DRV8876: IC protection issue

Hey Cherry,

The device should not be damaged by the procedure you describe, especially with just a 2.7A power supply.  Could you post details about that adapter, is it a TI adapter?  

Can you post a schematic showing what size bulk capacitance they used, and if they followed the Typical Application Schematic?  

Regards,

Jacob Thompson

Cherry,

I suspect the voltage on the supply is exceeding the abs. max when blocking and releasing.  Jacob had a good idea to check bulk capacitance to ensure supply is more stable.  Can you scope VM supply voltage when this damage occurs?

Regards,

Ryan

Hi Jacob and Ryan,

Thanks for your support.

• Adapter 65W, 2.0A is rated, but the manufacturer guarantees a maximum overload of 6A 20mS when used.
• large cap configuration is 100uF, 35V.
• VM voltage is input voltage 24V.

Below are some circuit shots of the power input and the DRV8876 peripheral design:

Regarding the waveform below, CH1 is the current  and CH2 is 24V.

Is it possible to directly cause IC damage when transient current is measured up to 5A? Or because of the large transient current and the large IC RDS(on), repeated testing will cause thermal damage?

Thanks and regards,

Cherry

Hey Cherry,

Let me confer with Ryan on the waveforms, should get back to you within 24 hours. 

In the meantime, I recommend looking at the DRV8874, as I believe it is pin to pin compatible with the DRV8876 but has a peak current limit of 6A instead of 3.5A and a much lower RDSon.  

Regards,

Jacob Thompson

Hey Cherry, 

Please remove R21, the 20-ohm resistor between VCP and VM, it is unnecessary and will negatively impact the charge pump performance.  

What appears to be happening is the overcurrent protection (OCP) is tripping, it is deglitching for 3ms, then trying to move, OCP, deglitch, etc.  The chip should be able to exist in this state indefinitely and protect itself from damage.  If you can post a scope shot that includes the nFAULT pin at the same time it should verify this.  

We suspect this should be some kind of an overvoltage event, and recommend trying a different power supply that is better at sinking reverse current.  Alternatively, try adding more bulk capacitance, maybe 200uF.  

Also, can you confirm if the chip in question is the DRV8874 or DRV8876?  Thanks

Regards,

Jacob Thompson

Hi Jacob,

Jacob Thompson said:

Also, can you confirm if the chip in question is the DRV8874 or DRV8876? 

It's DRV8876.

The nfault trigger signal is as follows:

CH1-current signal/CH2- 24V supply/CH3- nFault signal

The customer currently suspects that the IC OCP status is in Auto RETRY and therefore tries to continue running, but the circuit itself has an IMODE plug-in resistance of 62K, and an OCP without a power-up cannot be re-established.

Could you help double-check if the above waveforms is able to determine whether this's the case.

Thanks and regards,

Cherry

Hey Cherry, thanks for the update!

Please check section 7.3.3.2.2 Cycle-By-Cycle Current Chopping in the datasheet.  We suspect this is what is happening if the scope shot is zoomed in.  The device is attempting to chop the current during this blocked startup case.  With a 62K IMODE resistor that is Cycle-By-Cycle current current chopping mode, with nFAULT indicating current chopping and overcurrent events and a latched OCP response.  

Then we suspect the device is going into thermal runaway and the nFAULT is actually indicating a thermal trip and recovery, not an OCP.  

Try replacing the DRV8876 with a DRV8874 since it has a higher current limit.  Suspect that will solve the issue either way.  Could also try adding a small heatsink to the driver and see if that changes anything.  

Regards, 

Jacob