incinerated fet in a 24 fet infineon

[1000w]

Yesterday I went for a trouble free 9km test ride on this, viewtopic.php?f=28&t=34451&start=60 the controller was not even warm after 9km.
Then in the afternoon, motor and controller cool, I went up and down the street twice stopped the bike then went to take off again and the controller seemed to shut off.
I turned the power to the controller off to do a reset, and when I turned it back on a nasty loud arcing noise came out of the controller so I very quickly killed the power as my hand was still on the main switch.
Here is the result.




Does anyone have any clue to what could have caused this and where can I get IRFB4110 fets?
Cheers,
Matt.P.

[bigmoose]

What was your battery voltage?

For sourcing Arrow Electronics has them for a buck and a half apiece in unit quantities, but no stock.

Future has them for $3 apiece in 25 quantity and around $2 each in 200 part quantity. Both are good vendors with certified stock.

[1000w]

What was your battery voltage?

20s of A123, 66volts.

Cheers bigmoose.

[Gregory]

I bought some off Abraham (voicecoils?) Best choice for an aussie supplier. I need more....

[1000w]

I just talked to Aussie Jester and he has some spare 4110's so I'm sorted.
I would like to know what what happened though.
I don't know if it matters for the health of the controller, but I should add that I forgot to use the precharge resistor when the fet blew.

[1000w]

Ok I replaced 4 of the 8 fets on the yellow phase, then connected the battery, throttle and motor.
As soon as I twisted the throttle the other 4 fets on the yellow phase blew, I saw it this time as I had the cover off the controller, quite spectacular.
I have also opened the motor, which did get uncomfortable to hold my hand on heat wise in earlier testing, and it looks good inside.
Not cooked and no obvious areas of arcing.
How do I test if a phase is shorted or damaged? Could this be the problem causing the fets to pop?
I'm lost ATM.
Cheers.

[Jeremy Harris]

When the FETs blow on these boards they sometimes take out the gate drive resistor, too. Also, it's not uncommon for several FETs that look OK to have been seriously stressed by the failure, leading to their early demise.

[Gregory]

Can't help on testing.
But I did order 100 x irfb4110 after my last post because voicecoils was out. Should be here from usa in a week. Let me know of you need replacements.
Greg

[1000w]

When the FETs blow on these boards they sometimes take out the gate drive resistor, too. Also, it's not uncommon for several FETs that look OK to have been seriously stressed by the failure, leading to their early demise.

Thanks Jeremy, I guess I should have replaced all 8 fets on the phase.
Will the 4 new fets that I installed after the first failure now be stuffed because the other 4 blew with the second failure?
If the gate drive resistor is stuffed will this blow fresh fets?
Can someone tell me where the gate drive resistor is? How do I test it?

Can't help on testing.
But I did order 100 x irfb4110 after my last post because voicecoils was out. Should be here from usa in a week. Let me know of you need replacements.
Greg

Thanks heaps Greg. I have 8 more fresh fets from Aussie Jester as he gave me a 12 fet infineon to salvage parts from.
If I get through all those I may be in touch.

I have also opened the motor, which did get uncomfortable to hold my hand on heat wise in earlier testing, and it looks good inside.
Not cooked and no obvious areas of arcing.
How do I test if a phase is shorted or damaged? Could this be the problem causing the fets to pop?
I'm lost ATM.

Cheers,

[Thecaptain115]

Gate resistor is connected to the gate. Use the datasheet to determine which leg of the fet is the gate and follow the trace. From there, just start looking for shorts and melted parts. Use an ohmmeter to see if the resistor is shorted (0 ohm). If it is, replace with the same value resistor

[Jeremy Harris]

The gate resistor advice above is sound. If it were me I'd replace both the high and low side FETs is a phase that had suffered a failure, as the failure of one side is pretty much guaranteed to stress the living daylights out of the FETs on the other side.

[Gregory]

Edit... Ignore this - I'm wrong!

[heathyoung]

A TO-220 package usually is GDS (Gate Drain Source).

If you blew the rest of the FETS you may have had a shoot through (which can be spectacular!) Both high side and lowside FETS conduct at once, shorting the battery. Gate drive resistors or gate drive transistors can cause this messy situation. Check to see if any of the gates are energised when no throttle is applied - there will be a voltage on one of the phase leads to ground or B+.

[1000w]

To check the motor phases just put a multi metre on continuity setting , cycle through the phase combinations and check that it doesnt beep/resistance goes low.

When I do this on my motor and a spare motor every phase combination on both motors beep, isn't this normal???

Check to see if any of the gates are energised when no throttle is applied - there will be a voltage on one of the phase leads to ground or B+.

Ok connected up the controller again, tiny bit of smoke but no noise from the damaged fets. The blown phase, yellow, has 63 volt to the battery negative and the other two phases, blue and green, have 63 volts to the battery positive.
Does this help the diagnosis?
I really appreciate peoples help here as my electronic knowledge is fairly limited.
Cheers,
Matt.

[Gregory]

To check the motor phases just put a multi metre on continuity setting , cycle through the phase combinations and check that it doesnt beep/resistance goes low.

When I do this on my motor and a spare motor every phase combination on both motors beep, isn't this normal???

Sorry mate, my bad. It is normal they are all connected.

I'd better scurry off and learn some BLDC Theory 101.

Greg

[heathyoung]

To check the motor phases just put a multi metre on continuity setting , cycle through the phase combinations and check that it doesnt beep/resistance goes low.

When I do this on my motor and a spare motor every phase combination on both motors beep, isn't this normal???

Check to see if any of the gates are energised when no throttle is applied - there will be a voltage on one of the phase leads to ground or B+.

Ok connected up the controller again, tiny bit of smoke but no noise from the damaged fets. The blown phase, yellow, has 63 volt to the battery negative and the other two phases, blue and green, have 63 volts to the battery positive.
Does this help the diagnosis?
I really appreciate peoples help here as my electronic knowledge is fairly limited.
Cheers,
Matt.

Interesting. Not normal. Not good either.

OK - to determine if those are the gates energised or if those FETS are shorted, check the resistance between each phase and battery + and - ve.

Anything showing a reading here will be a problem (more cooked FETS). Also check between the gate (first pin of the FET) and the other pins (drain and source) - if you get continuity here, there is a very good chance that you will have problems with the gate drive resistors, and possibly transistors (as the full pack voltage has ended up here).

If these show no shorts, then you have a gate drive issue. Usually once you get to this point, it gets painful. The gate drive is usually a pair of transistors, arranged in something called a totem pole - these cheapies don't use proper gate driver IC's (at least not the ones I have fixed). What will happen is that one of the transistors will short (and when the other is energised, it will die as well) - worst comes to the worst, this will back-feed the microcontroller IO pins if not well protected.

Finding these transistors, and determining if they are shorted is a pain. You can determine if it is switching properly with an appropiatly sized resistor and an LED, with the halls connected (i actually have a pic microcontroller that fakes legitimate hall signals) and spin the wheel (no phases connected, just halls), with some throttle (so not zero PWM, but PWM only applies to one of the sides, not both - I think it is the lowside from memory) you should see the LED's flash when each MOSFET is meant to fire - any that don't flash, or stay stuck on - are problematic.