DRV2605L: I2C comunication failure

Hi,

Our audio experts are out of office for the holiday and will return the week of 11/28. Please be patient as responses are delayed.

Thanks,

Hi Ferran,

If I understand, this issue happens occasionally after power up where the device will not respond to I2C, however it will sometimes work?

this sounds like a known issue on these devices. If the slew rate of VDD is not fast enough the POR circuitry will not function properly and DRV2605L will not ACK at its I2C address. 

the requirement is the the VDD skew rate is >3.6kV/s. In your case of 3.3V VDD the supply should come up in under 916 ns. 

Regards,

Arthur

Hi Arthur,

The device works fine and from time to time fails. We have more than 50 devices with this IC and this error has been registered on all the devices at some point. We have tried to capture the Vdd waveform during the fail and we have not succeeded because the device did not fail during our tests.

These are the measurements of the slew rate, channel 1 (yellow) VDD signal, channel 2 (blue) VREG signal. We have a slew rate of 12.89 kV/s (256 us).

With these values of capacitors in VDD and VREG is it enough?

Do you have any other ideas we can try?

thanks,

Ferran

Hi Ferran, 

The capacitor values should be enough. I do not suspect that to be the problem. The rise time is also sufficient as you measured.

Will this issue happen when playing back a haptic effect? When the device transitions from driving->braking there will be a transient spike in current that can cause a dip on the Vreg. typically this doesn't cause an issue, but please check Vreg when playing a waveform, and then braking

Also can you share captures of the I2C communication? Showing the shape of the I2C bits.

When the I2C fails are both the I2C SDA/SCL stuck high? is SDA low?

What is the value of the I2C pullup resistor?

Regards,
Arthur

We are using the effects list from the waveform library and we don't have any problems. We tried effect id number 11, 12, 16 and 52.

When we reproduce a waveform, we don’t cause a dip on VREG. In the following screenshot you can see CH3 VDD signal, CH4 VREG signal, CH1-CH2 output pins when playing waveform 16.

We have others devices on the same I2C bus. When the DRV2605L fails, we have communication with the other I2C devices.

We measure when the DRV2605L fails and the SDA and SCL signals stay high.

In this image the signals of the I2C interface are shown: in yellow SCL and in cyan SDA.

We have two pullup resistors of 1.8K on the I2C lines.

Hi Ferran, 

Is it possible for you to test if the device is ACKing at any of the other possible 8 bit addresses?

when the device is in the failed state can you try writing to address 0x01 -> 0xff?

I would like to test further if this is the know POR issue.

Regards,
Arthur

Hi Arthur,

We have tried all 8-bit addresses and the device does not respond in any address.

We also write to register 0x01 the value 0x0FF and the device is still not reachable on the bus.

Regards,

Ferran

Hi Ferran, 

Arthur is out of office and will respond when he is back in office.

Thanks,

Hi Arthur and Daveon,

I am waiting for your answer. Any suggestion to fix the problem we have?

Regards,

Ferran

Hi Ferran, 

Sorry for the delay I have been catching up since the holidays/work travel. 

I will get back to you in two days. 

Regards,

Arthur

Hi Ferran,

Just wanted to check if you are still needing support. We obviously did not respond in a timely manner here. Please let us know.

Brian

Hi Brian,

yes, we still need help. We still haven't been able to solve the issue.

Regards,

Ferran

Hi Ferran,

1) Can you share the layout?  Id like to see how close C19, C20, and C21 are to the pins.

2) There could be a glitch in the VDD line that causes this.  I would put a scope on this pin, set this trigger on a falling edge and try to reproduce this.

-Jeff

1) Here is the layout of the DRVL2605L part:

2) We have been measuring at this point (pin 10) and we have not detected any voltage drop. We have adjusted the trigger to 2.9V, 3.04V and 3.12V, in none of the cases during the test did the oscilloscope trigger. We have not been able to reproduce the issue during this test.

Hi Ferran,

Apologies for the delay, Jeff is out of office but will return this week to continue the thread so please be patient as responses are delayed.

Regards,

So to clarify for #2, you experienced I2C drop out during this VDD monitoring test? 

-Jeff

During the test the I2C did not fail, this issue occurs occasionally.

Regards,

Ferran

Hi Ferran,

I would continue to monitor this VDD until you see the I2C fail.  I believe that should show a glitch that causes the I2C to break.  

-Jeff

Hi Ferran,

Does the issue never happen when you are probing the pin? It could be the case that the small amount of capacitance from the probe is improving the ICs performance. 

Regards,
Arthur