Author Topic: Freewheeling diodes are burning (Now other problems) (Read 8576 times)

hip424 · « **on:** May 11, 2020, 02:12:26 PM »

Hello,
I've built my first SSTC based on Kaizer SSTC II driver and bridge (6 turns of 4mm2 wire on primary, about 2500 turns of 0.15mm wire and the resonant frequency is about 133kHz) and I have problem with the freewheeling diodes (when I increase bus voltage to about 180VDC on variac, two of them are blowing). I designed PCB's for the driver and bridge separately and I'm wiring the gates of the IGBT's with some 0.75mm2 wire. The core for GDT is some random inductor core from SMPS (I'm gonna replace this because the ramp on the gates is a bit too high in my opinion but i don't think this could be the main problem, maybe I'm wrong). The IGBT's I'm using: IRGP50B60PD1 (from China and maybe here's the problem, I have legit IRFP460's but I don't want to damage them since I'm not sure what's the problem).
Freewheeling diodes: MUR1560.
I posted some photos:
1. Picture of my PCB's
2. My workplace (for higher bus voltage I'm using shielding)
3. Primary of the GDT at high side driver to GND (yellow) and gate of one IGBT (purple)
4. Gate voltage waveforms
5. High voltage spikes across the freewheeling diode (30VDC on bus)
Later I can post some videos on how this is running (and damaging)

I would be grateful if someone will help me to solve this problem.

Best Regards,
Daniel

ritaismyconscience · « **Reply #1 on:** May 12, 2020, 05:21:12 AM »

You're probably blowing them with voltage spikes. Try using TVS diodes to clamp the voltage spikes. Also your GDT needs to be rebuilt, too much droop. You're driving the transistors in the linear region, which is bad as it makes a lot of heat

davekni · « **Reply #2 on:** May 12, 2020, 06:17:04 AM »

Yes, definitely need a better GDT core. Also, looks like your nominal gate drive voltage is 10V. IGBTs need 15V, and are usually driven higher for Tesla coil use, 20 to 24V.

The spikes are due to inductance in your bridge wiring. It's best to have a two-layer ECB for just the power connections, using planes rather than traces. One layer can be VBus+ and VBus-, half and half, with local polypropylene film capacitors. The other plane is also half-and-half, with the split perpendicular to the first plane, for the two H-Bridge outputs. Gate wiring can be on a separate ECB or just a diode and resistor in series with the twisted-pair GDT output leads, soldered directly to the IGBTs. Gate wiring should have reasonably low inductance, but the bridge power wiring is much more critical for low inductance.

Besides VBus film capacitors on the H-Bridge, have the bulk electrolytic VBus cap(s) close and wired to the H-Bridge with low-inductance, paired or even twisted wires.

hip424 · « **Reply #3 on:** May 12, 2020, 08:34:32 AM »

Hi,
Thanks for replies. I'll order new GDT core and use TVS diodes but should I use two-way TVS diode like in the schematic below? (1.5KE550CA-LF will be OK?) When I'll rewind GDT I will post some updates

Best Regards
Daniel

Mads Barnkob · « **Reply #4 on:** May 12, 2020, 08:38:58 AM »

From: http://www.richieburnett.co.uk/temp/gdt/gdt2.html

Quote

Excessively drooping tops and bottoms to waveform.

- The tops and bottoms of pulses slope steaply down towards zero,
- Far too low primary inductance. Either too few turns or wrong core material used.
- Unusable because the MOSFETs would start to turn off towards the end of the pulses!

Use many more turns or choose a core type with higher Specific Inductance (Al).

I think you are simply burning your diodes out from improper driving of the MOSFETs, fix this first and we can take a look at the diodes afterwards if there is still a problem with proper gate drive.

Your PCB layout is not far from my SSTC1/2/3 and while it could be better and lower inductance, (how bad is stray indutance? http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/busbar-and-primary-circuit/ ) it will work. Add more snubbering capacitance on the DC bus if you can not have your electrolytic capacitor closer to the MOSFETs. See examples here: Snubbers: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/snubber-capacitor/ and electrolytic capacitors: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/dc-bus-capacitor/

hip424 · « **Reply #5 on:** May 12, 2020, 08:50:40 AM »

Hi Mads,
I will rewind my GDT and the electrolytic caps will be close to the DC bus. In terms of calculating stray inductance I'm not sure how can I calculate this since I don't have much knowledge in this field.

Best Regards
Daniel

Weston · « **Reply #6 on:** May 12, 2020, 10:21:29 PM »

You don't need to be able to calculate the exact value of stray inductance, you only need to know how to minimize it.

Stray inductance is proportional to the loop area. Look at your PCB layout and draw a line from the + terminal of the film capacitor to the cathode of the top freewheeling diode / collector of the top IGBT, from the anode of the top diode / emitter of the top IGBT to the cathode of the bottom freewheeling diode / collector of the bottom IGBT, and from the anode of the bottom diode / emitter of the bottom IGBT to the - terminal of the film capacitor. The area of this loop represents your loop inductance. Try to minimize this.

The inductance is also inversely proportional to the width of the traces, try to maximize that. As davekni mentioned, having your + and - power rails as copper planes is a good way to minimize inductance.

The loop inductance is the source of the voltage spikes during switching and can make the difference between parts blowing up or not. I have been meaning to do a more detailed forum post on this, but yesterday I posted some simulations on the other thread for voltage spikes for a "high inductance" layout and for a "low inductance" layout. You can see how big a difference it makes.
https://highvoltageforum.net/index.php?topic=1078.msg7678#msg7678

hip424 · « **Reply #7 on:** May 12, 2020, 10:30:48 PM »

Thanks for your answer Weston,

I get the point and so I have the question:
If I tin traces with some solder does it reduces stray inductance?
If no I will connect connections directly with wire.

