FET upgrade principles?

[Punx0r]

When looking to increase the current limit on a Chinese controller it seems like a good idea to replace the unknown-quality FETS with some quality parts.

What, though, are the important characteristics to observe when picking a replacement fet?

Voltage, current handling and RDSon seem obvious, but I'm guessing that going to big could harm the electronics that drive the fets?

The datasheets have a bewildering array of specs - can anyone suggest anything along the lines of "make sure the XX rating is similar to the original"?

These are the fets I wish to replace: http://pdf1.alldatasheet.com/datasheet-pdf/view/23894/STMICROELECTRONICS/STB140NF75-1.html (or at least, they are marked "140NF75").

Thanks,
Ant

 

[defuzo]

? High "dv/dt" value usetobe the subject, in switching pwm mode h-bridge set.

Whatever can drive HEXFET.

But better put some darlington inbetween some more complicated?

 

[boostjuice]

International Rectifier IRFb3077 is one of the best drop in replacement upgrade TO-220 FETs with a 75V rating that I know of.

When checking the eligibility of a replacement FET I also look at;

Some TO-220 FETs are package limited to 75A, others to 120A. In this case both the existing and potential replacement are 120A rated.

Qg (Total Gate Charge). A potential replacement FET should have this value kept not too much higher than the pre-existing FET being replaced, otherwise the drive circuitry may not be capable of driving the them fast enough - or at all - per switching transition (more important on fast electrical RPM motors [mid-drive/geared hub] than DD hub motors) where switching time becomes a larger loss component. If the drive circuitry has plenty of headroom you may be able to get away with much higher gate charge FETs, but without this knowledge confirmed it is best to play it safe. I use a rule of thumb that no more than 1.5x greater gate charge loading that the original FETs exhibited.

In this case, the IRFb3077 is almost identical to that of the STB140NF75-1 (160>220 nC vs 160>218 nC). All good here.

Also, make sure you compare RDSon at higher temps than the boldly published figure at room temperature. The FET's junction will be much hotter than it's casing :

IRFb3077 = ([2.2mΩ@25°C * ~2.8@160°C] = ~6.16mΩ junction temp at peak loading
STB140NF75-1 ([7.5mΩ@25°C * ~2.1@160°C] = ~15.75mΩ junction temp at peak loading

In this case, because the temperature coefficient transfer function follows similarly between both FET's, there is not much relative difference when scaled.

 

[Punx0r]

Thanks for that boostjuice - just what I was hoping for

defuzo, I'm just looking for a simple, straight swap of FETs. Nothing too complicated for this electronics newb