Author Topic: GDT keeps on killing IGBTs  (Read 6387 times)

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
GDT keeps on killing IGBTs
« on: February 21, 2020, 02:41:03 AM »
Hi,
I have a problem with my gate driver, as it randomly kills IGBTs. I've already had to replace 4 of them due to this problem.

Here's what my circuit looks like:



Here's a couple pictures of my setup:



The brown and green wires go to the GDTs



A picture of the waveform with 2 44nF capacitors simulating the gates of the 4 IGBTs being driven by the GDT:


(Scale: 10V/div & 2us/div)

The driver circuit is a chinese version which uses a CPLD, but the output stage uses the same parts as the UD 2.7 (It uses the FDD8424)
Here's a picture of it:





I think I might need to swap the TVS diodes for a lower voltage one. Anything else I need to change?

Edit: Unrelated question: I heard that you should use a fast diode instead of relying on the internal diode in the IGBT. Would using RURG8060 work, and would I need to put the diode on a heatsink? Thanks.
« Last Edit: February 21, 2020, 04:28:34 AM by ritaismyconscience »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 378
  • Karma: +22/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #1 on: February 21, 2020, 04:42:34 AM »
I see that your gate driver is shielded.  Is there a connection between the shield and gate-driver input power V- somewhere near where the power comes into the shield?  FPGAs can get confused by stray field spikes.

There could also be a bug in the FPGA code causing high-frequency oscillation or other problems with the gate drive waveform during operation.  It's difficult to catch waveforms during the failure event.  If your scope has a pulse-width trigger mode, try looking for short gate-drive pulses.

Without knowing the circuit, it's going to be difficult to find the problem.
David Knierim

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #2 on: February 21, 2020, 06:11:32 AM »
I'll try replacing the TVS diodes and I'll try adding 24V Zener diodes.

Edit: I came up with this. Will this work?



Also, I looked at other people's schematic, and they used 5.1 ohm resistors on each gate instead of doing what I did. Does this matter?
« Last Edit: February 21, 2020, 07:32:22 AM by ritaismyconscience »

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1387
  • Karma: +23/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: GDT keeps on killing IGBTs
« Reply #3 on: February 21, 2020, 02:44:44 PM »
The whole idea of a gate resistor is to have it as close as possible to the gate of the IGBT/MOSFET in order to quench ringing between input capacitance and parasitic inductance in the gate drive circuit.

You can not just move it to the primary side of your GDT and expect the same function. What you have done is entirely different.

I can not quite make out from your pictures, which wires goes where, but generally you have way too long leads going everywhere, have everything as short as possible. Everything else is just asking for even more inductance and "antennas" to pick up noise from spark EMI.

Your GDT is not killing your IGBTs, your changes to the gate drive circuit and far-from-optimal wiring is killing your IGBTs
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 311
  • Karma: +11/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #4 on: February 21, 2020, 08:00:10 PM »
It's also a good idea to add a speedup for the turn off resistance to help avoid shoot through etc.

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 378
  • Karma: +22/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #5 on: February 22, 2020, 03:39:10 AM »
Mads and Hydron are likely right about the real cause of IGBT frying:  clean short wiring and gate damping at the gates with speed-up diodes for the turn-off edges.  (I was too focused the controller per your question.)

One other issue that could be causing frying is too little lead-compensation in the controller.  The gate drive signal should ideally switch just before the zero-current times of the primary resonant circuit.  Too far off from that is hard-switching of the IGBTs, which fries them.
David Knierim

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #6 on: February 22, 2020, 06:49:22 AM »
Haven't tested it with voltage yet, so it's probably not the controller's fault

Offline dj.cosmo.esq

  • High Voltage Enthusiast
  • *
  • Posts: 14
  • Karma: +0/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #7 on: February 22, 2020, 07:41:18 AM »
Hey just a little input I have the same board which has given me hell on Earth (kinda). My question to you is are you using one GDT? Cause I have found some Chinese to English translations that say you are supposed to parallel the outputs since one is driving the GDT and the other is driving nothing. I ended up making 2 GDTs to drive 4x cm400ha-24h but being impatient I ran it off if the GDT1 and blew up the MIC chip on GDT2 but the IGBTs are ok. How are you getting your feedback signal? I was using a signal generator (fail) but Dave had a good idea and I went searching for a Kickstart circuit which actually was attached to a induction heater. Ive been trying to find the firmware to look at for the clpd board but no luck yet.

