Author Topic: DRSSTC low power testing killing IGBTs  (Read 10061 times)

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
DRSSTC low power testing killing IGBTs
« on: November 15, 2023, 12:45:37 AM »
So I recently built the microbrute drsstc from EVR and was having pretty good results. I was up to 23in arcs at 350A and 80VAC in on my variac.

I decided to make the breakout point a little longer and pointed upwards a bit more so as to avoid ground strikes. That’s when everything went wrong. Blew an IGBT (60n60 minibricks).

I replaced the bad IGBT, went back to the original breakout point, and then tried again. Blew both IGBTs this time. Thinking that the 60n60s were simply fakes or whatever (got them on eBay for cheap) I then swapped the IGBTs to the “Thunderbolt” IGBTs that EVR sells. I decided to remove the secondary and place a tin can in its place.

Tested at very low power (30vac in from variac), heard the buzzing from the Interruptor, and then blew one of these new and $$$ IGBTs.

Completely at a loss for what on earth could be happening here? I don’t even know where to begin with testing seeing as I only changed the breakout point before all of this started.

Before going into great detail of the system, what could cause these IGBTs to just suddenly begin failing (shorted) at low voltage?

-Matt

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #1 on: November 15, 2023, 03:01:35 AM »
Here are the gate drive signals with a function generator connected. (no IGBTs connected).  They look to be operating as expected, but I do see significant ringing. Gate drive resistors are 5ohm, maybe I should go to 10ohm? Might this be enough to be eating up IGBTs?

-Matt

Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #2 on: November 15, 2023, 04:51:16 AM »
Quote
Here are the gate drive signals with a function generator connected. (no IGBTs connected).  They look to be operating as expected, but I do see significant ringing. Gate drive resistors are 5ohm, maybe I should go to 10ohm? Might this be enough to be eating up IGBTs?
Vge ringing measurements without IGBT load are not useful.  Damping comes from series resistor into IGBT Cge.

For testing, run very low duty cycle, perhaps one enable pulse per second and <100us pulse width.  Even at low voltage, current can ramp up to a high level.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #3 on: November 15, 2023, 01:57:50 PM »
I’ve got one IGBT left, so I can connect it and measure again later today. Assuming that fixes the ringing, what would be the next thing to suspect? Bad TVS that’s failing to catch spikes on the bus? None of them are shorted, but that doesn’t mean that they aren’t open.

When I was testing previously, current got up to around 100A. I could see bus voltage switching close to zero crossing with current.

-Matt

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #4 on: November 16, 2023, 03:36:16 AM »
Looks like the gate signal with the one good IGBT connected is clean. Not sure where to go with this?

-Matt


Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #5 on: November 16, 2023, 04:08:09 AM »
Quote
Assuming that fixes the ringing, what would be the next thing to suspect? Bad TVS that’s failing to catch spikes on the bus? None of them are shorted, but that doesn’t mean that they aren’t open.
Yes, testing bidirectional TVS diodes requires enough voltage to check breakdown.

General thought is to run at very low duty cycle and post scope waveforms.  Perhaps I or someone else can detect an anomaly.

One possible issue, though perhaps not too likely, is occasional slow or failing oscillation startup at beginning of enable pulse.  Startup was a problem with my friends EVR Microbrute, though it did not result in damaged IGBTs.  As Mads has suggested for other builds, a resistor across bridge output helps with startup.  That fixed my friends Microbrute.  A resistor between Out+ and Out- nodes.  Makes output center of voltage-doubler supply.  Without resistor, slight leakage current differences between IGBTs can cause initial half-bridge output to be at one supply rail or the other.

Quote
Looks like the gate signal with the one good IGBT connected is clean. Not sure where to go with this?
Duty cycle is above 50% and positive voltage is above negative voltage.  Perhaps a capture zoomed out to include the initial edge at beginning of enable pulse would help show what is happening.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #6 on: November 16, 2023, 04:51:00 AM »
I see that now. Not sure what’s causing that. I’ll add the resistor and try again when some more IGBTs arrive.

Here is the zoomed out waveform. Still shows the irregularities.

Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #7 on: November 16, 2023, 05:29:40 AM »
Quote
I see that now. Not sure what’s causing that. I’ll add the resistor and try again when some more IGBTs arrive.
That bridge output resistor will not fix Vge issue.  Probably worth figuring out Vge problem before running with any bridge voltage.  Perhaps scope GDT secondary along with Vge and capture an even longer scope trace.  GDT secondary can't support an average DC voltage.  Longer capture of those two signals might hint at what is wrong.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #8 on: November 17, 2023, 12:32:28 AM »
Here is the measurement from the GDT secondary (in purple). Yellow is VGE on the good IGBT.

