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.

DRV8848: OCP in parallel mode

Part Number: DRV8848

I was able to test the DRV8848 in OCP (shunt resistors are shorted and current is exceeding 2A). I see the fault pin pulling low and the retry occurring at every 1.6ms -- all good when not in parallel mode.

When using parallel mode I see some different behavior. I understand each bridge has its own OCP monitoring and act independently, but I want to make sure what I am seeing is expected. 

Yellow is the output

Blue is the nFAULT pin

Green is the current (measured with a current probe).

First picture: 1ms/div time scale. You can see the nFAULT pin asserted for 1.6ms (low), which is expected. The output is also low. The next interval is retry, fault pin is high, the output is with an odd waveform and so is the current. The duration of this interval is not consistent - it varies a lot. The pictures of the scope are a single shot. 

Second picture is the same situation, but with 10us/div time base.

Is this behavior expected? Is it justified by the two H bridges interfering with each other?

Details of the setup:

VM = 8V

Load = 5A (electronic Chroma 63101)

Board: DRV8848 boosterpack

I appreciate the help!

  • Lenio,

    Are you putting the device into parallel mode per figure 7 in the datasheet?  Just want to make sure if you are doing this OR just directly connecting outputs together.  

    We have seen some goofy behavior when using electronic loads.  Do you happen to have some old school power resistors to do the testing?



  • Ryan,

    yes, the setup is just like fig 7.

    I followed your advice and replaced the electronic load with power resistors. The results are the same. It seems the bridges are not in sync when tripping, so the total current for OCP is not twice of a single bridge, but rather two stages of 2A each tripping at a different time. 

    Here is an interesting result in which the load is about 3A, therefore in between 2A and 4A. One of the bridges might be tripping while the other isn't, maybe? The nFAULT pin is oscillating, visibly through the LED on the board (blue line). Yellow is the output, green is the current.

    If I increase the load beyond 4A I get the same result I showed in the first post. I'm curious about that double pulse I highlight in the picture. I have seen it in all my attempts to test OCP in parallel mode. The current also looks really weird, with negative swings. Aside from parasitic inductances, the load is purely resistive.  nFAULT is constantly high. 

    Do we have OCP test results in parallel mode?

    Thank you! 


    I repositioned the current probe from my previous experiments. Before the current probe was at the output of the power supply. For this new set of tests, I put it very close to the load. I still see a negative swing, but the format is much more in sync with the output. One curious aspect that provides evidence to my theory of the two bridges acting independently is a degree of randomness in the curves. Here is a sequence of two scope shots in which nothing has changed, just two distinct attempts: 

  • Hi, 

    Thank you for your post. Our expert is reviewing your question and feedback to you on early of next week.



  • Hi Lenio,

    Apologies for late reply. I've been looking into your waveforms but I don't have a clear explanation. Looks to be an issues with the H-bridge not being in sync but I can't say for sure.

    For now, let's try to separate the load with the driver. Can you disconnect the load and keep the same set-up as before. Drive the outputs and measure the outputs. Are the outputs stable?

    If outputs are stable, the adding the load is causing the issues and we need to investigate why.

  • To close and document this thread, the issues occur in the 3 to 5A range, in which the DRV8848 trips on thermal shutdown rather than OCP. Once the current exceeds this range, OCP works as expected.