BQ40Z50-R2: Auto Ship Shutdown sometimes not working

Hello Sebastian,

The gauge should go to shutdown with those settings, as mentioned in the TRM if there is communication to the gauge then it will restart the timer. I suspect that sometime there was communication on the lines that restarted the timer, or the gauges went to sleep and then a voltage on the PACK pin woke them up again.

Sincerely,

Wyatt Keller

Hello Wyatt,

the customer reported that the batteries were just lying around in their boxing. The only thing they did was pressing the DISP button to check whether they are in shutdown mode yet. I checked some of these batteries and read "No Of Shutdowns = 0".

Best regards,

Sebastian

Update 1: 48h ago I set "Auto Ship Time = 2 days" and left them alone. Now I checked, if they went into shutdown. Only 7/13 were in shutdown mode.

Update 2: I left the batteries alone and checked again one day later, still 7/13 in shutdown mode.

Update 3: I chose the wrong tag in this question. The actual firmware version of these batteries is R1. In the release notes of R2 there's an added feature "Prevent unintended wakeup from shutdown". However from the test above I know that the problem is not that the batteries would wake up irregularly but that they don't even go into shutdown in the first place.

Update 4: For 13 batteries I set "Auto Ship Time = 1h" and left them alone. I checked after 65 minutes, every single one was in shutdown mode. After another hour still every battery was in shutdown mode. Woke them up. Waited 65 minutes. Every battery was in shutdown mode.

So it looks like the problem only arises with longer values of Auto Ship Time. Can anyone confirm or invalidate my theory?

Best regards,

Sebastian

Hello Sebastian,

Can you share your .gg file along with the voltages of the PACK pin and the cell voltages? We can review the settings to see if anything could influence the behavior.

Can you also share the schematic so we can review it as well?

I would also try updating your firmware to the R4 version to make sure it is not the firmware version you are using.

Sincerely,

Wyatt Keller

Hello Wyatt,

see the following attached files, which I exported from four batteries a few days ago:

AC2000155663_nio.gg.csv --> This battery did *not* go into shutdown mode.

AC2000155670_io.gg.csv --> This battery was in shutdown mode. I woke it up to export the data.

AC2000155729_io.gg.csv --> This battery was in shutdown mode. I woke it up to export the data. (I do not know however, why "Number of shutdown events" reads 0 here).

AC2000155731_nio.gg.csv --> This battery did *not* go into shutdown mode.

I don't have access to these four batteries anymore. So for the cell voltages: They were near the values of Lifetimes:Voltage:Cell x Min Voltage. For the PACK pin voltages I don't have the exact values. I remember them being in the range of SumOfCellVoltages +/- 50mV. If this is critical info to you, I can get back to it in a few days, when I have new batteries available for testing.

I will try updating to firmware R4 and then test the behavior. It will take some time because my batteries for testing are not available yet.

AC2000155663_nio.gg.csv

AC2000155670_io.gg.csv

AC2000155729_io.gg.csv

AC2000155731_nio.gg.csv

Thanks,

Sebastian

Hello Sebastien,

It looks like your .gg files show different PF Status data, this should only be saved when a PF occurs but I don't see any PFs present in the PF Status either. Is this leftover from testing or are the packs in a PF failure state?

The voltages are important since the voltage on the PACK pin will wakeup the gauge, we need to confirm there's no voltage high enough on the PACK pin to wakeup the gauge during shutdown.

Sincerely,

Wyatt Keller

Hello Wyatt,

when I checked the batteries they all had PF Status = 0, they were definitely not in a PF failure state. I assume the PF Status data values are leftover from our production process and not cleared correctly.

I'm not sure if I understood your suggestion to acquire the PACK pin voltages.

A) Read out the PACK pin voltage in BQ Studio. --> This will always give ~ SumOfCellVoltages as a result, because the battery will not go into sleep while it is attached to the EV2400.

B) Measure the voltage across Pack+ and Pack- on the battery plug of some batteries in shutdown vs. not in shutdown. (I can't measure directly on the PACK pin otherwise I have to break the casing) --> Is this voltage expected to be ~0 during sleep?

Best regards,

Sebastian

Hello Sebastien,

It must be leftover from testing, it caught my eye because it's deafferent between the two files.

