MOSFET and uC Fail with PWM Motors

[LACluc]

Hi all,
I am trying to drive a motor with PWM, and sometimes the mosfets burnt or the microcontroler fail the digital output (This port burns also).
I saw that working with 90% dutty is when mosfet more temperature gets, but if it is working 100% temperature is ok.
Microcntroler did not drive to the q3 because Vgs have to be more than 3.3V that's why I have Q13 Q15.
Q3 and Q4 are the power mosfets, N+P mos in soic 8 package.

I want to drive a motor with 2Amps and 24V.
2KHz PWM.
Microcontroler from NXP MKV11 family.
 
Can you help me please, I have burned lots of SH8MA3TB1 (Q4,Q3, N+P mosfet) and few uC !!! and I can discover exactly why.

Thanks

[LACluc]

SH8MA3TB1 DATASHEET

[LACluc]

Is there someone that can help me??

[xavier60]

From looking at the schematic, I see that the switching losses would be high even at a few KHz. I'm not going to do any calculations which I can only do crudely anyway.
The MCU damage is likely  being caused by a layout problem.
With the time and parts wasted trying to make  discrete circuitry work, you would have been better off using a made for purpose driver IC.
My pet part at the moment is the STGAP2D.

[xavier60]

Do you know the maximum motor current draw when stalled?

[xavier60]

Q13 and Q14 should have been omitted.

[xavier60]

One thing I missed, the turn off for the Nch MOSFETs will be slow.
First a long delay for the 10K resistor to discharge the Gate down to the plateau voltage plus about 10uS turning off.

[Avelino Sampaio]

Hi

This article can help:

http://prosje.be/Projects/BelastingSchakelen.html

[T_guttata]

Why don't you just use an N-channel MOSFET which is logic level rated? Then you would not need Q13, 14, 15, 16?
There are lots of MOSFETs which have very low Rds_on and are rated much higher than 6A!

[MadScientist]

I presume you have an allowance in the phases to prevent shoot through

[LACluc]

Hello,
Thank to all for answering.

* What do you mean when you talk about SWITCHING POWER LOSSE ?
And i do not underestand what is "allowance in the phases to prevent shoot through" 

*I decided not to use drivers for mosfets in order to have better stock options buying.
May be it was a mistake but now I can not change to it.

*maximun motor current = 2.8A

*I added Q13 and Q14 in order to push Vgs voltage because working with 3.3V mosfet over heat a lot.

*I change 10K resistor to 660ohm and it seem to work better, thanks for this comment.

*I think that all power mosfets are logic level, Vgth=2V


Now with 660ohm resist seem to work better. The only issue that I have is uControler ?
I do not knot why It fail, You can program at first use but is some moment you can not do it anymore althogh GPIOS are functioning ok. May be the debbung port burn??

[Zero999]

Switching losses occur because it takes a finite time for the MOSFETs to switch. Slower switching, at higher frequencies, results in higher the losses.

Shoot-through occurs when both the high and low side MOSFETs turn on simultaneously. There should be a delay between turning the high and low side on, to ensure both are not on at the same time.

Is 2.8A the maximum running current, or stall current? The stall current is normally much higher. If you don't know it, measure the resistance of the motor and work it out using Ohm's law.

Lower resistors will increase the switching speed.

The MOSFET threshold specified on the data sheet, is just the minimum voltage it starts to turn on at. A much higher voltage is required to meet the R_ON specification and achieve low conduction losses.

Use a MOSFET driver IC. There's a huge range of parts to choose from.

[batteksystem]

You can try to slow down the PWM if possible. Driving at a lower frequency may allow you to reduce switching loss.

[Doctorandus_P]

MOSfet gates have High DC resistance, but they also have a high capacitance, and during the chargeing of the gate capactiance a MOSfet only conducts partially and gets hot, because there is both voltage over, and current through the MOSfet.

For PWM'ming motor loads this time should be kept very short, and therefore high currents to (dis-) charge this capacitance are needed. Peak currents of 2A are not exceptional.

For this reason Gate driver IC's have been designed.
Often these gate driver IC's are designed for ah Half-H bridge and can control two fets, and they also have circuitry built-in to insert a "dead time" to make sure that the FET's are never opened at the same time.

[Carel]

I second the use of driver IC's. It simplifies everything: the 4 FET's are the same, high(er) voltage is kept away from the processor, the inbuild dead time.

Example: IR2104

[madires]

I fully agree, the high value gate resistors will slow down turn-on/off times and make the MOSFETs stay longer in the resistive operation mode. For low losses you want to switch the MOSFETs as fast as possible, while for lowering EMI you might prefer to hold back a little bit.

[Doctorandus_P]

In addition:

Gate voltages for MOSfets are generally limited to 20Vdc and they are very sensitive for over voltages.