• State TI Thinks Resolved
  • Locked Locked
  • Replies 13 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 803 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DRV8323R: Shorted half bridge CSD88599Q5DC, can’t reproduce.

Martin123
Prodigy 175 points
Other Parts Discussed in Thread: DRV8323R, DRV8323, CSD88599Q5DC, DRV8353, DRV8353R
Part Number: DRV8323R
Hi, we have used DRV8323RH in a lithium 48V system.
 
During field testing we had issues where at random times, both under high load and low load one mosfet would short and the 10A fuse would blow.
Just had to change the mosfet and system worked again.
 
We assumed that this was because of the high voltage and have now redesigned for DRVB8353RS with 100V mosfets.
 
Tried yesterday to confirm that the high voltage was the issue in a test bench with a single motor but could not get the mosfet to short, even with a full battery at 58.1V straight out of charger.
Varied the motor load from zero to 10+ amps.
 
So now we are concerned to what the issue was.
 
This is an overview of wiring with the issue:
Battery -> contactor -> motor1 -> motor2
                 contactor -> motor3 -> motor4
 
Contactor controls both + and – of the two strings and there is just one battery.
 
Wirelength battery to contactor is 2 meters
Wirelength contactor motor1 is 3 meters
Wirelength motor1 to motor2 is 2 meters
 
Wirelength contactor motor3 is 7 meters
Wirelength motor3 to motor4 is 2 meters
 
Can say for sure but probably just one motor has run when the short have occurred.
 
We have tried with a long cable(7 meter) from battery to motor in the lab to simulate the final setup.
This created a VPP of around 2V still no shorted mosfet.
With short wires approx. 50cm, VPP is 300mV.
  • 0
    TI__Mastermind 23045 points

    Hello Martin,

    Thanks for the question and for posting to the MD forum. I am not sure what the question is here or if I understood correctly, but if the previous system encountered a problem using the DRV8323 and the CSD88599Q5DC this was most likely due to the voltage ratings of the MOSFET. According to the CSD88599Q5DC datasheet this MOSFET has an absolute max voltage of 60V.

    On a worst case scenario you mentioned the battery provides 58V which is awfully close to the 60V max of the MOSFET, this means it did not have enough voltage margin in case if the motor generates any additional BEMF voltage or any parasitic inductance on the PCB causes a voltage spike (typically we recommend have a margin of 1.5 to 2 times the voltage). Now that you have moved to the DRV8353 and a 100V rated MOSFET you are most likely not seeing this issue because your new MOSFET is able to handle the voltage generated by the motor or any parasitic inductance that could have damaged the system previously.

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    Prodigy 175 points

    Hi Isaac, 

    Sorry if I did not explain it well.

    We wanted to check if it was the voltage that damaged CSD88599Q5DC mosfets, so we are now trying the reproduce the issue in the lab with the old controller that is based on DRV8323 and CSD88599Q5DC.

    We are not able to “destroy” the controller, so we are worried that switching to DRV8353 and 100V mosfet will not fix the issue since it’s not reproduceable with the old controller.

    After some searching, we read that long battery wires may damage ESC controllers, that’s why I included wire lengths. 

  • 0 in reply to Martin123
    TI__Mastermind 23045 points

    Hey Martin,

    Thanks for the clarification, I had misunderstood which system you were attempting to replicate the issue on. 

    Martin123 said:
    We have tried with a long cable(7 meter) from battery to motor in the lab to simulate the final setup.
    This created a VPP of around 2V still no shorted mosfet.
    With short wires approx. 50cm, VPP is 300mV.

    Can you explain where this voltage is being measured in the quote above?

    Was there any damage to the DRV8323 or just the MOSFET and the fuse?

    Could you provide a schematic?

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    Prodigy 175 points

    Hi Isaac

    Measurement is done at pcb inputs 3 cm from capacitors and DRV8323.

    There was no damage to DRV8323, just had to replace the CSD88599Q5DC that was shorted and fuse.

    DRV8323-DRV8323.pdf

  • 0 in reply to Martin123
    TI__Mastermind 23045 points

    Thanks for the info Martin! Sorry I was not clear enough the 2V and 300mV measurement is this for GHx, VCP, GLx, SPx, SNx? Just trying to understand where the length of this wire made a difference.

    Which MOSFET was damaged, high side or low side?

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    Prodigy 175 points

    The measurement is done at battery input, + and –. It’s the battery wires that is long.

    It was random what high side mosfet got damaged. When measuring from GHx to GND there was a short.

  • 0 in reply to Martin123
    TI__Mastermind 23045 points

    Hey Martin,

    I will review the information provided and get back to you tomorrow. Thanks!

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    TI__Mastermind 23045 points

    Hey Martin,

    I apologize I have not had a chance to review the schematic yet. Is it okay if I get back to you early next week? Most likely on Tuesday or Wednesday since we will be out for holiday on Monday the 17th.

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    Prodigy 175 points

    Hi Isaac, 

    no problem, we can wait.

  • 0 in reply to Martin123
    TI__Mastermind 23045 points

    Thank you Martin. 

  • 0 in reply to Isaac Lara
    TI__Mastermind 23045 points

    Hey Martin,

    You mentioned a fuse in your comments above but when reviewing your schematic I didnt find one and just a couple of relays. Are these the ones you were referring to?

    Have you been able to figure out which scenario caused your MOSFET to short and blow your fuse? Have you tried loading your motor and not allowing it to move? This will cause the current to increase and perhaps cause a short.

    Best,

    Isaac

  • 0 in reply to Isaac Lara
    Prodigy 175 points
    Hi Isaac,
     
    Sorry for the slow reply.
    There is a blade fuse on the supply wires.
     
    We might have found something, that have the highest probability now to be the reason for mosfet short.
    Our motors are hand wound, and some show an intermittent short to stator core.
     
    At lab the motor is not connected to ground but where we had the issue its grounded.
     
    This short is not easily reproducible as it comes and goes.
    There is 50-100% larger current draw when this error occurs when motor is connected to ground.
     
    We will continue with the new motor controller based on DRV8353R and change out the coils of our prototype motors.

    Back to wire length from battery, do we need to think about the length, or should it be fine?

  • 0 in reply to Martin123
    TI__Mastermind 23045 points

    Hey Martin,

    No worries, glad you are making progress to finding the source of the problem. Feel free to keep us posted if there is anything else you might need assistance with.

    The main key care about for the wire length from the battery would be to ensure that you have enough bulk capacitance in your systems to handle the added inductance from the wire length. Aside from that if they supply is steady and the cables are rated to handle the necessary current there should not be much of a problem.

    I hope this helps!

    Best,

    Isaac