I'm attempting to drive an H-Bridge that I made with 2 P-Channel and 2 N-Channel MOSFETs directly from an AVR microcontroller.

During my original testing, I applied the full 5V to the gates and it worked like a charm.

As I added PWM functionality, I'm now realizing that I possibly have a design flaw. I'm driving the motors by a potentiometer to control the PWM duty cycle.

I'm finding that if I increase the voltage that the H-Bridge is switching from 12V to 24V and decrease the duty cycle, the motor behavior is erratic as I move the potentiometer up and down and the N-Channel MOSFETS get very hot with no load.

Fully on and Fully off (5V or 0V at the gates) works great (cool MOSFETS).

Will changing the PWM frequency help? Should I go faster or slower? I'm able to hear a faint whine in the motors so it's within audible range.

Any other suggestions?

 

We need to see your schematic.

 

Good Point, Ron. I have attached the schematic. Thanks!

 

What frequency are you using and what are you cranking it upto. The N mosfets are getting hot then it implies the P mosfets are conducting.

Have a look at the waveform on the collector of your P driver transistors. If its slewing then it conducting during the transition period and possibly switching both N & P on at the same time.

Alternately try reducing the base and collector resistors by 1/2 and see if that makes a difference.

When I have used such a circuit, I have used mosfet drivers to drive the gates of both the N&P.

 

If you are only giving the irf530, 5V to the gate that is your trouble. The irf530 is a standard mosfet and needs to have ~10V gate voltage to turn completely on. If you need to have only 5V on the gate you must change to a "logic level" mosfet. Or use a mosfet driver and take the 12V supply from the motor supply. Only using 5V on the 530 gate is just barely turning it on, making it a big resistor.

With a micro you are better off using drivers any way. You can get a higher amperage level to the gates that way. More amps on the gate means faster switching. Faster switching gets you less heat in the mosfet.

 

Your p-channel gate pullup resistors are WAY too big. Power MOSFETs have a lot of gate capacitance. Your p-channels will take forever to turn off. What PWM frequency are you planning to use?
There are a lot of things to consider when building a discrete H-bridge. You have to build in dead time, so to don't get shoot-through conduction on the sides of the H. This is complicated by the fact that your level translator transistors (the NPNs) have storage time, and may take several microseconds to turn off. You should have snubber (flyback) diodes.
Have you considered buying an integrated H-bridge, or an H-bridge driver?

 

Thanks all. It seems it is in the design after all. I'm calculating frequency at 63khz.

I will try to change resistors and lower freq first. Thanks again.

 

if you are going to run this on 24V, you need to limit gate voltage. IRF530 absolute max Vgs is +/-20V.

 

if you are going to run this on 24V, you need to limit gate voltage. IRF530 absolute max Vgs is +/-20V.

Same for the IRF9640.

 

Does that mean I can only use 20v to drive motors or do I need to limit the voltage on the pull-up resistors?

 

Does that mean I can only use 20v to drive motors or do I need to limit the voltage on the pull-up resistors?

You need to limit the voltage on the pull-up resistors.

 

You could use a 7812 regulator from the 24V supply to get 12 volts for the gates, 12 volts is a good voltage to use.