Using 80NF10 in place of IRF3710

[mfnisar]

We tried to replace IRF3710 with STP80NF10, in our full bridge 72VDC inverter with switching frequency of 16kHz and IR2110 as drivers but the latter doesn’t work and gets damage during switching.

After several modifications we now have added a 1nF mylar capacitor across gate to source of the 8010 now the performance is much better but still the long term reliability of 80NF10 does not match 3710. What could be the cause, in terms of specs 80NF10 are superior to 3710.

Could anyone kindly spare some time to inspect the datasheet of both the devices.

 

[alexan_e]

It may have to do with the turn-on (td-on) and turn-off delay (td-off) of the devices.
The figures for 80NF10 are three times worse:

td-on=40ns vs 12ns
td-off=134ns vs 45ns

If the deatime is very small (or it doesn't exist) then you may get cross conducting because the speed of the mosfets is slower.
I'm not sure if this is the case but I think it is a factor to consider.

Alex

 

[mfnisar]

Thanks a lot Alexan_e. Have checked and verified the dead time it is 1.5uSec I think dead time is large enough.

 

[KerimF]

Usually I work with BJT. So when I have to replace a transistor with another one of higher power; I refine its driver to provide a higher current and preferably a faster one if the load was also increased (as power).
In case of MOSFET, I think the driving voltage (and its current) will likely need a sort of boost so that the 'transient' power (of the MOSFET and for the same load) will be again much like in the case of using the more sensitive MOSFET but of lower power.

Kerim

AddedAlex
I may say, I am also a bit surprised that adding 1nF made things better.

 

[alexan_e]

I'm not sure what is the problem, the 80NF10 has a higher Qg=140ns(typ) to 189ns(max) compared to the IRF3710 Qg=130ms(max) but this shouldn't be a problem.
Also the OP said that using a 1nF capacitor between gate-source makes the circuit work better, this capacitor would slow the operation of the mosfet even more.
Can it be related to the bootstrap capacitor, do you have the same problem with all mosfets (upper and lower half) or only the high side?

Alex

 

[mfnisar]

We have problem with both high and low side. As the circuit is full bridge one side high and low devices repeatedly gets damage

 

[alexan_e]

Is it always the same side that burns, have you tried to replace IR2110?
You can also replace the bootstrap capacitors, if something goes wrong with the upper mosfet and it burns then when the low side mosfet conducts in the same side it will be burned too.

Alex

 

[mfnisar]

Thanks a lot again Alexan_e. To add here is the fact that with same circuit board and IR2110 we have tried IRF3710, FDP2532, IRFP260, IRFB4710 all the MOSFETS works fine. The only devices causing problems is 80NF10, I have also tried to approach ST customer support but no reply from their side as well.

If there was some transient noise related issue then FDP2532 could have also caused problems.

The mystery remains unresolved

 

[alexan_e]

Can you be sure about the quality of the STP80NF10, can they be counterfeit components?
Have you used these mosfets in any other circuit?

Alex

 

[mfnisar]

These devices cannot be counterfeit components? As these were purchased from DIGIKEY. In the past however we have purchased these devices from different sources but the results were not different.

One more detail that I missed is, when these devices are operated as a single device these perform much better. When you parallel these devices the performance worsens, in our circuit 3 devices are paralleled.

We haven't used these in other circuits.

 

[mfnisar]

What I have concluded is by adding 1nF across gate-source lessens the effect of gate-drain parasitic capacitance. However when compared to other MOSFETS the performance of these devices is still not to the mark.

 

[mfnisar]

Why is the ST support no responding to emails?
Anybody with a clue....