Author Topic: Parallel IGBJT  (Read 616 times)

Offline alan sailer

  • High Voltage Technician
  • ***
  • Posts: 111
  • Karma: +7/-0
    • View Profile
Parallel IGBJT
« on: July 13, 2021, 10:55:49 PM »
I have a spare full bridge assembly using parallel FGA60N65SMD devices. I used it on my QCW coil to compare
two different bridge PCBs ( I didn't see anything significant ).

I was thinking about using it in a small DRSSTC just so it doesn't go to waste. In the QCW I set the OCD to about 150 amps
and everything was great. But I am curious about a reasonable current limit in a DRSSTC. I suspect that with the shorter pulse
length in a DRSSTC would allow much higher currents. One build referenced on this site claimed 250 amps but then killed the
devices after that.

So does anyone have a suggested current limit of parallel IGBJTs in a full bridge design. I plan to run double 110Ac to about 300 volts.

Cheers.

Online davekni

  • High Voltage Expert
  • ******
  • Posts: 1294
  • Karma: +62/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Parallel IGBJT
« Reply #1 on: July 14, 2021, 04:31:32 AM »
The weakest part of FGA60N65SMD is likely its diode, with thermal resistance 4.4x higher than the IGBT.  The part I use in parallel, STGW60H65DRF, is similar at 3.95x higher diode thermal resistance.  I ran some abuse testing of one STGW60H65DRF part, ring-down of an L/C circuit.  At 650A peak IGBT current for the first half-cycle, followed by 600A diode current for the next half-cycle, the part died shorted.  Dissection proved that the diode had fried and the IGBT still functioned.  It had previously survived for many cycles at 600A IGBT 550A diode.

So, FGA60N65SMD would not be good for pulse-skip mode where the diode conducts more of the time.  For normal operation, the limiting factor may be just after OCD tripping, where the diodes conduct the decaying primary current.

Current capability of paralleled pairs will depend on layout details.  Single parts will likely handle 250A, perhaps 300A.  A pair may handle 400A with reasonable layout, 500A with very careful matching.  Of course, these are just my guesses.

One other caveat:  Above presumes these are genuine FGA60N65SMD parts.  Counterfeit parts will die much sooner.
David Knierim

Offline alan sailer

  • High Voltage Technician
  • ***
  • Posts: 111
  • Karma: +7/-0
    • View Profile
Re: Parallel IGBJT
« Reply #2 on: July 14, 2021, 05:08:40 AM »
Thanks so much. I typically don't like getting anywhere near limits so I will be more than happy at 200-300 amps.
I bought the board from Alex Yuan so the pedigree should be good.

The parts were from either Digikey or Mouser. Not sure which but they were not Ebay. So they should be real.

Finally, thanks for the freewheeling warning. One of the things I'd like to try on this build is just that. The UD will be
a SimpleDriver that I also have lying about and it does just about everything.

Cheers.

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 413
  • Karma: +15/-0
    • View Profile
Re: Parallel IGBJT
« Reply #3 on: July 14, 2021, 11:46:24 AM »
The rated Icm for a single part is already 180A, so 200-300A with a parallel pair should be fine, with the noted caveat about the diode.

Even if you have some uneven switching losses from the parallel parts (likely unless you're forcing current sharing or have a very good layout and luck with matching) it's still probably useful even if only to share a bit of the load on the diode (wouldn't hurt to try and match these a bit, and make sure they're thermally connected to try and equalise the die temp).

In my case one of the reasons I switched to a FGY75N60SMD is that the diode is much heftier - 0.48 K/W vs 1.1 K/W thermal resistance!

Offline alan sailer

  • High Voltage Technician
  • ***
  • Posts: 111
  • Karma: +7/-0
    • View Profile
Re: Parallel IGBJT
« Reply #4 on: July 14, 2021, 04:21:17 PM »
Hydron,

Interesting information. I have a question about the term matching. Are you talking about electrical impedance
matching or matching the two devices  by their measured characteristics ie pairing similar devices?

In either case, what are the techniques of matching?

Cheers.

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 413
  • Karma: +15/-0
    • View Profile
Re: Parallel IGBJT
« Reply #5 on: July 14, 2021, 06:08:28 PM »
I was meaning picking parts with similar electrical characteristics.

In my comment about sharing switching losses I was mostly referring to being lucky (it would be a pain to measure dynamic characteristics), but you could try and do some loose matching of some DC characteristics, e.g. Vce(sat) of the IGBT or Vf of the diode.

