H-bridge mosfets blowing

I made a motor controller using IR2113's as the h-bridge controllers. This circuit has been working fine and I have done a lot of testing on it. Normally I would turn off the PWM pulse before plugging in the 24V for the motor (the motor draws about 10 amps), but this time I forgot and ended up plugging in the motor when the PWM was at a maximum. Of course this caused the plug to spark as it made contact with the battery terminal, and the mosfets on my motor controller blew.

Is it likely (or possible) that the mosfets blew as a result of plugging in the motors' power supply when the PWM was on already? I have checked every other aspect of the circuit and cannot find a problem (and as I said, the controller has been working fine for a while before this happened)

You forgot to tell which MosFet transistors you are using.

Boncuk

Boncuk said:

You forgot to tell which MosFet transistors you are using.

Boncuk

Click to expand...

Transistors are STP55NF06

The Ir2113 probably took pretty good care of you but nothing good ever happens when you have those big arcs around an h-bridge. I can only guess that it may be the inductance in the power line that created the break down. Do you have big electrolytics across the 24 volt supply at the H-bridge?

The transistors should easily handle the 20A you are pulling.

However they require fast recovery diodes across the inductive load (motor).

If you start the motor at full power (100% duty cycle) the motor current is six to eight times the nominal current (full load) resulting in extremely high back EMF, which might destroy the transistors.

They are rated 60V which might be exceeded quickly upon start up.

I measured the back EMF at a 24VDC solenoid of 5W with peaks up to 130V! Motors might behave more nasty!

Boncuk

Do you have big electrolytics across the 24 volt supply at the H-bridge?

Click to expand...

I dont at the moment, but should I (as general good practice)? What sort of size capacitance?

However they require fast recovery diodes across the inductive load (motor).

Click to expand...

I am using 1N4148 diodes - are these sufficient? They did not blow when the mosfets did.

If you start the motor at full power (100% duty cycle) the motor current is six to eight times the nominal current (full load) resulting in extremely high back EMF, which might destroy the transistors.

Click to expand...

How do I measure the back emf? Do I just connect an oscope across the Drain-Source terminals of the mosfet and look for voltage spikes?

> I dont at the moment

When you PWM an inductive load, the back-emf must go somewhere. Usually, with a large supply capacitor, the back-EMF energy is simply stored back in the capacitor. Without a capacitor of suficient value, the supply voltage will raise until something burns.

> I am using 1N4148 diodes - are these sufficient?

Definitely not, use Schottky diodes, rating depends on how you drive the bridge.

peufeu said:

Definitely not, use Schottky diodes, rating depends on how you drive the bridge.

Click to expand...

Sorry I was looking at the wrong diodes - I was actually using a 1N4001... is this adequate? I see it's not a Schottky diode

With a good synchronous driver you can do without the diodes, the MOSFET's body diodes will do the job just fine (check the ratings in the datasheet), it will just be a little bis noisier and you'll lose a little efficiency, nothing really bad.

1N4007 is probably much worse (and slower) than the MOSFET body diodes anyway. So the 1N4007 are probably lowering your efficiency.

WHat happened is probably this :

When you connect the battery, it does not establish a perfect contact instantaneously : you'll get a few cycles of intermittent contact. When 24V is applied to the motor, the current rises in it. When the battery contact breaks for an instant the motor current cannot stop instantaneously, the energy has to go somewhere. Without a suitable capacitor on the 24V line, this will generate a huge voltage spike (negative or positive), perhaps hundreds of volts, which will destroy your MOSFETs.

To calculate the capacitor we need to know the motor's inductance and PWM frequency.

Don't use Schottky diodes as fast recovery diodes.

Instead use MR850 series fast recovery rectifiers. The MR856 can stand 600V at 200A pulses.

Using an electrolytic cap to recover the back EMF it has to be rated not to exceed the maximum allowable voltage.

200 and 385V electrolytics are huge (diameter/height 40/55mm for 1,000µF/200V and diameter/height 40/105mm for 1,000µF/385V)

The MR856 measures LXW of 9.5X5mm.

Use your scope to measure EMF peaks directly at the motor.

Boncuk

Thanks for the replies, it is very helpful
I replaced the mosfets that blew yesterday and was testing the back emf peaks this morning... I have 2 motors and a controller for each - I connected the one motor/controller up to test for the back emf, while leaving the other motor controller still connected to the 24V supply but not to its own motor. I slowly increased the PWM duty cycle and noted how the back emf increased from about 20V to 60V (when the duty cycle was only 50% roughly). Strangely though, the controller that was *not* connected to the motor blew (one mosfet actually caught fire) when the back emf reached about 60V. I checked in the data sheets and saw that the Vds rating was 55V - does this show that back emf is my problem? Even though it was induced by the other motor controller's motor?

Well, without a suitable capacitor to decouple your +24V rail and absorb the back EMF spikes, depending on the quality of the battery contacts (vibration-caused intermittent open circuits, etc), wire inductance, battery internal resistance, battery capacity to react to transients, etc, your +24V is only pure theory, since during MOSFET switching, it could vary from minus a few hundred to plus a few hundred volts, so yes, anything that sits on it will blow...

What is the motor inductance and maximum current ?

A schematic might be in order so we can all get on the same page. Are you using a 24 volt power supply or batteries?