DRV8818: DRV8818 First step after crossing zero has current oscillation

I think I mistakenly set it to resolved - it is not

Bryan,

We got it back open now.  Can you tell me more about the settings of the device?  Is this on our EVM or your own board?  

The DRV8818 is a fairly old device.  If this is a new design, I would suggest looking at the DRV8424.  

If that is not an option, please help me with your device configuration and we will try to understand what is going on.

Regards,

Ryan

Hi Ryan, It is an older design.  Attached it the schemschematic.docxatic.  Thanks

4555.schematic.docx

Hi Ryan, Any thoughts on what could cause this?  Let me know what you think.  Thanks Bryan

Bryan,

I am going to need some more time to review this.  Please give me until the end of the day.

Regards,

Ryan

Bryan,

Issue only occurs when the current is rising is increasing in quadrants 1 and 3.  For this device, there is only one decay mode that is used and that is SLOW decay...does not matter how the DECAY pin (PFD in your schematic) is configured.  

Out of curiosity, what does the waveform look like if you connected the SRn pin to VCC.  This will force fast decay on current increasing.  

Also, I know this current waveform may not look "ideal", but a lot of times the waveform does not and operation of the motor is unaffected.  Are you noticing more noise in the DRV8818 solution vs. Allegro?  

Regards,

Ryan

Hi Ryan, Thanks for the response.  Motor performance does not seem to be affected by the current oscillation in the first step but there is risk involved with using this in production without fully understanding it.  Waveforms are below with SRn connected to VCC:

single step from zero current to first step:

Bryan,

Current is not oscillating as before, but the tradeoff is higher current ripple.  

I suspect it is related to the blanking time around the zero current crossover.  Some of our parts have "dynamic tblank" as described in the following datasheet.

https://www.ti.com/lit/ds/symlink/drv8834.pdf

Section 8.3.1 describes how this works.  

Unfortunately, this is not a feature on the DRV8818 and blanking time is fixed for the entire range by external RC components as described in the datasheet.  As a test, you could adjust the blanking time lower with the external RC components.  Lower blanking should help here.

Regards,

Ryan

The original design had a 680pF and 56K for the RC, below is with the R changed to 20K:

single step from zero, 200ma/div !  First step should only be about 300ma, this is over 2X

This is obviously worst, will try increasing the RC next

RC now changed to 680pF and a 100K:

single step from zero, 100ma/div

Oscillation peak is lower but now the design has an audible high frequency squeal.

I can't understand why others do not have this same issue, it's a pretty standard circuit. 

Any thoughts ?

Okay Ryan, Thanks for your help!