Motor Controller 29 Failure Analysis

During Driving practice Today, me and my teammate discovered that one of our motors had a sudden loss of reverse ability. I quickly assumed that the motor controller was at fault. Then came the smell of electronics burning. We quickly unplugged the Controller and put it out on the ledge of our outside window to cool it down. When we retrieved it and cracked it open, we discovered that one of the drive MOSFET’s had failed catastrophically. (A MOSFET is a device that switches electricity with electricity.) I believe the specific part is a Fairchild FDS4935BZ N-Channel PowerTrench MOSFET, for anyone who is curious. As I am sure you are aware, these motor controllers burn up all of the time, my team has been through as many as three or four in the last year. All of which were sudden failures. I can only assume that the same part failed. Which would mean a crazy amount of current was drawn through the MOSFET. All I can think of right now is a design flaw or an external issue which caused a direct short to ground.

Does anybody have experiences with motor controllers catching fire or smoking?

Has anybody taken them apart after they have failed?

Does anybody have any ideas as to why they fail so frequently?

Thank You,

-56E

Picture of the failed PCB. Note the failed SO-8 chip on the top Left of the board.

[ATTACH]8895[/ATTACH]
FullSizeRender.jpg

Yes, it happens. You damaged one of the MOSFETs.

Yes
Motor Controller Problems - post #9

Students abuse the motor wires (indirectly, things like picking up motors by the wire). The motor wires fray then short. The MC29 then makes magic smoke.

Not a huge deal, The chip itself costs about $1. Not beyond economical repair as long as the PCB itself wasn’t damaged in the fire.

I have a feeling that our motors were used and abused before we got ahold of them. This would explain the short required to blow out the MOSFET. But, the PTC in the Cortex should have kicked in and stopped this from happening. I don’t think it did. Perhaps the current spike was far to fast for the PTC to do anything. Even if this is the case, the MOSFET is rated at 50 Amps pulsed current.

-56E

The short is usually where the wires enter the motor, this is before the PTC so it doesn’t help.

Given H-bridge vunerability, having it embedded in a 29 motor controller ($10), rather than a 393 motor ($20+10) or a cortex ($400) is highly preferable. Otherwise, you get problems like this. Thankfully, you only have to worry about the internal H-bridges on ports 1 and 10.