I've never actually bothered trying to match IGBTs myself (the only parallel parts I've used have had forced current sharing using a split MMC), but If I were to have a go at it I'd clamp the devices to a heatsink to keep the temperature equal, and run say 10A through them, checking the forward voltage and/or the Vce(sat) (the latter would need to be at a fixed gate voltage) and just pick pairs that are closest to each other. Quite possibly not worth doing though - they may already be close enough to share usefully as is, especially if they're from the same batch or something.

Offline alan sailer

  • High Voltage Technician
  • ***
  • Posts: 111
  • Karma: +7/-0
    • View Profile
Re: Parallel IGBJT
« Reply #6 on: July 14, 2021, 06:28:17 PM »
Hydron,

Thanks for the clarification. I have a loose background in microwaves so to me matching means impedance.

I'll think about testing. But will probably not. The devices are all well soldered into the bridge assembly, with snubbers, TVS,
GDTs and all. It would be a pain to pull them.

Like I mentioned the assembly was an experiment to see if Loneoceans PCB was any different than Alex Yuans.
The easiest way to check this was just to build up two complete bridge assemblies. I was getting a lot of (expected)
transients in a phase shift QCW build and wanted to see if board layout would help. It didn't...

Cheers.

Online davekni

  • High Voltage Expert
  • ******
  • Posts: 1294
  • Karma: +62/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Parallel IGBJT
« Reply #7 on: July 14, 2021, 07:04:34 PM »
If your FGA60N65SMD parts were all from a single order, there is a good chance they are well enough matched.  Even if from different orders, it is not too likely to include enough variation in forward drop to be problematic.

I was thinking more about matching impedance in layout.  The dynamic impedance (slope of V/I curve) is about 0.013 ohms on the FGA60N65SMD typical graph.  At 159kHz for example (1M rad/sec), 13nH has the same impedance magnitude.  If emitter and collector connections come from one side to the first IGBT of a pair, then on to the second IGBT, it wouldn't be too hard to get 13nH more inductance to the second IGBT.

BTW, once I used multiple MMCs for forced current sharing.  However, it wasn't for IGBT sharing, rather to force sharing in sections of a distributed primary in my low-frequency QCW experiment.
David Knierim

Offline alan sailer

  • High Voltage Technician
  • ***
  • Posts: 111
  • Karma: +7/-0
    • View Profile
Re: Parallel IGBJT
« Reply #8 on: July 14, 2021, 07:49:58 PM »
David,

I'll give the PCB a look and see if anything like that was done. According to some old wire bond tables 13nH is 2cm.

Cheers.

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1823
  • Karma: +47/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Parallel IGBJT
« Reply #9 on: July 14, 2021, 09:19:36 PM »
This is a interesting application note from SEMIKRON and a quite extensive test: https://www.semikron.com/dl/service-support/downloads/download/semikron-application-note-modules-in-parallel-operation-with-central-and-individual-driver-board-en-2017-01-27-rev-00/

The guide lines with even inductance paths, symmetrical load paths and ensuring same environmental properties (same batch IGBTs and same heat sink etc.) is all something we have known for years. Its been well-known back from paralleling diodes that this is necessary ( https://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/rectifiers/#references )

I was surprised by the papers conclusion that pulse testing shows the worst characteristics and full on inverter operation is more stable due to less delta-T and high load inductance. But can we translate that into an analogy of DRSSTC vs. QCWDRSSTC and how these are made with single bricks vs. many small parallel and low on-time vs. long on-time? Its apparently not just about lowering switching losses from high frequency parts, but also that the long on-times keep it stable (given that its not grossly abused and burns to the ground)
https://kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 413
  • Karma: +15/-0
    • View Profile
Re: Parallel IGBJT
« Reply #10 on: July 14, 2021, 11:41:07 PM »
Good read, thanks for pointing it out. Also makes me think that I should maybe play around with double pulse tests too when playing with PCB design trying to tame transients, easier than having a coil running next to you during development!

Online davekni

  • High Voltage Expert
  • ******
  • Posts: 1294
  • Karma: +62/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Parallel IGBJT
« Reply #11 on: July 15, 2021, 02:41:26 AM »
Quote
But can we translate that into an analogy of DRSSTC vs. QCWDRSSTC and how these are made with single bricks vs. many small parallel and low on-time vs. long on-time?
Thank you for the great app-note.  For phase-shift QCW, I suspect most of the considerations are valid, just scaled way up in frequency.  For ZCS bridges (most DRSSTC designs), many of the switching time issues are less critical, such as different gate plateau voltages (due to low current at the switching point).  That is likely why I get away with direct paralleling of sets of 10 IGBTs without separate gate resistors.  On the other hand, DRSSTC has high load frequency, same as switching frequency.  (Compared to the app-note where 3kHz switching was shown with ~50Hz load frequency.)  This high load frequency makes emitter and collector inductance matching still important even under ZCS conditions.

