openinverter forum

Hi!

I already have my controller running in the car. It works very well and smoothly.
The problem I have is that the power stage gets too hot.
I am in the stage of tuning the motor/controller values, I would appreciate if you could help me to define which parameter is the one that can influence more in the overheating of the power stage fets.
At this moment I am checking if I have the correct deadtime with the oscilloscope.
The power stage is implemented with 24 stb75nf75 mosfets, so 4 in parallel.
I don't have much data about the motor other than its maximum speed, voltage, power and number of poles.

thanks a lot

Checking deadtime is a good place to start.

Edit - deadtime and pwm frequency are the obvious parameters affecting power dissipation. The slower the device switching the higher the deadtime needed and the greater the switching losses. With higher switching losses a lower pwm frequency can be beneficial. FETs should be pretty fast though, (if driven right).

Could really do with more information to be able to offer advise though.

What power are you trying to put through the FETs?
How are you driving the FETs?
How are you cooling the FETs?

Pete9008 wrote: ↑Tue Dec 06, 2022 11:50 am Checking deadtime is a good place to start.

Edit - deadtime and pwm frequency are the obvious parameters affecting power dissipation. The slower the device switching the higher the deadtime needed and the greater the switching losses. With higher switching losses a lower pwm frequency can be beneficial. FETs should be pretty fast though, (if driven right).

Could really do with more information to be able to offer advise though.

What power are you trying to put through the FETs?
How are you driving the FETs?
How are you cooling the FETs?

hi!

thank you very much for your time.
All the hardware of the controller is commercial, it is a Danaher ACD4805 from a golf cart to which I removed the Texas microcontroller and wired a bluepill.
That's why I know that it is a matter of setting parameters until it works well, since originally everything worked perfect, same motor, same fets, same drivers, etc.

The PWM frequency of the original controller is 8KHz, so I'm at 8.8KHz.

It is difficult for me to determine the deadtime, according to what I saw in the datasheet of the mosfets, I have the rise time, fall time, Ton and Toff, is it ok to say that the minimum deadtime would be the risetime plus the fall time?

Each set of 8 fets (4 for + and 4 for -) is driven by an IR2010S.

The main factors that will influence heat dissipation are:
* Deadtime. I believe deadtime should be the greater of "ton + rise" of "toff + fall" under ideal circumstances, but I'd be setting it a bit higher.
* Drive strength. the rise and fall times are based on an ideal driver, If your driver is less powerful, you might need more deadtime and you will get more heat.
* Frequency. A lower frequency means less heat. 8.8kHz should be a good value though.
* Your cooling system. Obviously the better your cooling, the lower your temperatures.
* Power factor. If your control loop isn't well optimized, it's quite possible that you have lots of current flowing theough the FETs but not actually producing any useful output.

Deadtime should be easy to check; just increase it and see if the problem goes away, if it doesn't, put it back.

It seems to me that you don't have control over all of these factors so focus first on deadtime. Once you've confirmed that deadtime is okay, start looking at your AC vs DC current. If you have a lot of AC current recirculating but not doing useful work, you may need to improve your angle/voltage control.

Is this induction or FOC?