Inverter MOSFET

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
See the pic.
IRFP450 & IRF540
These are in an inverter I am attempting to fix.
My question is the good are showing different on resistance.
Could it be that in these some are partially faulty or not.
How can I check the on resistance to guarantee that they will last once fixed with new ones. Should I replace all of em or not.
Any thoughts appreciated. :D
 

Attachments

beenthere

Joined Apr 20, 2004
15,819
The definitive check of a FET is the body diode. If you can see the junction between source and drain (with N channel, the diode cathode is toward the drain), the device should be good. They fail all the way short or all the way open - nothing in between.

You will probably have to pull them all to read the body diodes. Boiling won't work.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Umm beenthere.. this is seperate, the boiling thing ended up pretty funny.

All the diodes in the good ones are OK, they all show equal Vf.

When testing MOSFET with meter, the trick I uses tell me if the mosfet is switching and in most of 'em this works, gives equal Vf. But these ones some are different, The Rsdon, are different, could it be the MOSFET are partially faulty.
 

beenthere

Joined Apr 20, 2004
15,819
If the gate is blown (is this common?), couldn't the diode still be intact?
Just from personal experience over 12 years or so, all the failures I observed left all three terminals as a dead short. I was switching 3600 watts, so maybe that is why.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Any fet with ANY Rds will cook when Rifaa sticks them in the microwave. ;)
Now the microwave thing huh!!. You lots are always trying to find an angle. :D
Don't worry you will keep seeing these things from me occasionally, this really does spice up my technical life.
None of you ever do these kind of things.
I cooked 2 pc MPM3004 H bridges trying to built a bi directional drive to my winder. It was only after that I thought of PWM it.:p

About the MOSFET's the gate's are intact. The faulty ones have shorted gates or S & D.
They are switching but the different Vf tells me the on resistance are different even though I am charging the gates to the same Vg.
I just use diode range of my fluke to charge the gate. and them measure the DS.
I think I will discard all of em to be on the safe side, But It wouldn't hurt to check em with a supply. would it?
 

retched

Joined Dec 5, 2009
5,208
Now the microwave thing huh!!. You lots are always trying to find an angle.
Woah.. We dont need to FIND an angle, you GIVE us everything we need! ;)

As for the MOSFETs, crap-can them. Better off replacing them all. You dont want a customer because it blows in two hours.
 
Could be that different measured resistances between the drain and source is because of different gate charges on the different devices.

You could charge each gate up to 10V and see if the resistances are all similarly low.
Then you could short the gate to the source and see if the resistances are all similarly high.

Always measuring resistance with the positive lead on the drain, since these are both N channel devices.

You might want to put a high-value resistor on the gate drive while testing, to make sure you can see any blown gates. Those are huge delicate gates with 20V limits. ESD could get them especially if it's dry where you are.
 

retched

Joined Dec 5, 2009
5,208
He is not testing them in circuit, and while your method of checking does work, It doesn't reliably detect remaining life. And good practice would be to replace every one for the customers sake.

He has only 2 types of fets, and an inverter will put them through their paces.

If you are trouble-shooting the device, then by all means test them to see where the fault occurred.
 
He is not testing them in circuit, and while your method of checking does work, It doesn't reliably detect remaining life. And good practice would be to replace every one for the customers sake.

He has only 2 types of fets, and an inverter will put them through their paces.

If you are trouble-shooting the device, then by all means test them to see where the fault occurred.
I don't know how you'd detect remaining life on a FET. If the gate isn't blown and if you put +10V from gate to source and see mOhms from drain to source it should be fine, I think.

If you don't control the gate then they could measure all sorts of resistances due to different gate charges, so the person thinks some of them are blown.

If it's mission-critical then replace them, but if it's mission-critical I figured they wouldn't be posting here.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Hi Sarg....
As far as I know you are the one to look for any MOSFET problem.
So here goes.
The inverter I am fixing have a little road bump.
U see this thing have two types of FET's.
The IRF540 is used to step up the DC Voltage and IRFP450 are used at the inverter output drivers. These are driven by a SG3525 PWM and act as 50Hz High power Amps giving out 230VAC. Similar to an audio push pull stage look the circuit has.
Any ways I am going to replace the 540 with IRFP260
and IRFP450 will be replaced by IRFP460.

My concern is the later as the Vgs voltage is different. So I like to know what would be the consequence should I install them and power it up. Will I blow them or not. :D
I have attach the data for you.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
Going from the IRF540 to the IRFP260, you will more than triple your total gate charge (72nC -> 234nC), which will slow down the switching times.

The SG3525 can source or sink up to 500mA. That's not a lot if you are switching a whole bunch of MOSFETs. You might have excessive power dissipation in the MOSFETs.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
So u are saying that 540 are at the DC to DC converter which boosts the 12V to 230V DC or so will have problem if I put IRFP260 instead.
And I thought the 450 and 460 will create a problem due to higher gate voltage.
Now you are creating a bit of a new problem for me Sarg
 
Top