You will not be able to measure with BQStudio since it would not go to shutdown. You would need to measure with external measurement device. If the FETs are on it should be the stack voltage essentially. If the voltage is always above the V Startup voltage to exit shutdown the gauge won't enter shutdown mode. Another possibility is if you have a large capacitance on the PACK pin (pack+/pack- pins) the voltage may be staying high for long enough to prevent the autoship mode.

Sincerely,

Wyatt Keller

Hello Wyatt,

I measured the voltage across Pack+/Pack- of five batteries in sleep mode and in normal mode. The voltages were around 14.0 to 14.1 volts in normal mode and ~ 300mV less in sleep mode. The values make sense because in sleep mode the CHG-FET is OFF so the voltage drops a litte because of the body diode of the CHG-FET.

The TRM reads like this: The battery would first go from NORMAL to SLEEP and then to SHUTDOWN ( "... shut down after staying in SLEEP mode without communication for a preset time interval specified in the Auto Ship Time"). 

Statement 1 (what I understand from the TRM and our schematics):

In our config SLEEPCHG is set to zero, so in sleep mode the CHG-FET will be OFF and the DSG-FET will stay ON. If DSG-FET = ON, then PACK pin voltage will be the stack voltage minus ~ 300mV. (So... well above 10V).

Statement 2 (your statement):

"If the voltage is always above the V Startup voltage to exit shutdown the gauge won't enter shutdown mode"

If both statements are true, then the battery would never enter shutdown mode. 

Now I have three questions:

1. Might one of the two statements be wrong and if so, which one?

2. If both are true, why is it that if I set AutoShipTime=1h every battery I tested enters shutdown reliably?

3. What is V startup even? The TRM mentions V startup but not which value it has or how it is configured. Is "V startup" =  "Charger Present Threshold"? If so, then it is configured to 10V in our batteries.

Best regards,

Sebastian

Hello Sebastien,

I think there was misunderstanding with the voltages, if there is an external source or large capacitance on the system, the gauge may enter shutdown, but then exit, or it will not even enter if the voltage on PACK is above V(startup). The gauge will shutoff the DSG FET when it attempts to enter shutdown, so the voltage should drop at that point.

Anything hardware related is defined in the datasheet, V(startup) is the hardware level voltage the gauge checks for in order to startup. The TRM mostly defines the firmware operation, it has very little descriptions for hardware.

Sincerely,

Wyatt Keller

Hello Wyatt,

ok, found it in the hardware DS:

So the voltage on the pack pin should be below 1.8V to definitely be able to enter shutdown and above 2,45V to leave it.

>> The gauge will shutoff the DSG FET when it attempts to enter shutdown, so the voltage should drop at that point.

Thanks for the good info so far! 

To find out more about the pack voltage behavior I can simulate the battery trying to enter shutdown by manually turning of the DSG FET and measuring the voltage across pack+/-. I can see it dropping to < 10mV in an instant (did not measure with a scope yet). But what's the required speed? I.e. how long will the gauge wait after it turned DSGFET off until it tries to enter shutdown?

Would you agree that to see whats really happening on the pack pin I'd have to crack open the battery casing and measure directly between the pack pin and VSS (device ground)? Or would you consider measuring between pack+/- as sufficient?

Best regards,

Sebastian

Hello Sebastian,

From reviewing the schematic I don't think there would be a large difference between PACK pin and the pack+/pack- port since it's only got 10k ohm in the path and no capacitors on the PACK pin node. I would put a scope on it just to make sure everything looks okay.

Sincerely,

Wyatt Keller

Hello Wyatt,

for reference I had a look at the pack voltage of a battery that did not have the shutdown problem yet. One graph is the moment of shutdown and the other ist when turning of the DSG FET manually. They seem alright to me, but I don't know the required timing constraints anyway.

Pack voltage in the moment of the shutdown takes ca. 0.55s to reach 1.8V

Pack voltage when turning off the DSG FET manually takes 630ms to reach 1.8V

I will make the same measurements when I get my hands on batteries which are known to show the shutdown problem.

I'll be on vacation for the next two weeks, so if possible this thread can stay on hold.

Best regards,

Sebastian

Hello Sebastian,

No problem, when you post on the thread again it will open it on our side.

I don't believe that time period would be an issue to wakeup the gauge, it would be interesting to see packs that are having trouble and if the time delay is much longer.

Sincerely,

Wyatt Keller