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #9 on: November 17, 2023, 01:08:41 AM »
I decided to use my diff probe to see how VGE looks. Looks perfect to me?




Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #10 on: November 17, 2023, 01:09:58 AM »
Quote
Here is the measurement from the GDT secondary (in purple). Yellow is VGE on the good IGBT.
I'm guessing one of the driver chips is bad.  I'd scope those two outputs next.  And perhaps captures of both sides of coupling capacitors C10 and C11 between driver U4 and GDT primary to make certain those are not shorted.

BTW, the instructions and schematics my friend received from EVR are somewhat inconsistent.  Is your GDT 1:1:1 as per instructions?  (Schematic shows 1:2:2.)  There could be other differences between my friend's and yours.  My recommendations here are presuming your coil is similar to the one my friend purchased a couple years ago.  Schematic is copyrighted, so I can't post a copy.

Presuming C10 and C11 are good and both 2uF, GDT inductance must be quite high.  How many turns on your GDT?

Quote
I replaced the bad IGBT, went back to the original breakout point, and then tried again. Blew both IGBTs this time.
When one IGBT fails shorted, that stresses the other IGBT.  (Usually destroys other IGBT, but occasionally not.)  It is quite possible that other IGBT was partially damaged.  It then failed completely, which then destroyed the replacement IGBT.
Another possibility is that some of the 1.5KE220CA TVS diodes failed first.  Depending on avalanche energy capability of the IGBTs being used, TVS diodes can cause more issues than they fix.  If TVS diodes fail shorted, that causes opposite IGBT to die driving the short.  That may then cause enough current to burn open the failed TVS diode bond wires.

Quote
I decided to use my diff probe to see how VGE looks. Looks perfect to me?
Oh.  Was there any power to half-bridge during this scoping?  Presuming not, was scope probe ground clip connected to emitter?  Diff probe is a better way to measure, but I'm puzzled as to how the previous traces ended up the way they are.

Vge is low.  Perhaps there is still a driver chip issue?
« Last Edit: November 17, 2023, 01:15:17 AM by davekni »
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #11 on: November 17, 2023, 01:22:03 AM »
My GDT is 1:1:1 and I see is supposed to have 10 turns. I actually have 15 since that’s how many fit. Could that be a problem? The capacitor is two 1uF in parallel, so 2uF total.

I swapped both driver ICs and same result with diff probe.

Previous measurements, I actually had ground clip tied to board ground. I can clip to emitter to see if that makes a difference. No power to bridge during any measurements.

-Matt

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #12 on: November 17, 2023, 01:29:51 AM »
Here is the VGE with ground clip on emitter.

Forgot to mention that I’m not using the onboard MMC. I upgraded to a larger one…Having said that, my first blown IGBT was with the on board MMC, so I’m leaning towards something else. I’m also using a different primary and secondary (much larger and robust than what was called for).

Any chance the upgraded MMC is the problem? It’s 0.165uF rated for 6400v.


Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #13 on: November 17, 2023, 03:57:38 AM »
Quote
Here is the VGE with ground clip on emitter.
Looks much better, though still a bit low.  Microbrute does produce somewhat low Vge, ~14V.  Perhaps scope probe compensation or other calibration would verify if voltage is really ~12V as shown.

Quote
Forgot to mention that I’m not using the onboard MMC. I upgraded to a larger one…Having said that, my first blown IGBT was with the on board MMC, so I’m leaning towards something else. I’m also using a different primary and secondary (much larger and robust than what was called for).

Any chance the upgraded MMC is the problem? It’s 0.165uF rated for 6400v.
Can't think of any reason that would cause issues.

Quote
My GDT is 1:1:1 and I see is supposed to have 10 turns. I actually have 15 since that’s how many fit. Could that be a problem?
Extra turns does have a down-side of increased leakage inductance.  Slows down Vge edges a bit.  Probably not a big concern here.  If you want to improve GDT, winding with two twisted pairs will further reduce leakage inductance:
https://highvoltageforum.net/index.php?topic=1854.msg13949#msg13949

