So recently I was experimenting with some 100W LED's (30-35 VDC @ 3A) and a "1200w" DC-DC boost converter that goes up to 85V.

While at an input of ~12 VDC and an output of ~26 VDC, I accidentally shorted the output, and the main MOSFET blew and is now shorted out. I un-soldered it from the board, to see the model, but while I was able to find the spec sheet, I was not able to find the MOSFET for sale for a reasonable price (and I only need one).

I put another MOSFET that I had laying around in it (I don't know the part number), and it worked, but I'm almost certain that it isn't rated for such high voltages, as it was creating a lot of heat (no load) at anything above maybe 30 VDC and started dropping when it got to ~70 VDC.

The boost converter is rated up to 85 VDC (which is the limit of the original MOSFET).

Out of a random moment of insanity, I purchased a IRFP4468 which is rated for a slightly higher voltage (100v VS the 85 of the original) and a higher amperage.

The original MOSFET is an NCE85H21TC

Will the IRFP4468 work in place of the original MOSFET, or did I waste my money?

I followed where the pub goes and traced it in the included picture. I don't have the schematic and drawing it by hand would take days.

The converter had two 15A fuses in parallel in the input, but I bypassed them and enlarged the positive rail with more solder, as it was heating up and causing a voltage drop. I'm pretty sure that even without the bypassed fuses, the MOSFET would've died, as the cables I was using are extremely thin, about half of a spaghetti noodle, including the coating, so I'm sure I wasn't passing the ~15A output needed to run through those cables in order for the 30A input fuses to blow.

Thanks for the help in advance.



  • \$\begingroup\$ It's almost the same but much higher input pF \$\endgroup\$ – Tony Stewart EE75 Jan 18 '17 at 2:34
  • 1
    \$\begingroup\$ Please edit the question, use the tool and draw a schematic \$\endgroup\$ – Voltage Spike Jan 18 '17 at 4:32
  • \$\begingroup\$ Please provide spec sheet for NCE85H21TC. | THe IRF... has a reasonable chance of working, but more information will be useful. Installing it and riunning at close tp zero load intially should allow some scoping of what is happening wrt drive etc. \$\endgroup\$ – Russell McMahon Jan 18 '17 at 9:44
  • \$\begingroup\$ NCE85H21 data sheet and [IRFP4468 data sheet](eet**](ncepower.com/Upload/MOSFET/NCE85H21TCdatasheet-16543235225.pdf) and ). | The IR part is general;ly superior and has a good chance of working well. Most relevant specs are better or no worse. HOWEVER, as Tony says the IRF part has substantially higher input capacitance and this MAY affect operation deep-ending on the drive circuitry. ... \$\endgroup\$ – Russell McMahon Jan 18 '17 at 9:57
  • \$\begingroup\$ ... Also, the NCE part claims to have extremely good single pulse avalanche energy capability. This probably does not matter AND may not be true. The IRF part has their typically very detailed information on single and repetitive avalanche energy withstand and the NCE part has a single claimed parameter for single cycle only. AND you hope this is largely irrelevant in this circuit regardless. | SO: The IRF part has a good chance of working well, YMMV, AC&NR ... :-) \$\endgroup\$ – Russell McMahon Jan 18 '17 at 9:59

Given what you have provided, there is simply no way to tell.

On the one hand, your replacement FET has a nominal Rds(on) about 30% lower than the original, which means it should (all things being equal) dissipate about 30% less power. But, of course, not all things are equal. As Tony Stewart commented, the replacement FET input capacitance is nearly 3 times greater. This means that it's entirely possible that the gate driver will not provide a fast enough turn-on and turn-off, and dynamic heating may be much greater. Or maybe not. Without details of the gate drive it's all a big mystery.

And just as a passing comment, beefing up traces with solder isn't that effective, since the bulk resistivity of solder is about 10 times that of copper. You'll get better results if you lay down bare copper wire on the trace and solder that in place.

  • \$\begingroup\$ Thanks for that clarification. I have a picture where I traced the pcb from the gate transistors and found their part numbers. Sadly one of the wires goes into a no-name chip which I assume is programmable. I'll try to change one of the links and put the picture. \$\endgroup\$ – Bwinzey Jan 18 '17 at 4:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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