to acknowledge, once again, that you had some really good results with long sparks and are one of few to have made a coil with both external and internal primary coils.

Does this mean that you are moving on full time to DRSSTC and the Chinese drivers?

cheers for the compliments. I think i'm going to persevere with this coil, if i don't it will always bug me! but after this coil i will be moving to drsstc.

After basically having a bitch about killing the igbts i took a few days to rethink what the problem might be. I knew the driver was ok and that (although with some ringing) gdt was giving a good output so that left one last thing to check....... the gate resistors. I was using cheap, and i'm talking dirt cheap chinese resistors. I had four of them in parallel because i didn't have the correct value. three of the resistor bundles were reading correct @ 6ohm but one was reading about 7k7..... The individual resistors gave a reading of 9k, 405k, 0.6 Megaohm and 1.12 Megaohm.

After knowing this i made a fullbridge of irfp460 and it seems to be running ok, although current limited. So tomorrow i'll head to my local electronics shop and buy the parts necessary for a higher power run, hopefully if everything goes ok i'll order another set of the 60n60 mini blocks and test everything before going to higher voltages.