Simulations would be great.  However, I've had poor luck with the couple IGBT models I've downloaded - slow simulation and occasionally convergence issues.
David Knierim

High Voltage Forum

Re: Parallel IGBJT
« Reply #11 on: July 15, 2021, 02:41:26 AM »

 


* Recent Topics and Posts

post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
Today at 05:34:54 AM
post prodc9 interrupter plexi case
[Dual Resonant Solid State Tesla coils (DRSSTC)]
costas_p
September 24, 2021, 07:14:53 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Uspring
September 24, 2021, 05:37:56 PM
post Re: Tesla plate
[General Chat]
Benbmw
September 24, 2021, 05:20:01 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Netzpfuscher
September 24, 2021, 07:26:30 AM
post Tesla plate
[General Chat]
Duane B
September 24, 2021, 01:54:49 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 23, 2021, 09:08:23 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Uspring
September 23, 2021, 07:09:23 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 23, 2021, 05:46:38 PM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 23, 2021, 11:12:25 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Netzpfuscher
September 23, 2021, 08:49:51 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 23, 2021, 06:15:34 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 23, 2021, 05:48:40 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Uspring
September 22, 2021, 07:21:12 PM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 22, 2021, 10:41:20 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 22, 2021, 07:07:48 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 22, 2021, 05:26:40 AM
post Re: Measuring TC secondary parameters
[General Chat]
MRMILSTAR
September 22, 2021, 05:14:30 AM
post Measuring TC secondary parameters
[General Chat]
Duane B
September 22, 2021, 05:01:22 AM
post Re: Load resistance and discharge terminal for my 14-stage CW multiplier
[Voltage Multipliers]
MRMILSTAR
September 21, 2021, 10:33:34 PM
post Re: Futurist's DRSSTC 0.5
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
September 21, 2021, 07:21:50 PM
post Re: Load resistance and discharge terminal for my 14-stage CW multiplier
[Voltage Multipliers]
klugesmith
September 21, 2021, 05:24:35 PM
post Re: BrOdin coil - High power Big sparks!
[Dual Resonant Solid State Tesla coils (DRSSTC)]
futurist
September 21, 2021, 04:18:47 PM
post Load resistance and discharge terminal for my 14-stage CW multiplier
[Voltage Multipliers]
MRMILSTAR
September 21, 2021, 06:30:44 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 21, 2021, 06:04:52 AM
post Re: BrOdin coil - High power Big sparks!
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
September 21, 2021, 03:56:28 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
johnnyzoo
September 20, 2021, 09:24:10 PM
post Re: BrOdin coil - High power Big sparks!
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Uspring
September 20, 2021, 05:07:43 PM
post Re: Large Rectifier Tube
[General Chat]
Benbmw
September 20, 2021, 04:12:27 PM
post Re: Proper phase lead adjustment at the input
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Chrisader
September 20, 2021, 10:40:33 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 20, 2021, 10:39:59 AM
post Re: BrOdin coil - High power Big sparks!
[Dual Resonant Solid State Tesla coils (DRSSTC)]
futurist
September 20, 2021, 10:38:31 AM
post Re: Large Rectifier Tube
[General Chat]
klugesmith
September 20, 2021, 07:07:55 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 20, 2021, 06:03:57 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 20, 2021, 05:31:57 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 20, 2021, 04:59:17 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 20, 2021, 04:28:03 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 20, 2021, 12:53:39 AM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
johnnyzoo
September 19, 2021, 11:09:37 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
AstRii
September 19, 2021, 09:47:14 PM
post Re: Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 19, 2021, 08:28:05 PM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 19, 2021, 08:19:01 PM
post Idea for QCW DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
AstRii
September 19, 2021, 07:58:09 PM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 19, 2021, 03:16:40 PM
post Re: Driverless MOSFET SSTC
[Beginners]
AstRii
September 19, 2021, 01:32:47 PM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 19, 2021, 10:15:24 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
Magneticitist
September 19, 2021, 06:32:18 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
davekni
September 19, 2021, 06:06:44 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
AFreshLad
September 19, 2021, 04:14:18 AM
post Re: Help with SSTC feedback circuits with 4046 IC and Schmitt Triggers
[Solid State Tesla Coils (SSTC)]
Magneticitist
September 18, 2021, 11:33:37 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