Blew out power mosfets at low power?

[GoingFastTurningLeft]

I have a simple motor interface circuit I designed to control the motors in a massage pad with the intent of getting tactile feedback from a racing simulator. Its an AVR connected to some FQP30N06L (60V LOGIC N-Channel) power mosfets via 1.5K resistors (what I had sitting around). Motors have a +12VDC common and are connected to the drain terminal, source is ground. Motors are about 100 ohms.

It was working great for a while (i wrote some good code for wall impacts ), but this last time I turned it on I noticed the motors were running a lot higher than usual even when sending it a 1% duty ratio signal at 35Hz. Oscilloscope showed the drain was hovering around 5V (not 12 as I'd expect) and dropping to ground when it was supposed to, and the AVR's pwm signal was fine. 4 out of 5 MOSFETs are showing near 0 ohms from gate to source, the other shows near 1M. My code never let it go above 70% duty ratio, I've never observed it go over 50%.

I've narrowed it down to 2 things:
1. Static electricity from the pad's foam somehow killed the mosfets the last time I got up/sat down.
2. I made some horrible error in my design. These are power mosfets rated 60V and 32A and are supposed to be able to run off 5V logic signals. Rdson is 0.045 ohms. The mosfets I chose should be overkill for this.

Any ideas what happened? I'm leaning towards #2.

[dyarker]

Zero Ohms gate to source sounds like voltage spike damage. Do you have reversed diodes across the motor?

Less likely, is switching speed fast enough to prevent FETs from getting hot during on-off and off-on transitions?

Just a couple ideas. Cheers,

[GoingFastTurningLeft]

I haven't added any protection. These are not big motors, all they do is spin an off-center weight to induce vibration, I don't know what the inrush is but steady state they draw about 115mA @ 12VDC. The only other parts are a small disc cap across motor terminals. Originally they were wired through a switch that either connected them to 12V (high speed) or to 12V through a 1/2W resistor (low speed). I am re-using the original power supply, rated 12V @ 800mA.

Maybe I'm underestimating these motors...

[CeaSaR]

Small or not, they are still inductive and will cause a voltage spike when
the power is removed. Even the smallest "pager" motor will spike if not
clamped by a diode.

An ounce of prevention....

CeaSaR

[haklesup]

While I can't rule out an ESD (static electricity) event it would have to be a pretty big zap to kill 4 out of 5 FETS in circuit. One at least big enough to notice and feel (many ESD events are below the human detection level but big enough to kill a loose part). In circuit even without explicit protection clamping diodes would still be protected by the clamp diodes built into the chip you use to drive the gates (depending on exactly what you used) and parasitic capacitance from the traces and other components also would soak up some of that energy. ESD is of greatest risk for loose parts.

60V is for Vds and 32A is for Id. What is the spec for Vgs? that's what likely got exceeded to kill the gates. Some FETs have built in protection diodes but many do not. Its often a good idea to use a zener from gate to ground or a regualr diode from gate to Vdd to give any excess voltage a place to go if it should appear. A reverse biased diode across the motor will clamp inductive kick back at the source as well.

Such is the difference between a circuit that works and one that is reliable. What FET P/N were you using, can you link a datasheet?

[GoingFastTurningLeft]

I just pulled them out and checked them again, looks like I was confused testing them in circuit. DS is shorted on 3, everything is shorted on the 4th. May have overheated it removing it.

Its an AVR connected to some FQP30N06L (60V LOGIC N-Channel) power mosfets via 1.5K resistors.

http://www.fairchildsemi.com/ds/FQ%2FFQP30N06L.pdf


Vgs = 10 V. My AVR is running on 5V.

All I have are some signal diodes at the moment, I'm going to throw those on the control board itself and hope my replacements won't kick the bucket.

Whats really strange is that I blew Channel 2-5. I'm using channels 1, 2, 4, 5.