[okashira]

Hi guys

I have a 2012 AZD transit connect with only 82 miles on the clock.
When I got it, it did not run or charge, and I found two cells at 0v.
I replaced one of the cells, and the other cell charged up and holds a charge fine.

Btw, I made the replacement cell out of 21 18650 cells in parallel from a 2016 Model S ;-)
pics: https://1drv.ms/f/s!AuqwBZb5n5Jqgu1KRHYx_pEuam5Vuw


Anyway, after fixing everything else and putting the pack back together I got the van so it can drive, ac works, etc.
But pack 2 is throwing errors and wont come online. So the van runs in limp mode and won't charge.

Here are the codes:
-----ESS Pack 2 Info Codes-----
P0AFC-49 Hybrid Battery Pack Sensor Module
P0B3F-F1 Hybrid Battery Voltage Sense A Circuit Intermittent / Erratic


-----VCU DTC-----
P1A28 Battery Control Module "B" Critical Fault (same as P1A20)
P1A8B HV Discharger Status Circuit High


Anyone have any parts or insight?

 

[okashira]

After another 4 hours of experimenting and diagnosis Ive narrowed it down to one bad BMS module board.

It looks grungy because ive been trying to clean off the conformal coating.

It has two AD7280 BMS chips
and a 32 bit NEC micro controller

The van has a total of 16 of these boards, only one is bad and giving the voltage sense code.

Does anyone know anyone who might have some of these boards?

 

[kennybobby]

Looks like twelve cells per module board, all the cells are good? Looks like 3 parallel bleed resistors per cell for balancing--can you read the resistor value?

First thing i would do is use a multimeter to check for a short thru one of those little yellow ceramic chip capacitors--they are notoriously prone to chip and crack, and create a shorted path. Quick and easy to check and not hard to fix if one is bad.

The board looks simple enough to trace out and draw a schematic, but getting the code out of the NEC chip would be tough. It looks like a resistor divider network to measure voltage and FETs to drive the bleed balancers, with a high- and low-voltage isolation barrier and serial communication across the barrier (opto couplers).

All this is very similar to that used on the Tesla boards with one important exception related to redundancy and fault-tolerance--Tesla has it, but i don't see it on the Azure board.

p.s. Thanks for posting the picture, can you identify the other ICs on the board?

 

[okashira]

Looks like twelve cells per module board, all the cells are good? Looks like 3 parallel bleed resistors per cell for balancing--can you read the resistor value?

First thing i would do is use a multimeter to check for a short thru one of those little yellow ceramic chip capacitors--they are notoriously prone to chip and crack, and create a shorted path. Quick and easy to check and not hard to fix if one is bad.

The board looks simple enough to trace out and draw a schematic, but getting the code out of the NEC chip would be tough. It looks like a resistor divider network to measure voltage and FETs to drive the bleed balancers, with a high- and low-voltage isolation barrier and serial communication across the barrier (opto couplers).

All this is very similar to that used on the Tesla boards with one important exception related to redundancy and fault-tolerance--Tesla has it, but i don't see it on the Azure board.

p.s. Thanks for posting the picture, can you identify the other ICs on the board?

Good idea on checking caps... will do.
Though I think if one is shorted the BMS would report 0v for a cell :-( which it does not.

I'll just plug it in and measure the voltage at the 12 cell inputs to the BMS chips

 

[okashira]

Looks like twelve cells per module board, all the cells are good? Looks like 3 parallel bleed resistors per cell for balancing--can you read the resistor value?

First thing i would do is use a multimeter to check for a short thru one of those little yellow ceramic chip capacitors--they are notoriously prone to chip and crack, and create a shorted path. Quick and easy to check and not hard to fix if one is bad.

The board looks simple enough to trace out and draw a schematic, but getting the code out of the NEC chip would be tough. It looks like a resistor divider network to measure voltage and FETs to drive the bleed balancers, with a high- and low-voltage isolation barrier and serial communication across the barrier (opto couplers).

All this is very similar to that used on the Tesla boards with one important exception related to redundancy and fault-tolerance--Tesla has it, but i don't see it on the Azure board.

p.s. Thanks for posting the picture, can you identify the other ICs on the board?

I measured the resistance of the bleed resistors got 50ohm.

I just plugged the bad board into the module and measured the voltages at the test points near the analog devices BMS chips, and they all read correct ~3.59V. So no bad caps. The voltage is making it to the BMS chips.
So for some reason the microcontroller is getting a bad value.

I might try to replace the BMS chips?

here are higher res pics of the board (and it's sister board from the other pack)
https://1drv.ms/f/s!AuqwBZb5n5Jqgu1hJtIip--wssZzuw

Here are the test points where I checked good voltage. These go to the voltage sense pins on the analog devices BMS chips

 

[okashira]

Just ordered two of these from Mouser.
I guess not too much harm in trying to desolder and replace those two chips.

https://www.mouser.com/Search/Produ...TZvirtualkey58430000virtualkey584-AD8280WASTZ

 

[okashira]

No dice. I replaced the two AD7280 chips with AD7280A, and it throws the same code. :-(
I need to find someone who has a transit connect battery pack they would be willing to scrap or even a single PCB.

 

[kennybobby]

Are the cells labelled by letters, e.g. "cell A, cell B,..." or are they referred to by numbers, "cell 1, cell 2, cell3, etc..? Does the ESS report each cell voltage such that you could identify which cell or circuit has the defect (temperature sensor and voltage sensor)--maybe it is a reference voltage issue since it affects 2 different type sensors?

Worse case you could measure every resistor and see if they compare with the value. e.g. the balancer resistors are "151" or 150 Ohms, but there are 3 in parallel so the total across is only 50 Ohms. A bad resistor in a voltage divider circuit would give incorrect data and throw the DTC. It's tough to troubleshoot without a schematic, that's why i try to trace everything i can and draw it up--someday it will need to be repaired...

 

[okashira]

The diagnostic software does not show all voltages. only min and max.
I posted pics of everything I can see.

The temperature sensor error is because i swapped the AD7280 for the AD7280A, and Analog Devices changed how the thermistor voltage is referenced. It is something I can fix.

I am going to see if an "ALERT" is being triggered by either ad7280 and sent to the uC.

 

[gjsolar]

Hi, Just wondered how this progressed ? I've just started with an isolation fault on pack 1 of my Transit Connect Battery and so plan to investigate further.