Power MOSFET blew up when increased duty cycle from 5% to 10%

[XaviPacheco]

I've got these waveforms so far, input and output of the gate driver. I haven't been able to capture Vge yet. The new IGBT hasn't failed yet. (That's why I'm now referring to Vge instead of Vgs) Seems like the MOSFET I used before didn't tolerate the kickbacks from the motor. I'll report back when I get Vge measurements.

[xavier60]

The Collector waveforms might show something also.

[AndersJ]

Your pwm frequency is high.
The motor inductance may be so high that you do not get a square wave output.
In that case the transistor is operating in its linear region, thus running hot.

[XaviPacheco]

Why is the PWM frequency high? The oscilloscope says about 10 kHz (9.7 kHz), and that’s what I actually chose it to be.

[AndersJ]

10 KHz may be high for a:
- Mosfet, depending on gate capacitance and drive impedance
- "Large" inductance like a motor.

I'm only guessing.
You are the only one that can verify this.

I have pwm'ed small motors at around 300 Hz.
Higher frequencies did not work well.
That was 27 years ago.
I don't remember the details.

Why did you choose 10 KHz?

[xavier60]

For the amount of wire that DC  motors have in them, the inductance can be rather small. If the PWM frequency is too low, the motor will draw large discrete current pulses which causes high I squared R losses in the motor windings and everything else in the circuit path and poor load regulation.
As the frequency is increased, the current becomes continuous with lower peaks and load regulation improves.  Going unnecessarily high will cause loses to increase with no real advantage.
I have found the minimum to be 5Khz with the motors that I have used.
I guess it depends on the motor. 300Hz might be suitable for some.

[XaviPacheco]

Actually you've made a good point regarding the PWM frequency. I didn't find anything that tells you the suitable PWM frequency straight forward, because it depends on many factors. I just did a kind of research and came up with 10 kHz. I've attached the nameplate of the motor I'm controlling now.

[XaviPacheco]

I haven't mentioned that I plan to measure speed and compensate duty cycle according to the load.

[Wolfram]

How do you change the duty cycle? If you abruptly step the duty cycle by a certain amount, the motor will draw a proportional fraction of its stall current until its speed reaches the new value.

[T3sl4co1l]

Indeed, the very first thing you're doing by adjusting PWM is adjusting the flux into the winding inductance.  Get this wrong and current overshoots and *pop*.

The best way (the only true way) to control an inductive load, is to control the current into it.  A Hall affect current sensor could be used to measure motor current, and the PWM adjusted to maintain that at some set current.  This inner loop is very fast (comparable to the clock frequency), so it can protect the transistor even in the event of a short circuited load!

The current, in turn, is set by a second loop that controls whatever parameter is of interest: output voltage (especially for power supplies), torque (torque is very close to current, actually, but a torque servo would address bearing and windage losses), RPM (very close to voltage, except for winding DCR and magnetic leakage), etc.

The two-loop architecture is very easily implemented in a microcontroller, or can be done discrete with a dual op-amp.  Note that you need a relatively fast MCU to monitor current safely: ATmega won't cut it, the ADC is just too damn slow.  But Xmega will, or pretty much anything else out there (dsPIC, ARM..).

Tim

[XaviPacheco]

As shown in my schematic, I'm using the TLI4970 current sensor to monitor the current. How effective it is, I still don't know. I'm working on it. I'm using a L432Kc nucleo for the project.

[T3sl4co1l]

Right, but that's monitoring source current, so it doesn't sense motor current when the transistor is off.

Source side peak current sensing is alright for protection, but you need quick response.  Usually a comparator and flip-flop (consider UC3842's block diagram).  Some MCUs have hardware configurable for this, but no way in hell you'll get away with doing it in software (e.g., analog comparator triggers interrupt, or let alone ADC polling current signal*).

*Unless it were something like a dsPIC that can sample and process quite fast (single digit µs response time).

Average current mode control allows you to respond more slowly, at least if you have some minimum inductance in series with the motor.  Which is a good idea if the motor is on a long cable, by the way -- extra filtering, just make it a buck supply in the first place and run the motor on clean power.

Since you're already using a Hall effect sensor, moving it to the motor is trivial.

Tim

[XaviPacheco]

Let me see if I get the idea.. do you mean moving the sensor to the drain/collector leg?

[T3sl4co1l]

Probably between DC+ and motor, since that will have less noise.

Tim

[XaviPacheco]

Ok. I will take it into consideration for the next design, as I already have the PCB with the sensor in the source leg. Thank you.