BQ27520-G4: I2C transfer NAKs

Hello Jonathan,

If you use the TI toolchain (EV2400 and BQStudio) do you see this error or is it only with your host system? It is helpful to narrow down root cause if it is the gauge or host side/timing issue.

Sincerely,

Wyatt Keller

Hello,

Unfortunately I do not have an Eval Kit with which to test this. Is there something I can do with my setup that would be helpful?

Thank you,

Jonathan

Hello Jonathan,

Generally these types of issues are from the host side violating an I2C timing spec. If it does not occur with our toolchain (doesn't have to be our EVM, it can by your board and EV2400) EV2400 is required to setup this gauge in order to upload the chem ID and pull the golden image.

Sincerely,

Wyatt Keller

I see. I'll work to order one of the EV2400 (they're out of stock at Digikey, but in stock at Mouser).

In the meantime, is there any utility to looking at the protocol analyzer traces? I'd imagine that would show any timing issues that may exist? I've attached the Saleae trace if that's of use.

Thank you,

Jonathan

Hello Jonathan,

It looks like the Saleae file didn't attach, if you can share this file I will take a look to see if there are violations.

Sincerely,

Wyatt Keller

Hello,

When I try to attach the file, I get a 500 (Internal Server Error):

So here is a Google Drive link:

https://drive.google.com/file/d/1cpCVgkz_h3WTwFnV6EDqWWfYk3xeX1vZ/view?usp=sharing

Thank you,

Jonathan

Hello Jonathan,

I looked through the log but it is difficult to tell if there are any timing violations since it is only the digital data and not analog as well. It's hard to tell what the slew rate is and other level thresholds are passed.

Sincerely,

Wyatt Keller

Hello,

Thank you for the feedback. I think you're correct, the issue is likely with the bus:

Considering we're at 100kHz, these clocks should be far more crisp. I have 1k pull-ups, since there are 5 devices on the bus, and the voltage level is 1.8V. I'll see about increasing these to 470ohm, and perhaps check the capacitance.

Thank you,

Jonathan

Hello Jonathan,

The waveform time constant could also be from the Saleae sampling rate since this is selectable. I would make sure you're at a high sampling rate so you can see the true ramp times. It is likely that the bus capacitance/pull-up resistance combination needs to be adjusted.

Sincerely,

Wyatt Keller

Hello,

Apologies for the late reply - I was traveling last week. You were indeed correct, I should have noticed this! With the sampling rate turned up, the waveforms look like what I expected. Here is a screenshot of a NAK:

Here is the screenshot of the same command, succeeding:

Thank you, Jonathan

Hello Jonathan,

Are you able to find the timing violation? Can you share the raw Saleae files?

Sincerely,

Wyatt Keller

Hello,

Here is the raw Saleae file: https://1drv.ms/u/s!Al_uynZRNxW6gssqd5LsJC7QZOkNMg?e=VgJkPT. Let me know if this works. In this capture, I tried to read the Firmware Version twice - the first one succeeded and the second one failed.

Thank you,

Jonathan

Hello,

I'm curious if you saw anything here that looks out of spec? To my eyes, it looked OK (rise time / fall time), although I very well could have missed something. The fact that it's so predictable (working every other time) is a little wacky to me, if it's a timing thing like that.

Thank you,

Jonathan

Hello Jonathan,

Most of the time if the EV2400 is working and host is not it is timing/structure violations. From all of my previous experiences it has been due to the host side.

I tried downloading the file but I think it is blocked by the firewall, can you attach directly on the forum?

Sincerely,

Wyatt Keller

Hello,

Unfortunately when I try to attach it here, it seems to fail every time. The file is rather large (100+MB, compressed). I've uploaded it to Google Drive as well, in case that works: https://drive.google.com/file/d/1HW4Bszf2Cyx9oxjZom_gWl-B9FvD2Gd-/view?usp=drive_link. 

I don't currently have an EV2400, I'd hoped not to need it but if it's necessary we'll go ahead and order it.

Thank you,

Jonathan

Hello Jonathan,

I still cannot access the link, it is blocked. Is it possible to share a shortened version of the file directly?

EV2400 is required in order to configure a flash gauge for production in order to upload the chem ID and pull the flashstream/srec for production programming.

Sincerely,

Wyatt Keller

Hello Jonathan,

I don't see any obvious timing issues in the Saleae capture you shared. Are you able to test with EV2400 as well? I will do a quick check on an EVM tomorrow.

Sincerely,

Wyatt Keller