Replacing single mosfet on KU123

[nasunasu]

I have a KU123 controller which looks like it died the other day. I troubleshooted and removed the bad mosfet (type HY1707, another forum post said the KU123s use STP75NF75 as well).

My question is can I replace the bad one with a STP80NF10 (I have this on hand) and still be OK. That one is rated for higher voltage so shouldn't be the limiting factor.

Thanks.

 

[amberwolf]

It would probably be ok, but it might be safer to replace all of teh ones on the same phase leg (all the highside or all the lowside) with the same kind, so they share current properly.

 

[dnmun]

the 1707 is a knockoff i assume so you did not find a data sheet for it?

the 75nf75 is very common, and i assume the one you have is 80V, 100A from the label.

how did you test the mosfet you decided was dead or was it blown up already?

i would measure the body diode of each of the remaining mosfets to see how close they are in voltage and compare that to your 80nf10.

 

[nasunasu]

the 1707 is a knockoff i assume so you did not find a data sheet for it?

the 75nf75 is very common, and i assume the one you have is 80V, 100A from the label.

how did you test the mosfet you decided was dead or was it blown up already?

i would measure the body diode of each of the remaining mosfets to see how close they are in voltage and compare that to your 80nf10.

yes i believe it the 1707 is a knock off, i have found them on aliexpress as 75V 80A with the one I have 80V 100A.

my system was not working so i spun the wheel backward with lots of resistance, unplugged the three phase wires and it spun freely. isolated it to two mosfets on the positive/green phase line using a DMM. not sure if there was a way to differentiate which one of the two was bad but the first one i desoldered had continuity on all three pins and then continuity was lost on the other that was still soldered to the controller.

i tried it with the new mosfet unsoldered and the system works now. how do you measure the diode body voltages?

i think as a temporary fix i might solder the new mismatched mosfet until i get the right one

 

[dnmun]

you use the diode tester on your DVM. for an n-channel mosfet like this, the drain is always the higher voltage so you put the red probe on the drain and the black probe on the source.

when the diode tester pushes the current out of the probes, it will not flow because the mosfet is off as you would expect, and the meter reads open circuit.

if you reverse the probes, with the red on the source leg and the black on the drian leg then you can measure the forward bias of the body diode.

the forward bias will reflect the doping characteristics of the substrate of the mosfet which gives you a clue if it is similar to the others.

if it works you can just use it as it is. i expect the mosfet you replaced it with is fast enuff to work, so no need to replace it.

 

[nasunasu]

appreciate your help with this.

the new mosfet is reading 0.475 and the one it is paired with reads 0.488. The remaining 10 on the other legs are all about 0.470

 

[dnmun]

if the one that reads .488 is identical part to the others which read .47 then you should also replace the one that reads .488V also.

put another of the new mosfet side by side with the 80nf10 you used so both are identical. i thought you had 6 mosfets, not 12 so use identical side by side and the one at .488 is damaged too.