BQ76940: Cell Over Voltage protection While Cell Balancing is ON.

Hi,

Today and tomorrow (Nov. 28-29) are U.S. holidays. The team will be back in office on Monday, and will get back to you then.

Thank you for your patience.

Regards,

Max Verboncoeur

Hello Monarch,

Do you have more information about this? Any logs or a schematic to share? How are all of the cell voltages changing?

Best Regards,

Luis Hernandez Salomon

Will share schematic in personal message.

Hello Monarch,

I did not receive the schematic yet.

Can you explain the image? When is balancing occuring?

Best Regards,

Luis Hernandez Salomon

Hi Luis,

I am attaching another image showing all the cell voltages.

Cell balancing was active and controlled by the BQ78350. In the attached image, you can see that the voltage of Cell 5 and Cell 6 jumped from around 4174 mV to 4204 mV.

HHello Monarch, 

Is this during charging?

Or when the batteries are at idle?

Something I notice from the schematic, is that you have 5.6-V Zeners on the VCx inputs (D4/D7/etc). We suggest to have Zeners that are at least twice the cell voltage, to avoid unintentional bias of the Zeners.

For your test I may suggest to remove the Zeners around Cell 5/6 and see if there's any improvement. 

Best Regards,

Luis Hernandez Salomon

Hi Luis,

The batteries are in idle condition.

I have removed the diodes across the VCx inputs, but the result has slightly worsened, with no significant improvements.

Hello Monarch,

We have received your request and will work to answer it as soon as possible. 

Our normal support team is OOO and will be returning 12/10 to help with any ongoing inquiry. 

We appreciate your patience and understanding.

Best, 

-Luis Torres (PME

Hello Monarch,

Can we get logs showing the cells that are being balanced and the cell voltages during balancing and when they are not balancing?

Best Regards,

Luis Hernandez Salomon

Hi Luis,

Shared LOG Via Private Message.

Thanks!

Is it possible to get a log showing which cells are being balanced and when are they being balanced?

Best Regards,

Luis Hernandez Salomon

Shared as Private Message / Additionally, is there a way to obtain data on which cells are being balanced from the host?

Hello Monarch,

My guess here is that there may be a poor cell connection to the battery cells themselves.

A poor cell connection may cause higher resistances in the path that are more easily noticeable when cell balancing is active. The alternative theory would be that there is too high of a capacitance on these pins, but I think that's more unlikely.

Your host should be able to figure out which cells are balancing from reading the registers of which balancing cells are enabled, the BQ76940 does not control any cell balancing on its own.

Best Regards,

Luis Hernandez Salomon

Hi Luis,

I noticed in the shared log that the cell voltages fluctuate between high and low values at exactly 10-second intervals.

I would like to understand how a loose or poor connection could cause this specific time duration. Is there a particular pin of the AFE that needs attention, or is there another factor I should consider?

Could you please suggest the register address that I can use to determine which cells are currently being balanced?

Hello Monarch,

When you balance, you have to enable balance with the cell balancing registers, so to know if you are balancing you can check what is the host writing to in the registers. With the BQ78350, I am not too sure however. That is a gauging device, so it may show registers somewhat different from there.

The fluctuations may be due to when the cells are being balanced.

As to how a poor connection could affect measurements. The part will draw a bit of current when taking measurements, if there's a bad connection it may add a higher resistance from the cell to the VCx pin, which would show up on the measurement reported.

Best Regards,

Luis Hernandez Salomon

Hi Luis,

I think Monarch is pointing that exact 10 seconds interval between high & low value fluctuations can't be caused by loose /poor connection, so connection may not be a reason here. And seeking your attention to identify another probable root cause.

Hi folks, given the holiday season here, many people are out of office, including Luis.  

I'm not super familiar with this case or these parts myself, but regarding the 10sec interval, I see you are using bq78350, that device will usually manage the long-term balancing schedule.  It re-evaluates which cells will be balanced at a regular interval (I think default in the BQ78350-R1A is every 20sec, not sure if the default is different for other versions of BQ78350).  That value is programmable by the customer (up to 255sec), so you might check its value in your system (register = Cell Balance Interval).  I assume the BQ78350 may change which cells are balanced after each interval.  Given that some cells were measured high during balancing, it may then switch to balancing those cells on the next interval.

I don't have access to the schematic, so can't see what series resistances and caps you are using on the cell input pins of the BQ76940.  I believe the BQ76940 disables all balancing briefly, then measures cell voltages, then re-enables balancing again, at a fairly fast periodic rate.  If the RC time constant between the cells and the VC pins is too long, then the transient doesn't have time to settle out when balancing is disabled briefly, before the measurement is taken.

To debug this, I'd suggest as an experiment that you change the Cell Balance Interval to its max setting (255sec), so I expect it will stay in one mode for 4min+.  That then gives you time to walk around various points on the PCB with a scope, see the voltage at the VC pins and across the series R's going to the VC pins, and see what the transient looks like on the cells adjacent to one being balanced. 

It might be that your RC time constant is too large (your scope images may help diagnose this), so you may need to use lower R or C values.  Or as Luis noted, you have a poor connection at a particular point on the PCB, so the effective resistance is larger than the schematic would indicate.

Thanks,

Terry