Quote
I was up to 23in arcs at 350A and 80VAC in on my variac.
Might be just a bit too high current for this design.  Especially if your Vge is even lower than typical.  Microbrute controller has no phase lead.  IGBTs switch slightly after zero current, so hard-switch off-to-on.  That adds some stress and power dissipation.  Low Vge could be causing IGBTs to come out of saturation at peak current.  (Did you measure current or base it on measured R23 POT voltage?)
There are other driver chips that can handle over 15V.  Also some with FET-only output stages that would provide closer to full supply voltage to Vge.  You could increase Vge slightly by tweaking 15V towards 16V absolute maximum rating of UCC3732X chips.  One easy patch is to add a diode in series with LM7815 ground lead.  Adds ~0.6Vf of diode.
This post lists a few 20+V driver chips.  Mostly duals for UD2.7 etc. drivers:
    https://highvoltageforum.net/index.php?topic=2400.msg17650#msg17650
There may be pin-compatible drivers for UCC3732X with either higher voltage capability and/or FET-only output stages.  I'd recommend a DigiKey search.  Even if purchasing elsewhere, I find DigiKey's search easiest to use.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #14 on: November 17, 2023, 04:45:57 AM »
Current was measured with a Pearson 110. I do have the OCD set and it did trigger a few times during tuning, but failures all occurred well below the limit. After I had things tuned and hitting around 23-24”, no more OCD triggering.

I guess the next thing is going to be getting new IGBTs installed and managing some scope shots with power applied to the bridge?

If all else fails, I do have a UD2.7 driving a cm300dy-12nF for my fusor power supply. I’ll probably duplicate that setup and use for this coil.

EVRs stuff seems very finicky and expensive to own.

-Matt

Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #15 on: November 17, 2023, 05:13:52 AM »
Quote
but failures all occurred well below the limit.
That does seem odd.  Makes my previous guesses unlikely.
Possible that some previous event fried TVS diode(s) open.  Looks like IXGN60N60C2D1 isn't avalanche rated.  That could make TVS diodes important to clamp any switching transients.  (I forget how good or not the bridge ECB layout is.)
Or perhaps it is still possible that occasional startup failure caused IGBT failure.  Vge tends to look like a very-low-frequency ring (2uF with GDT inductance) if oscillation doesn't start.  That doesn't often fry IGBTs, but it does put them in linear range for extended times.  Resistor across bridge output helps make startup reliable.

Quote
I guess the next thing is going to be getting new IGBTs installed and managing some scope shots with power applied to the bridge?
I can't think of any other good option (except your next though).  Add the output resistor first.

Quote
If all else fails, I do have a UD2.7 driving a cm300dy-12nF for my fusor power supply. I’ll probably duplicate that setup and use for this coil.
UD2.7 is a better driver.  Phase lead and larger Vge swing allow pushing current higher for a given IGBT.  Bridge output resistor is still needed (unless IGBT leakage currents match well).

Quote
EVRs stuff seems very finicky and expensive to own.
My only direct experience with EVR was purchasing a MIDI interrupter kit in 2019.  (Was in a hurry to get something for demonstrating my DRSSTC at a local science fair.  Got my DRSSTC running 10 days before the fair, so had little time left.)  Quite disappointed with its performance.  Designed and built my own for the next year.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #16 on: November 17, 2023, 02:22:29 PM »
I’ve noticed at times that startup requires a much longer on time than I usually run. I usually keep on time below 100uS, but have to start up near 250uS (from time to time) to get it to start. The first set of IGBTs were lost at low power and I could hear a thumping from my variac. OCD was set to 150A, so it wasn’t that.

Second set was lost very inexplicably since they had been going strong for many runs.

Last two sets lost at low power, too low to see any discharge. (Last set was with secondary removed).

Maybe startup is somehow killing them?

What size/type resistor should I use for start up?

-Matt

Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #17 on: November 17, 2023, 07:28:06 PM »
Quote
What size/type resistor should I use for start up?
I think 5k to 10k is common.  Generally lower is better until power dissipation becomes a nuisance.  Lower resistance can overcome higher IGBT leakage current.
David Knierim

Offline FonziDaytona

  • High Voltage Technician
  • ***
  • Posts: 108
  • Karma: +0/-0
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #18 on: November 17, 2023, 08:49:25 PM »
I think I read that this would go from the CE connection (half bridge) to either positive or negative rail depending on which IGBT comes on first, right? So if positive rail IGBT comes on first, then resistor goes to negative rail?

-Matt

Offline davekni

  • Executive Board Member
  • *******
  • Posts: 2987
  • Karma: +141/-2
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: DRSSTC low power testing killing IGBTs
« Reply #19 on: November 17, 2023, 09:49:11 PM »
Quote
I think I read that this would go from the CE connection (half bridge) to either positive or negative rail depending on which IGBT comes on first, right? So if positive rail IGBT comes on first, then resistor goes to negative rail?
As long as you know for sure which IGBT comes on first, yes, what you say is accurate.  Provides double the startup pulse amplitude.  However, I don't think Microbrute is guaranteed to always start in the same state.  (Unlike UD2.7 which biases comparitor input slightly negative.)  Without knowing initial state, resistor between out+ and out- makes output start in center.  Then either IGBT turning on generates a half-voltage startup pulse.  Even most UD2.7 users don't bother measuring or figuring out which IGBT(s) turn on first, so just use a resistor across bridge output.
David Knierim