Best Regards
Daniel

ritaismyconscience · « **Reply #8 on:** May 12, 2020, 10:59:57 PM »

Adding solder will decrease the resistance but not the inductance.

Try making connections shorter

hip424 · « **Reply #9 on:** May 14, 2020, 02:47:32 PM »

Hi
I've wound a new GDT transformer on the core with 7350nH AI rating. Here I'm posting some of the pictures of waveforms and test setup. 30VDC on bus: yellow is the primary of the GDT to the ground and the purple is at the gate of one of the IGBTs. The amplitude is about 10V. I'm also using 15V zener diodes in the right configuration at the gates of the IGBTs. Unfortunately the voltage spikes at the diodes are bigger (sorry for not posting the picture). And I'm worried that this time if I will crank the voltage the diodes will blow even at lower voltage. I also buyed some TVS diodes (1.5KE550CA to be exact) and could you tell me if these diodes will behave better than MUR1560?

Best regards
Daniel

Fumeaux · « **Reply #10 on:** May 14, 2020, 05:21:49 PM »

Hi,

I guess the yellow graph is the Gdt output? I think that the voltage for the driver IC's drops too low during switching. Do you mind scoping the 12V on the driver IC? If thats the case a big cap might fix it.

Do you currently have any pseudoload at the bridge output? That could result in some spikes too.

You should twist the wires from the Gdt in the corresponding pairs. Twisting Green and White for example might even be counterproductive.

hip424 · « **Reply #11 on:** May 14, 2020, 05:46:39 PM »

The yellow graph was the primary of GDT. Now the yellow is 12VDC powerline to driver ICs and it looks OK to me. It gets a little bit of ringing at higher voltage on the bridge but overall it looks good. Purple graph is gate of one of the IGBTs. What do you mean by saying pseudoload? I have my primary of the tesla coil connected to the bridge output.

Best Regards
Daniel

Fumeaux · « **Reply #12 on:** May 14, 2020, 06:56:16 PM »

Yeah the primary is great. I just wanted to make sure the bridge output isn't just floating, but that there is an actual load.

Quick question, what exactly did you change for that improvement? Just a bigger cap?

hip424 · « **Reply #13 on:** May 14, 2020, 07:02:55 PM »

I didn't changed nothing. I have 1000uF cap at the output of the 12V regulator. On the bridge also nothing changed. I also don't know of what improvement you are talking about (from what moment). Since my first post I just changed GDT and some wiring. For now that's it. I just wanna know can I safely run it on higher voltage without blowing freewheeling diodos (my main question).

Update:
I ran my Tesla Coil at about 330VDC and one of the IGBTs blew up (never buy chinese electronics). I've replaced them with IRFP460 and it works nicely at 330VDC. I will post here some pictures later. When I will run this on direct mains I will post some update.

Thanks guys for all your help
Best Regards
Daniel

hip424 · « **Reply #14 on:** May 16, 2020, 03:42:42 PM »

Hi
My topload at higher BPS gets very hot (smoke comes out of it or maybe from glue). What can cause that problem? The topload is made out of aluminium "donut".

So I wanted test it at increased BPS. I increased BPS and slowly increased voltage. And BOOM two of the transistors blew up (the current from mains was about 7A). Any ideas why this happened? I suppose that two of them turned on at the same time and caused short circuit. But why? Tomorrow I will replace IRFP460's with IXFH20N50P3's. And I have one more question: Can I put them without thermal pads? I'm using anodized radiator and it doesn't conduct at the surface.

Best Regards
Daniel

flyglas · « **Reply #15 on:** May 16, 2020, 08:09:46 PM »

Hi Daniel,

i would always use a thermal (isolation) pad. The anodised layer is very thin (a few µm) and not sufficient to withstand the voltage for a long time.

hip424 · « **Reply #16 on:** May 16, 2020, 08:18:25 PM »

Hi flyglas,
Thanks for your answer. I will continue using thermal pads.
And please someone help me solve the case of burning transistors.

Best Regards
Daniel

davekni · « **Reply #17 on:** May 16, 2020, 08:36:20 PM »

Concerning top-load heating and smoking, have you measured continuity from the top load to ground (through the secondary winding)? If there's a bad connection (gap) at the top of the coil, it may still function normally, but the arc across that gap will fry glue etc.

Another less-likely possibility comes to mind: induction heating of the top-load. Most Tesla coils have a section of spaced winding at the top of the secondary, or spacers from the secondary top to top-load. That provides a return path for magnetic flux. If you don't have such space, perhaps there could be enough current induced into the top-load to heat it.

hip424 · « **Reply #18 on:** May 16, 2020, 08:44:49 PM »

Hi davekin,
Thanks for your answer. In my case the second option is more probable cause I don't have much space from secondary to the topload. I am posting picture of how it looks under topload.

Best Regards
Daniel

ElectroXa · « **Reply #19 on:** May 16, 2020, 11:00:37 PM »

Hi
Couldn't topload heating be caused by the fact the topload is acting like a shorted turn in your design, thus heating a lot ?

I found this answer here :

Quote

Some top load electrodes behave as shorted turns and those should be used only on longer secondary coils, perhaps with length about 3 times their diameter.

( http://donklipstein.com/sst.html#o )

High Voltage Forum · « **Reply #19 on:** May 16, 2020, 11:00:37 PM »

Author Topic: Freewheeling diodes are burning (Now other problems) (Read 8576 times)

High Voltage Forum

Recent Topics and Posts