I am designing a switch-mode welding power supply. The secondary current is 130 A at 24 V max and the switching frequency is 20 kHz. Using Schottky diodes for full bridge rectification results in around 500 W power dissipation whereas our max power is 3500 W. Also, the diodes for this rating are expensive so cost is an issue.

The diode I was considering before going for alternatives https://media.digikey.com/pdf/Data%20Sheets/International%20Rectifier%20PDFs/UFB200FA20.pdf

I was looking for an active rectification solution and the appropriate MOSFETs are IRFP3206. While looking into the datasheet I found out the body diode of MOSFET is rated for 200 A Continuous Source Current (as the diode is connected to the source). Reverse Recovery Time is 33 ns typ. Maximum Reverse Recovery Current, IRRM is 2.5 A (I don't know what that means).

The question is can I use these diodes for normal rectification? And if I were to use these MOSFETs for active rectification what properties should I be aware of about these body diodes?

The datasheet link is below and a graph about the body diode is included in the datasheet (Fig 21) https://www.infineon.com/dgdl/irfp3206pbf.pdf?fileId=5546d462533600a401535628d64a1ff0.

  • 2
    \$\begingroup\$ I found out the reverse protection diodes are rated for 200A <--It's not a reverse protection diode; it's a body (or bulk) diode. \$\endgroup\$
    – Andy aka
    Jan 10 at 12:29
  • 1
    \$\begingroup\$ IRRM stands for "Maximum Reverse Recovery Current" ... \$\endgroup\$
    – Antonio51
    Jan 10 at 13:19
  • 2
    \$\begingroup\$ How did you end up with 500W dissipated across the diodes with 130A through them? That sounds an order of magnitude off. Also note that this is why most high-current SMPS use a center-tapped secondary instead of a diode bridge: a bit of extra copper instead of the extra diode drop. \$\endgroup\$
    – TooTea
    Jan 10 at 13:42
  • \$\begingroup\$ @TooTea DC conduction related losses would probably be < 140W for this diode at 130A. But it's possible that there are some switching losses that add to it. Still 500W seems very high. \$\endgroup\$
    – user4574
    Jan 10 at 16:30
  • \$\begingroup\$ @TooTea 1v delrop per diode for 4 diodes, 1Vx4x130A=520W \$\endgroup\$ Jan 11 at 7:14

1 Answer 1


If gate and source are connected together there appears to be no reason you can't use the MOSFETs as a normal rectifier diode.

When I design a converter, I typically look at...

  • reverse recovery charge
  • reverse recovery time
  • forward voltage
  • leakage current

Its hard to do a direct comparison because the test conditions in each datasheet are different for each parameter. But overall the IRFP3206PbF body diode specs look as good as (if not better than) the UFB200FA20 diode.

One thing to note:
If you look at the datasheet, it says that you can put 200A continuous through the MOSFET, but there is a note #1 attached that says, "Bond wire current limit is 120A".

So, if you used this part for a 130A converter you would be beyond that 120A limit, which is probably not a good idea.

  • \$\begingroup\$ Yes I looked at that point. I plan on using two in parallel. Eight of those MOSFETs is still cheaper than four UFB diodes. And small size too. \$\endgroup\$ Jan 11 at 7:18
  • 1
    \$\begingroup\$ Also how can we pass 200A continuous current through the MOSFET if the bond wire current limit is just 120A? \$\endgroup\$ Jan 11 at 7:23
  • 3
    \$\begingroup\$ @Umer You can't, reliably. I couldn't find any package without that 120 A bond wire limit rating: an example of advanced specmanship even considering those *note 1*s. \$\endgroup\$
    – greybeard
    Jan 11 at 9:00
  • 1
    \$\begingroup\$ (See IR on current limits, too.) \$\endgroup\$
    – greybeard
    Jan 11 at 17:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.