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #8 on: February 23, 2020, 05:43:17 AM »
My question to you is are you using one GDT?

I'm using 2 GDTs, one to drive each half of the fullbridge. Seems to work so far.

How are you getting your feedback signal?

Haven't tested it with voltage yet. However, using a 1:1023 feedback transformer should work.

Also, I fixed my driver and I saw this on my oscilloscope:



Does this look normal? For some reason the voltage spikes to 30V at the start
« Last Edit: February 23, 2020, 06:36:29 AM by ritaismyconscience »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 378
  • Karma: +22/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #9 on: February 24, 2020, 04:37:33 AM »
When you say "haven't tested it with voltage yet", are you saying that the IGBTs fried with no collector voltage, only gate drive?  That would be extraordinary.

Was the scope trace taken at the IGBT gate-to-emitter?  Normally there is undershoot there, but not overshoot.  Typically the gate series resistor is high enough for critical damping with the GDT leakage inductance for the low-to-high transition.  For high-to-low, a diode bypasses some of the gate drive resistance, making it underdamped, generating undershoot.
David Knierim

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #10 on: February 24, 2020, 07:08:30 AM »
Looks like I must have connected the scope leads backwards then, because it probably caught the negative cycle. Anyways, would -30V across gate and emitter be enough to kill my IGBTs? It looks like my zener diode circuit I came up with was too slow to deal with the spike.

When you say "haven't tested it with voltage yet", are you saying that the IGBTs fried with no collector voltage, only gate drive?  That would be extraordinary.

Yes, but it's probably because I messed up the circuit in my first attempt. I've also used a hammer to reveal the die, and there aren't any burn marks.
« Last Edit: February 24, 2020, 07:12:26 AM by ritaismyconscience »

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #11 on: February 25, 2020, 04:23:50 AM »


Looks like you were right, here's the waveform on the positive cycle.

Anyways, is -30V going to kill my IGBTs or should I not worry about it?

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 378
  • Karma: +22/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #12 on: February 25, 2020, 06:20:04 AM »
Mads has said in other posts that -30V is fine for momentary undershoot, even though it's beyond the IGBT's specification limit.  I have no personal experience with such.

The zener diodes are likely fast enough, but not low-enough impedance to clamp the gate signal.

To avoid damaging your circuit, especially once bus voltage is applied, be careful about grounding sensitive nodes to your scope.  Even if the scope is floating (not connected to line ground), capacitance from the scope "ground" to the line power can be enough to disrupt circuit operation.  So, "ground" only low-impedance nodes that don't have intentional AC signals on them, such as the low-side emitters or high-side collectors (bus voltage - and + nodes), not gate signals or bridge outputs.  If you "ground" a floating scope to any high voltage such as + bus voltage, be careful not to shock yourself from the scope case!
« Last Edit: February 25, 2020, 06:25:35 AM by davekni »
David Knierim

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #13 on: February 25, 2020, 06:27:40 AM »
Mads has said in other posts that -30V is fine for momentary undershoot, even though it's beyond the IGBT's specification limit.  I have no personal experience with such.

To avoid damaging your circuit, especially once bus voltage is applied, be careful about grounding sensitive nodes to your scope.  Even if the scope is floating (not connected to line ground), capacitance from the scope "ground" to the line power can be enough to disrupt circuit operation.  So, "ground" only low-impedance nodes that don't have intentional AC signals on them, such as the low-side emitters or high-side collectors (bus voltage - and + nodes), not gate signals or bridge outputs.  If you "ground" a floating scope to any high voltage such as + bus voltage, be careful not to shock yourself from the scope case!

I'm going to use a second current transformer to monitor the current and I'm going to stick a wire through the gate transformer to keep track of gate voltage. This way my scope is completely isolated from the voltage.

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1387
  • Karma: +23/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: GDT keeps on killing IGBTs
« Reply #14 on: February 25, 2020, 08:28:28 AM »
Anyways, is -30V going to kill my IGBTs or should I not worry about it?

http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/

Quote
Maximum Gate-to-Emitter Voltage (VGE)

The gate voltage is limited by the thickness and characteristics of the gate oxide layer. Though the gate dielectric rupture is typically around 80 volts, the user is normally limited to 20 or 30V to limit current under fault conditions and to ensure long term reliability.

It is normal practice to drive IGBTs at +24VDC with the Steve Ward universal driver and clones of it that most use. The reason for the gate voltage limit is not so much for protecting the gate itself, it will first break down at some 80 Volt. Higher gate voltage means higher currents can be conducted through the Collector-Emitter. We take advantage of this by pushing the gate a little over its rated voltage to allow us to conduct higher currents through the IGBT, at the cost of higher switching losses. As explained by manufacturers in the following.

It is important to note however that IGBTs exhibit relatively high gain even at high gate-emitter voltage. This is because increasing the flow of electrons by increasing the gate-emitter voltage also increases the flow of holes. The gain of a high voltage power MOSFET however is very insensitive to gate voltage once fully on.

http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #15 on: February 26, 2020, 05:55:55 AM »
Thanks for all of your help, looks like my setup is working now

Here's what the waveforms look like on 25VAC (the bus voltage is double the AC voltage, so the bus voltage is 70V DC)



1 div = 20A current

This is my first time building one of these, so I might need more help in the future
« Last Edit: February 26, 2020, 08:06:35 AM by ritaismyconscience »

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1387
  • Karma: +23/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: GDT keeps on killing IGBTs
« Reply #16 on: February 26, 2020, 04:01:18 PM »
This is my first time building one of these, so I might need more help in the future

Congratulations on having it oscillate correctly. Before you ask, low voltage testing can have seemingly large switching transients due to output capacitance of the IGBT being high at low C-E voltages. So do not fear large switching transients below some 100-200V DCbus

We are always here to help you with your Tesla coils, you provide a good amount of information and pictures to make it easier to understand your problems  ;)
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #17 on: February 27, 2020, 03:27:07 AM »
I tested it at half bus voltage today (170V) and I got this:



There seems to be some switching spikes (1 div = 100A). Does this look normal?

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 378
  • Karma: +22/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #18 on: February 27, 2020, 05:27:56 AM »
Those transients do look high.  The actual current through the primary resonant circuit likely doesn't have such current spikes.  It's more likely capacitive or inductive coupling in your scoping current transformer or its wiring.  Do you have a picture and any details of your sense transformers, both the one used for scoping and the one for feedback?  What load resistor are you using on the scope transformer?

It's hard to tell in the picture, but it appears that the spikes occur shortly after the current zero-crossing points.  The H-Bridge switching is softer if done just before current zero-crossing (by increasing lead compensation in the gate-driver circuit).  That way the Tesla primary current makes the voltage transition during the IGBT dead-time.  Lack of dead-time would be another cause of large switching spikes.  Do you have a sketch and/or images of how your gate-drive transformers and resistors/diodes are wired now?
David Knierim

Offline ritaismyconscience

  • High Voltage Technician
  • ***
  • Posts: 135
  • Karma: +3/-0
    • View Profile
Re: GDT keeps on killing IGBTs
« Reply #19 on: February 27, 2020, 07:22:17 AM »

^ Feedback transformer (1:1023 for both)


^ 1:1023 current transformer + 10 ohm resistor (connects to scope)


^ Circuit design


^ A picture of the driver I'm using. It doesn't seem to have phase lead compensation. Is it possible to modify the circuit to add this?
« Last Edit: February 27, 2020, 07:28:20 AM by ritaismyconscience »

High Voltage Forum

Re: GDT keeps on killing IGBTs
« Reply #19 on: February 27, 2020, 07:22:17 AM »

 


