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.

DRV8850: motor current measurement issue

Expert 5890 points
Part Number: DRV8850
Other Parts Discussed in Thread: DRV8873

We are using DRV8850 motor driver in our project. We are measuring current flowing through the motor, using VPROPI output. Schematics attached below. VMTR is 5.5V. The motor is connected to J2 and it’s a simple brushless DC motor.

In the already working project we observed that there is some asymmetry while measuring current. While the motor is running in one direction, the measured current seems to be up to 30% higher than the measurement for reverse current. Of course we suspected the motor itself in the first place. We switched the load to 3 Ohm resistor in series with current meter instead of motor. The results on VPROPI where the same: up to 25% difference between one direction and the other. For example: in one direction the measurement on VPROPI gives us about 1.4A, in the other we can see 1.8A (value close to the expected one). On the attached current meter the results where almost identical, around 1.8A in both directions.

The measurement was repeated on few different PCBs (the same production line, same batch, same product, different instance). Some of them show this issue, some of them gives us results with smaller difference, close to each other.

My question is:

Is it possible, that the difference in measurements on VPROPI output is caused by some internal characteristics of each motor driver, like Rds(on) difference between the FETs? Maybe it’s some kind of issue during PCB soldering/assembly process? Faulty batch? Or maybe those measurements are in the range of tolerance for this driver?

 As some additional info I would like to add, that we took 2 PCBs – one with strong measurements difference between current in both directions and one without this issue, soldered them out and then soldered to those 2 PCBs, but DRV8850 from PCB 1 was mounted on PCB2 and DRV8850 from PCB2 was mounted on PCB1. The issue “transferred” to the other PCB with the driver, so it ensured us, that the issue is connected to DRV8850 itself.

Thanks for your prompt help.

  • Bart,

    Thank you for swapping the ICs between the good board and bad board. That is one item what I want to check.

    Do you have to the waveforms to show the VPROPI difference?

  • Hi Wang,

    tomorrow we should be able to provide the waveforms. If you have any feedback meanwhile, let me know.


  • Wang, 

    please see the requested examples of the VPROPI measurement difference below. Just to remember: the load is the same 3 Ohm resistor, which should give the measurement around 1.83 A (the supply voltage is 5.5V). The value of the resistor connected to VPROPI is 1kOhm (1% tolerance). Each waveform shows the measurement while the H-bridge is acting in one direction and then the other.

  • Bart,

    There is a previous post talking about this issue.

    Basically, it is part by part variation on VPROPI current sensing.

  • Hi Wang,

    I would appreciate that we look specifically into this case, instead of referring to others.

    The statistical data that you referred to was taken at 2A and shows an offset in the mean between OUT1 and OUT2.

    We were also measuring with current value of about 2A (1.8A). However I do not see the difference as an offset between OUT1 and OUT2. Please look at the slide from the documentation of the driver attached.


    What is concerning looking at waveforms, that were sent before, is the difference between the current shown on VPROPI during one direction and the other. For example on the oscillogram below:

    The green line indicates the mean value of measurement on VPROPI pin for the FORWARD direction (operation 1 described/shown on attached page from datasheet, the left side of the Figure 14). The working conditions are exactly as described: the current flows through the H1 FET, through the load and through the L2 FET. The measurement shows around 1100 mV which simply translates to 2.2 A.

    The red line indicates the mean value of measurement on VPROPI pin for the REVERSE direction (operation 1 described/shown on attached page from datasheet, the right side of the Figure 14). Again here: the current flows through H2 FET, through load and through L1 FET. The measurement shows around 780 mV which simply translates to 1.56 A.

    We see quite a difference in our measurements just changing the direction of the current flow. The average value is about 1.88A, more or less, which is about right /close enough to expected value. The current measured with external current meter shows around 1.84 A in both directions.

    To sum this up:

    I expected the measurements in both directions to be quite close to 1.84A. In reality the measurements can be up to 25% higher or lower than expected value. Usually, in one direction the measurement on VPROPI is higher, in the other direction it is lower than expected. I just want to confirm, that this 20-25% deviation from the real current value is normal for this driver?

    For your information: we tested around 150 devices (each one have one DRV8850 chip) and some of them seems to deviate only around 5% from the expected value and some are reaching the 25% deviation from the expected value. I just want to be sure, that we do not have a faulty batch of ICs and these measurements are normal. I am just a little afraid, that if this measurements are out of limits, this drivers’ batch can have some other issues.

    Thanks for looking into this and clarifying.

  • Bart,

    VPROPI is derived from the current through either of the high side FETs. With different directions, the VPROPI senses the different high side FETs. DRV8850 doesn't have the matching circuit to keep them same. Yes. "this 20-25% deviation from the real current value is normal for this driver."

    New integrated FET devices, such as DRV8873,  have a better current matching. But, DRV8873 may not fit your application input voltage range.

  • Thanks Wang, understood. We can accept these accuracy levels in such case.

  • Thank you for your feedback and test result.