QCW IGBT Selection
zzz_julian_zzz, Thu Jul 31 2014, 04:34PMHi,
I'm planning to use an IGBT switch for my full bridge QCW ~350kHz, just a quick question before i buy the parts.
2 IGBTs came visible to my requirement:
1. IRGP50B60PD1PBF - Some coilers swears to it, good feedback heard from them, due to its low losses
2. FGH60N60SFD (NOT SMD)- Never seen a coiler used it before, but it has higher current rating than the one above, lastly, cheaper.
I'm planning to set the OCD about 300 Amps., plus I plan to parallel them. I hope you can help me decide guys, because i'm buying for bulk and i don't want to lose money to wrong / lesser components. Thanks in advance! happy coiling to everyone!
Re: QCW IGBT Selection
Mads Barnkob, Thu Jul 31 2014, 04:56PM
I have used the IXGN60N60 minibloc IGBTs in a medium DRSSTC at 70 kHz and a small at 300 kHz, see my site for details on the coils.
I think people stated blowing these up at 600-700A, as for QCW duty I have no experience.
One of the important parameters are also the thermal characteristics, where these two you linked are about identical, but you could also aim for some with even lower ratings which is even more important in a QCW with its long on-time.
Mads Barnkob, Thu Jul 31 2014, 04:56PM
I have used the IXGN60N60 minibloc IGBTs in a medium DRSSTC at 70 kHz and a small at 300 kHz, see my site for details on the coils.
I think people stated blowing these up at 600-700A, as for QCW duty I have no experience.
One of the important parameters are also the thermal characteristics, where these two you linked are about identical, but you could also aim for some with even lower ratings which is even more important in a QCW with its long on-time.
Re: QCW IGBT Selection
Dr. Dark Current, Thu Jul 31 2014, 06:14PM
Compare
1) Switching loss (@ the same Ic)
2) Thermal impedance @ the required pulse length. When you're at it, you can also calculate the temp rise.
In both cases less is better.
Dr. Dark Current, Thu Jul 31 2014, 06:14PM
Compare
1) Switching loss (@ the same Ic)
2) Thermal impedance @ the required pulse length. When you're at it, you can also calculate the temp rise.
In both cases less is better.
Re: QCW IGBT Selection
teravolt, Thu Jul 31 2014, 07:11PM
I use these HGTG27N120BN they have a 500 watt rating and they are for switching purposes. I get mine from ebay
teravolt, Thu Jul 31 2014, 07:11PM
I use these HGTG27N120BN they have a 500 watt rating and they are for switching purposes. I get mine from ebay
Re: QCW IGBT Selection
Steve Conner, Thu Jul 31 2014, 09:11PM
Check out the new Ixys XPT parts too. But beware, the bigger ones don't have antiparallel diodes.
Steve Conner, Thu Jul 31 2014, 09:11PM
Check out the new Ixys XPT parts too. But beware, the bigger ones don't have antiparallel diodes.
Re: QCW IGBT Selection
Steve Ward, Wed Aug 06 2014, 02:44AM
Julian, those 2 parts look like they could be exact replicas of eachother, i think either would be perfectly suitable for the application.
If you want to run more than a few % duty cycle, then cooling becomes more challanging, and you may want more devices in parallel to spread the heat out.
Here is my rough recommendation for those parts assuming ~350khz and 450V max:
Low duty, ZCS only : 120A per device (peak)
High duty, ZCS only: 60A per device.
Low duty, Phase shift (hard switching): 60A per device
High Duty Phase shift: 30A per device.
My calculations for the additional switching losses for phase shift bridge control suggested the power dissipation is a bit over double what it would be for ZCS. When you factor in the higher Vce at more amps, it says you should run them at half the current when hard switching. Again, these are very rough estimates.
Also note that simply paralleling IGBTs does not guarantee perfect sharing, so the more you parallel the more you should derate them. You could use individual gate drives and use extra means of forcing current share if you think its worth it (i do in some cases). Also, going beyond 2 parallel devices starts to get harder to force them to share because of layout inductance mismatch.
Steve Ward, Wed Aug 06 2014, 02:44AM
Julian, those 2 parts look like they could be exact replicas of eachother, i think either would be perfectly suitable for the application.
If you want to run more than a few % duty cycle, then cooling becomes more challanging, and you may want more devices in parallel to spread the heat out.
Here is my rough recommendation for those parts assuming ~350khz and 450V max:
Low duty, ZCS only : 120A per device (peak)
High duty, ZCS only: 60A per device.
Low duty, Phase shift (hard switching): 60A per device
High Duty Phase shift: 30A per device.
My calculations for the additional switching losses for phase shift bridge control suggested the power dissipation is a bit over double what it would be for ZCS. When you factor in the higher Vce at more amps, it says you should run them at half the current when hard switching. Again, these are very rough estimates.
Also note that simply paralleling IGBTs does not guarantee perfect sharing, so the more you parallel the more you should derate them. You could use individual gate drives and use extra means of forcing current share if you think its worth it (i do in some cases). Also, going beyond 2 parallel devices starts to get harder to force them to share because of layout inductance mismatch.
Print this page