* Recent Topics and Posts

post Re: MMC for a 32kvac load
[Capacitor Banks]
Weston
Today at 03:06:20 AM
post Re: Isn't there a better way to make a custom toroid?
[General Chat]
johnf
Today at 01:33:47 AM
post MMC for a 32kvac load
[Capacitor Banks]
Rowdy
Today at 01:02:34 AM
post Re: Isn't there a better way to make a custom toroid?
[General Chat]
Rowdy
May 24, 2020, 11:47:26 PM
post Re: GDT cable specs
[Transformer (Ferrite Core)]
costas_p
May 24, 2020, 11:09:44 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
Rowdy
May 24, 2020, 10:39:45 PM
post Re: GDT cable specs
[Transformer (Ferrite Core)]
station240
May 24, 2020, 09:55:49 PM
post Re: Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
davekni
May 24, 2020, 09:55:48 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
Mads Barnkob
May 24, 2020, 08:16:40 PM
post Re: GDT cable specs
[Transformer (Ferrite Core)]
thecyberman
May 24, 2020, 05:38:27 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
Rowdy
May 24, 2020, 02:33:55 PM
post Re: Freewheeling diodes are burning (Now other problems)
[Solid State Tesla Coils (SSTC)]
hip424
May 24, 2020, 02:31:39 PM
post Re: Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
iraquois
May 24, 2020, 02:08:53 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
May 24, 2020, 12:12:36 PM
post Re: Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
ritaismyconscience
May 24, 2020, 07:56:21 AM
post Re: Freewheeling diodes are burning (Now other problems)
[Solid State Tesla Coils (SSTC)]
ritaismyconscience
May 24, 2020, 03:09:57 AM
post Re: Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
davekni
May 23, 2020, 10:24:55 PM
post Re: GDT cable specs
[Transformer (Ferrite Core)]
johnf
May 23, 2020, 09:30:29 PM
post GDT cable specs
[Transformer (Ferrite Core)]
costas_p
May 23, 2020, 09:06:39 PM
post Re: Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
Mads Barnkob
May 23, 2020, 08:22:59 PM
post Re: Freewheeling diodes are burning (Now other problems)
[Solid State Tesla Coils (SSTC)]
davekni
May 23, 2020, 07:47:49 PM
post Re: Building a bigger DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
May 23, 2020, 07:41:41 PM
post Working Principle Of Neon Transformer Circuit
[Transformer (Ferrite Core)]
iraquois
May 23, 2020, 02:09:26 PM
post 600 Watt GEL Battery Charger Teardown - Signet Systems HB600-24B
[Electronic Circuits]
Mads Barnkob
May 23, 2020, 12:46:23 PM
post Re: Building a bigger DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ritaismyconscience
May 23, 2020, 04:14:52 AM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ritaismyconscience
May 22, 2020, 01:15:36 AM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mugi30
May 21, 2020, 11:55:37 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
May 21, 2020, 07:46:28 PM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mugi30
May 21, 2020, 09:15:57 AM
post Re: How to get rid of stuff?
[Sell / Buy / Trade]
johnf
May 21, 2020, 06:35:31 AM
post Re: How to get rid of stuff?
[Sell / Buy / Trade]
MRMILSTAR
May 21, 2020, 05:23:40 AM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
futurist
May 21, 2020, 12:16:04 AM
post Re: How to get rid of stuff?
[Sell / Buy / Trade]
klugesmith
May 20, 2020, 11:22:34 PM
post Re: Powerful easy to build MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
May 20, 2020, 12:15:43 PM
post Re: Portable Q(uarantine)CW Tesla Coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
SteveN87
May 19, 2020, 02:08:20 PM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mugi30
May 19, 2020, 12:05:51 PM
post Re: Powerful easy to build MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
May 19, 2020, 11:49:02 AM
post Re: Powerful easy to build MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Mads Barnkob
May 19, 2020, 09:26:06 AM
post Re: Portable Q(uarantine)CW Tesla Coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Weston
May 19, 2020, 07:55:54 AM
post Re: What's the best way to trigger these things?
[Capacitor Banks]
klugesmith
May 19, 2020, 01:55:11 AM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ritaismyconscience
May 18, 2020, 11:54:27 PM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mugi30
May 18, 2020, 10:10:09 PM
post Re: New projet OmegaDR
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mugi30
May 18, 2020, 10:05:14 PM
post Re: What's the best way to trigger these things?
[Capacitor Banks]
johnf
May 18, 2020, 09:03:11 PM
post Re: What's the best way to trigger these things?
[Capacitor Banks]
John123
May 18, 2020, 07:50:08 PM
post Re: Powerful easy to build MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
May 18, 2020, 04:15:16 PM
post Re: Freewheeling diodes are burning (Now other problems)
[Solid State Tesla Coils (SSTC)]
hip424
May 18, 2020, 03:13:14 PM
post Re: How to get rid of stuff?
[Sell / Buy / Trade]
klugesmith
May 18, 2020, 06:31:25 AM
post Re: Building a bigger DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ritaismyconscience
May 18, 2020, 03:18:40 AM
post Re: What's the best way to trigger these things?
[Capacitor Banks]
davekni
May 18, 2020, 02:26:27 AM

Sitemap 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 
SimplePortal 2.3.6 © 2008-2014, SimplePortal