BQ78350-R1: Wrong values from BQ chip

Hi Martin,

Are you using PEC on the SMBus transactions?

Matt

Hi Matt,

yes we are using PEC. We normaly get only a few pec errors, but we don't get any error when we get these wrong values.

In the battery status the EC bits are most of the time set, which indicates an unknown error

Martin

Hi Martin,

Have you seen this issue on multiple devices?

Hi Matt,

we have seen that issue on multiple devices

Hi Martin,

The EC bits indicate something is likely going wrong in the SMBus communications. It can indicate incomplete SMBus command transactions or that SMBus data transfers were not completed. It may be good to capture the SMBus communication on a scope to see if there is anything wrong on the bus.

Matt

This is how it looks like when I create the graph in BQStudio. EC Bits are all zero

And this is how it looks like in my program:

I't reading the Battery status register every 250ms like in the BQStudio

Hi Martin,

I agree the SMBus waveforms look the same except for the 0x7 Unknown error code returned. I checked with the firmware engineer on what might be causing the error code. He said BQStudio has inherent delay of 20-30ms between each command to prevent overloading the bus. So I wonder if the bus is getting overloaded?

Matt

Hi Matt,

in the above scenario I send the above message only every 250ms so it should not be a problem. I tried to overload the bus already once. I get a lot of timeout (10ms) because the BQ78350 does not answer, but it was not reproducable that I get wrong values.

Martin

I was able to reproduce the wrong voltage issue (Pack voltage is around 19V):

Correct reading:

The voltage reading is the first read. The second is the current, the third is the relative state of charge and the fourth is the battery status. It happens during the first steps of the program.

I checked the PEC and I've seen that the PEC does not match with the calculated one:

http://www.ghsi.de/pages/subpages/Online%20CRC%20Calculation/index.php?Polynom=100000111&Message=C64D

Neigther for the correct voltage message nor for the wrong one.

0x17 must be prepended the data to calculate the PEC: 0x17C64D

I found this thread: https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/537235/bq78350-reading-wrong-values

It might be something similar?

Hi Martin,

Is the device already awake from Shutdown mode before these first steps of the program? Do the bad reading only happen during these first steps or do they continue after this?

The old thread did not have a conclusion, so I am not sure if the issue is the same. 

I checked the PEC calculations for the waveforms you sent yesterday and they look correct. It is difficult to see the new waveforms because the image is smaller. Here is the calculator I use: https://crccalc.com/ 

Best regards,

Matt

Hi Matt,

the screenshots are made during startup of my program. There is a reset of the BQ78350 command inside. Otherwise I was not able to reproduce it. But as it can be seen in the above logs, it occurs from time to time

If you open the image above and then double click on it, it zooms in with high quality and so you can see every byte. I rechecked the PEC and it seems to be correct for both messages, the wrong voltage and the correct voltage reading

Hi Martin,

That's good to hear. Let me know if you see any more error readings after this change.

Matt