High Voltage Forum

Re: DRSSTC low power testing killing IGBTs
« Reply #19 on: November 17, 2023, 09:49:11 PM »

 


* Recent Topics and Posts

post Re: Many Multiple Mini Capacitor
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
Today at 08:06:55 PM
post Re: My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
Today at 08:04:51 PM
post Re: My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
LoFoTroFo
Today at 06:17:25 PM
post Re: Single board for SSTC and DRSSTC operation
[Solid State Tesla Coils (SSTC)]
davekni
December 01, 2024, 10:45:49 PM
post Re: My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
December 01, 2024, 09:53:13 PM
post Re: My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
LoFoTroFo
December 01, 2024, 03:42:44 PM
post Re: Last cap in CW always blowing up
[Voltage Multipliers]
JoeBusic
December 01, 2024, 03:30:50 PM
post Scientific Atlanta (Cisco) 1 GHz Combiner and PSU Teardown
[Electronic Circuits]
Mads Barnkob
December 01, 2024, 11:43:09 AM
post GU-81M Hartley Driven VTTC
[Vacuum Tube Tesla Coils (VTTC)]
janno288
December 01, 2024, 10:46:59 AM
post Single board for SSTC and DRSSTC operation
[Solid State Tesla Coils (SSTC)]
Simranjit
November 30, 2024, 09:10:38 PM
post Re: My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 30, 2024, 07:29:22 PM
post My DRSSTC1 rebuild / Strange GDT ringing
[Dual Resonant Solid State Tesla coils (DRSSTC)]
LoFoTroFo
November 30, 2024, 01:47:58 PM
post Re: Many Multiple Mini Capacitor
[Dual Resonant Solid State Tesla coils (DRSSTC)]
rikkitikkitavi
November 29, 2024, 02:10:44 PM
post Re: Many Multiple Mini Capacitor
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 28, 2024, 09:34:29 PM
post Many Multiple Mini Capacitor
[Dual Resonant Solid State Tesla coils (DRSSTC)]
rikkitikkitavi
November 28, 2024, 07:52:42 PM
post Re: Oscilloscope recommendation for SSTC?
[Solid State Tesla Coils (SSTC)]
Simranjit
November 26, 2024, 11:34:05 PM
post Re: The evolution of a solid state Tesla coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Anders Mikkelsen
November 25, 2024, 01:03:46 AM
post GU-81M VTTC 2.7MHz Help
[Vacuum Tube Tesla Coils (VTTC)]
janno288
November 24, 2024, 01:53:22 PM
post Re: My SSTC's IGBT's blow up when toroid is added
[Solid State Tesla Coils (SSTC)]
AstRii
November 21, 2024, 07:17:16 PM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic Circuits]
petespaco
November 20, 2024, 12:11:18 AM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic Circuits]
betalab99
November 19, 2024, 05:41:27 PM
post Re: Testing and teardown of a commercial induction heater
[Electronic Circuits]
Anders Mikkelsen
November 19, 2024, 01:33:48 PM
post Re: Testing and teardown of a commercial induction heater
[Electronic Circuits]
DashApple
November 19, 2024, 08:02:30 AM
post Re: First DRSSTC, Full Bridge PCB & IGBT Selection question.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 19, 2024, 05:31:47 AM
post Re: 160mm DRSSTC II project | Questions
[Dual Resonant Solid State Tesla coils (DRSSTC)]
futurist
November 19, 2024, 02:33:42 AM
post Re: Testing and teardown of a commercial induction heater
[Electronic Circuits]
dejuli2
November 18, 2024, 11:19:28 PM
post Re: First DRSSTC, Full Bridge PCB & IGBT Selection question.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Benjamin Lockhart
November 18, 2024, 08:43:20 PM
post Re: Testing and teardown of a commercial induction heater
[Electronic Circuits]
DashApple
November 18, 2024, 06:07:11 PM
post Re: Is intentional overlap in the secondary theoretically possible?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
MinuteMylar
November 18, 2024, 05:04:09 PM
post Re: First DRSSTC, Full Bridge PCB & IGBT Selection question.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Beggernator.
November 18, 2024, 12:17:41 PM
post Re: Testing and teardown of a commercial induction heater
[Electronic Circuits]
dejuli2
November 18, 2024, 09:25:38 AM
post Re: First DRSSTC, Full Bridge PCB & IGBT Selection question.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 18, 2024, 01:09:10 AM
post Re: 160mm DRSSTC project
[Dual Resonant Solid State Tesla coils (DRSSTC)]
paulsimik
November 17, 2024, 11:36:52 PM

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