Designing a low voltage protector

Thread Starter

arman74

Joined Jul 1, 2017
30
Hey guys.
We designed a driver for BLDC motors recently and we replace ours with the destroyed one. We were searching for why the original driver would damaged and we have just one reasonable answer for this.These drivers may work 1 month or some years,so the thing damaging these drivers is something rare.
Below is the schematic of the h bridge controlling the motor.We guess that a spike on the power supply cause the p channel mosfets burn.

The n channel gate and the base signal comes from a microcontroller with 5 v regulated supply.So if the 24v drops to 18volts or less the micro controller is still sending signal to the mosfets.The n channel gate-source voltage is enough to keep it in saturation area with low Rds(on) but the pmos source gate voltage drop around 3 volts and much more Rds(on)and burns.
So we need a circuit to break the 24v in case of low voltage detection and it must work as fast as it can.What do you suggest?
 

Attachments

crutschow

Joined Mar 14, 2008
23,347
So if the 24v drops to 18volts or less the micro controller is still sending signal to the mosfets.The n channel gate-source voltage is enough to keep it in saturation area with low Rds(on) but the pmos source gate voltage drop around 3 volts and much more Rds(on)and burns.
How did you determine that?
The LTspice simulation below shows that the P-MOSFET gate-source voltage (yellow trace) is constant at over 12V down to below a 17V supply voltage, since the driver transistor acts as a constant current for the 1.5kΩ resistor.

So the supply voltage dropping to 18v is not the problem.

Do the N-MOSFET and P-MOSFET signals for the transistors on the same side have a non-overlap delay so that they are never momentarily on together?

upload_2019-9-18_8-18-42.png
 
Last edited:

Thread Starter

arman74

Joined Jul 1, 2017
30
So the supply voltage dropping to 18v is not the problem.
Thanks for the simulation.So I did a simulation with pspice when the 24 volts drops to 5 volts the mcu still works(it works even with 2.5 volts),so is this the problem?
Do the N-MOSFET and P-MOSFET signals for the transistors on the same side have a non-overlap delay so that they are never momentarily on together?
Yes they do.The motors work properly.Maybe 10% of them will crash and the crash time varies from even 1 week to years.For years we can say that the device is old,but a lot of them wont survive more than 1 year.
 

crutschow

Joined Mar 14, 2008
23,347
Thanks for the simulation.So I did a simulation with pspice when the 24 volts drops to 5 volts the mcu still works(it works even with 2.5 volts),so is this the problem?
Possibly.
Why is the voltage dropping?
Perhaps you need a better power supply.

If you need to monitor the voltage then you could do that with the mcu if it has an A/D input.
 

Thread Starter

arman74

Joined Jul 1, 2017
30
Why is the voltage dropping?
As you said maybe the power supply is not doing well.
If you need to monitor the voltage then you could do that with the mcu if it has an A/D input.
Is the mcu a/d fast enough to do this?I was thinking of a comparator oppamp that is connected to the mca external interrupt so i can have a delay after dropout detection?Any suggestions?
Can you guess what other reasons would destroy the circuit?
 
Top