Author Topic: IGBTs for Electronic Loads?  (Read 2105 times)

0 Members and 1 Guest are viewing this topic.

Offline t1dTopic starter

  • Super Contributor
  • ***
  • Posts: 1255
  • Country: us
IGBTs for Electronic Loads?
« on: November 27, 2019, 05:47:30 am »
I have never used IGBTs, so I watched this Great Scott video.


He says (@ 6:33) IGBTs can handle higher voltage and current than MOSFETs. That made me wonder if they might be better for E-loads than MOSFETs. What do I need to consider?

Thanks for the education!
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: IGBTs for Electronic Loads?
« Reply #1 on: November 27, 2019, 05:56:24 am »
Check the SOA, an [analog] electronic load needs DC SOA.

IGBTs have the smallest dies of all silicon transistors = lowest power dissipation for a given rating, besides other characteristics that make them undesirable in such service.

Despite that, I have surprisingly seen some IGBTs with DC SOA curves.  I guess if you want to...

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: IGBTs for Electronic Loads?
« Reply #2 on: November 27, 2019, 06:51:26 am »
It is possible if you get just the right kind of IGBT, otherwise no.

Most power IGBTs(And also MOSFETs) out there are actually internally built in the form of a large arrays of tiny transistors on the same die. Unfortunately IGBTs are particularly temperature and tolerance sensitive so when operated in the linear region the many tiny transistors inside will not share the load evenly, causing some small areas to get hotter that causes them to turn on even more and take more current, get even hotter...etc going into a thermal runaway until one small spot overheats and destroys the transistor. All this can happen at only a fraction of the rated dissipated power. Similar thing can happen with MOSFETs that are heavily optimized for switching operation.

Usually MOSFETs are less sensitive to this hotspot thermal runaway and so are better suited for DC operation. If a given MOSFET is designed for being operated at DC then the Safe Operating Area (SOA) will show a curve for DC on it and usually this will be mentioned on the first page of the datasheet along with perhaps calling it a Linear FET. Usually such FETs are significantly more expensive as they are more difficult to produce and generally require larger silicon dies

But a transistor won't instantly blow up on DC if it has no DC SOA specified in the datasheet. Some switching transistors are runaway resilient enough that they actually work just fine in a wide linear region. The manufacturer won't grantee this so you have to buy about 10 to 100 of these transistors and stress test them in your own lab all the way to destruction to find out how far up the SOA chart you can get before they explode. There are a few transistor part number out there that are known to be good at this.
 

Offline t1dTopic starter

  • Super Contributor
  • ***
  • Posts: 1255
  • Country: us
Re: IGBTs for Electronic Loads?
« Reply #3 on: November 27, 2019, 07:02:25 am »
Most power IGBTs(And also MOSFETs) out there are actually internally built in the form of a large arrays of tiny transistors on the same die. Unfortunately IGBTs are particularly temperature and tolerance sensitive so when operated in the linear region the many tiny transistors inside will not share the load evenly, causing some small areas to get hotter that causes them to turn on even more and take more current, get even hotter...etc going into a thermal runaway until one small spot overheats and destroys the transistor. All this can happen at only a fraction of the rated dissipated power. Similar thing can happen with MOSFETs that are heavily optimized for switching operation.
Good info. Thanks!
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17429
  • Country: us
  • DavidH
Re: IGBTs for Electronic Loads?
« Reply #4 on: December 03, 2019, 01:45:15 am »
Despite that, I have surprisingly seen some IGBTs with DC SOA curves.  I guess if you want to...

Since it is not something they are designed for, they are not tested for DC SOA so I would not trust the DC SOA curve unless the manufacturer specifically discussed it.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 7454
  • Country: pl
Re: IGBTs for Electronic Loads?
« Reply #5 on: December 03, 2019, 08:46:06 am »
Believe it or not, Toshiba used to make complementary N/P-channel linear IGBTs for "high power amplifier applications" like the GT20D101/GT20D201.

Used to :P

edit
A good place to look for information on linear transistors of any sort is audio forums like diyaudio.com. You could perhaps find some known-good N-channel IGBT still in production if you look around, if you really can't live without one ;)

AFAIK P-channel IGBTs aren't made anymore. Not sure why, it would seem that since NPN power transistors are better than PNP, P-ch IGBTs could be better than N-ch, but apparently not :-//
« Last Edit: December 03, 2019, 09:12:18 am by magic »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17429
  • Country: us
  • DavidH
Re: IGBTs for Electronic Loads?
« Reply #6 on: December 04, 2019, 06:26:44 pm »
Believe it or not, Toshiba used to make complementary N/P-channel linear IGBTs for "high power amplifier applications" like the GT20D101/GT20D201.

Used to :P

They did but I think it was always a dumb engineering idea.  Maybe marketing sold it to them.

The chief advantage of IGBTs over MOSFETs is that they are less expensive for a given voltage and current however linear applications are limited by power dissipation which is proportional to die area so using IGBTs was always going to be more expensive than MOSFETs or especially bipolar transistors.

Quote
AFAIK P-channel IGBTs aren't made anymore. Not sure why, it would seem that since NPN power transistors are better than PNP, P-ch IGBTs could be better than N-ch, but apparently not :-//

I think the issue is that NPN and PNP power transistors are relatively close in performance for a given die area compared to n-channel and p-channel MOSFETs; p-channel MOSFETs are like 3 times the size of an otherwise equivalent n-channel MOSFETs.  So in an IGBTs, swapping the n-channel MOSFET for a p-channel MOSFET is a larger disadvantage than swapping the NPN element for a PNP element.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf