Joined
·
1,382 Posts
P0B3B Battery Voltage Sensor Internal Circuit Error.
There are now several examples on this forum and the FB groups of Lithium cars suffering from fatal BMS board failures which render the IMA completely inoperative.
Owners are then left struggling trying to obtain expensive spares parts for these rarer Lithium cars and then swapping them out which is fairly tricky and involves removing the IMA pack etc.
There are 4x 10 cell BMS boards in the Lithium IMA pack labelled A,B,C,D each with a unique CAN $ID.
Honda Part numbers. (Note strange number order, more on that later)
A $204 = 1K440-RW0-013 (3812102-500)
B $201 = 1K410-RW0-013 (3812102-470)
C $202 = 1K420-RW0-013 (3812102-480)
D $203 = 1K430-RW0-013 (3812102-490)
In two recent cases the code reader HDS or a decent expensive equivalent has identified a failed D board which when replaced did not correct the issue.
Only replacement of the A board fixed the issue. So why is this?
Possibly......
1) Both boards had failed but only D was identified as faulty by the HDS for some reason.
2) Only the A board had failed but it was miss-identified by the software as the D board.
The fact the HDS and other codes readers are misidentifying the board suggest Honda's own ID OBDII DTC database is in error.
The other code reader manufacturers also use the same database licensed from Honda, so the fact they are wrong as well is understandable.
So if we get an HDS D board indicated failure it seems likely an A board might actually be required.
So this leaves us with a problem of finding spare BMS boards as they have to be replaced like for like (A for an A etc)
Once the available stock of breakers yard A boards is used up people are stuck.
I have checked with Honda UK today and the BMS boards are classed as part of the IMA battery and are not available separately.
New IMA battery = £6500 UK or $8500 LOL.....
Now I did do quite a bit of work during 2019 on the CAN ID's of the 4 BMS boards which range from $201-204
There are two physical types of PCB A & B inside the 4 BMS units (A,B,C,D) which have different battery tap connector pinouts.
So we have two of each PCB A & B type in our four board total.
Looking very carefully at a full set of 4 boards (ID 201, 202, 203,204) the only differences in the components on the pcbs are in the small areas below.
![Green Circuit component Font Electronic component Electronic engineering]()
Each of the 4 boards has a unique arrangement of 103 (10k) resistors in the pictured areas which generates the CAN ID.
Ok it's actually four x 10k possible resistors pulling 2 inputs high or low (0-5V) to give 2 bits and four possible ID configurations.
Only two resistors out of the possible four are populated on any board.
The resistor config is noted by the onboard cpu at first power on and then the board has that ID fixed.
![Product Green Font Circuit component Audio equipment]()
I was able to change the ID of a test board using the above idea and soldering/desoldering these resistors.
In the pic above the truth table for these two points A + B is as follows. High being 5V and Low being GND.
A ID $204 [A = High B = Low] (PCB Type B)
B ID $201 [A = Low B = Low] (PCB Type A)
C ID $202[ A = Low B = High] (PCB Type B)
D ID $203 [A = High B = High] (PCB Type A)
Note the weird ordering and you can see how the HDS software might be logically expecting ID$204 to be D but it is in fact A
Extrapolating this further we might expect the HDS software to believe that....
$ID201 is A but it is in fact B
$ID202 is B but it is in fact C
$ID203 is C but it is in fact D
$ID204 is D but it is in fact A
If this theory holds true the software will miss identify the faulty board in every case!!!!
i.e. If it identifies board C as faulty we know it is board D!!!
So the bottom line in all this is we can change the CAN ID of boards by swapping resistors if we run out of any particular one..
Run out of A boards and we can change a C board (same PCB type) to the CAN ID we need $202 to $204 and vice versa .
* You cannot change an A or a C to a B or a D and vice versa, as they are different PCB A/B connector layouts. *
A can become C and C can become A.
B can become D and D can become B.
We basically potentially doubled our supply of spare boards. (Note I have not actually tried a modified board in the car, but don't see why it shouldn't work.)
I will edit/refine this post and add more info as I dig it out and the member in Germany with the problem currently updates us with the result.
There are now several examples on this forum and the FB groups of Lithium cars suffering from fatal BMS board failures which render the IMA completely inoperative.
Owners are then left struggling trying to obtain expensive spares parts for these rarer Lithium cars and then swapping them out which is fairly tricky and involves removing the IMA pack etc.
There are 4x 10 cell BMS boards in the Lithium IMA pack labelled A,B,C,D each with a unique CAN $ID.
Honda Part numbers. (Note strange number order, more on that later)
A $204 = 1K440-RW0-013 (3812102-500)
B $201 = 1K410-RW0-013 (3812102-470)
C $202 = 1K420-RW0-013 (3812102-480)
D $203 = 1K430-RW0-013 (3812102-490)
In two recent cases the code reader HDS or a decent expensive equivalent has identified a failed D board which when replaced did not correct the issue.
Only replacement of the A board fixed the issue. So why is this?
Possibly......
1) Both boards had failed but only D was identified as faulty by the HDS for some reason.
2) Only the A board had failed but it was miss-identified by the software as the D board.
The fact the HDS and other codes readers are misidentifying the board suggest Honda's own ID OBDII DTC database is in error.
The other code reader manufacturers also use the same database licensed from Honda, so the fact they are wrong as well is understandable.
So if we get an HDS D board indicated failure it seems likely an A board might actually be required.
So this leaves us with a problem of finding spare BMS boards as they have to be replaced like for like (A for an A etc)
Once the available stock of breakers yard A boards is used up people are stuck.
I have checked with Honda UK today and the BMS boards are classed as part of the IMA battery and are not available separately.
New IMA battery = £6500 UK or $8500 LOL.....
Now I did do quite a bit of work during 2019 on the CAN ID's of the 4 BMS boards which range from $201-204
There are two physical types of PCB A & B inside the 4 BMS units (A,B,C,D) which have different battery tap connector pinouts.
So we have two of each PCB A & B type in our four board total.
Looking very carefully at a full set of 4 boards (ID 201, 202, 203,204) the only differences in the components on the pcbs are in the small areas below.
Each of the 4 boards has a unique arrangement of 103 (10k) resistors in the pictured areas which generates the CAN ID.
Ok it's actually four x 10k possible resistors pulling 2 inputs high or low (0-5V) to give 2 bits and four possible ID configurations.
Only two resistors out of the possible four are populated on any board.
The resistor config is noted by the onboard cpu at first power on and then the board has that ID fixed.
I was able to change the ID of a test board using the above idea and soldering/desoldering these resistors.
In the pic above the truth table for these two points A + B is as follows. High being 5V and Low being GND.
A ID $204 [A = High B = Low] (PCB Type B)
B ID $201 [A = Low B = Low] (PCB Type A)
C ID $202[ A = Low B = High] (PCB Type B)
D ID $203 [A = High B = High] (PCB Type A)
Note the weird ordering and you can see how the HDS software might be logically expecting ID$204 to be D but it is in fact A
Extrapolating this further we might expect the HDS software to believe that....
$ID201 is A but it is in fact B
$ID202 is B but it is in fact C
$ID203 is C but it is in fact D
$ID204 is D but it is in fact A
If this theory holds true the software will miss identify the faulty board in every case!!!!
i.e. If it identifies board C as faulty we know it is board D!!!
So the bottom line in all this is we can change the CAN ID of boards by swapping resistors if we run out of any particular one..
Run out of A boards and we can change a C board (same PCB type) to the CAN ID we need $202 to $204 and vice versa .
* You cannot change an A or a C to a B or a D and vice versa, as they are different PCB A/B connector layouts. *
A can become C and C can become A.
B can become D and D can become B.
We basically potentially doubled our supply of spare boards. (Note I have not actually tried a modified board in the car, but don't see why it shouldn't work.)
I will edit/refine this post and add more info as I dig it out and the member in Germany with the problem currently updates us with the result.
