CMU board notes, eprom, cell numbering

[kiev]

Starting a thread to collect this information.

CMU board with labels,

[kiev]

Index placeholder; Hope to post up some schematics also.

[kiev]

reading the eprom using an arduino board:







ref: https://www.speakev.com/attachments/dat ... ng.116112/

[edit: ignore this eeprom data it is incorrect]

The (lower 256-words) memory contents from Rupert's posting on the speakev forum:



Background: Rupert read out the contents of the lower 256-words of the eprom from 2 different boards. He thought his issue was in CMU08 based upon really high voltages shown in Canion cell data, but he was using an incorrect pack layout diagram of the CMU numbering. So he pulled CMU02 instead of 08, and ordered a replacement board 08 from a dismantler (which was also an 02).

Notes: the differences are highlighted; the replacement shows a value "2" in several rows, which could be the CMU ID number, but the original is blank in those locations.

The eprom holds 512 words of 16-bit data. So the memory addresses would range from 0x000 to 0x1FF. The rows that contain the majority of the data seem to be organized to contain a cell address number in bits 7-9 of the address:
i am adding brackets around the [cell address] bits
Hex binary address
0x010 000[0 00]01 0000 MUT or CAN buss cell A, at the most positive terminal of a module
0x050 000[0 01]01 0000 B
0x090 000[0 10]01 0000 C
0x0D0 000[0 11]01 0000 D, at the most negative terminal of a 4-cell module (CMU06 or 12)

if i extend this bit scheme, then a read of the upper 256-words should have cell data in these rows:
0x110 [1 00]01 0000 MUT or CAN buss cell E
0x150 [1 01]01 0000 F
0x190 [1 10]01 0000 G
0x1D0 [1 11]01 0000 H, at the most negative terminal of an 8-cell modules.

[kiev]

updated the photo with additional components.

Found some talk on the www that trying to read the eeprom in-situ can introduce spurious data.

On our board, When the power is applied as was done by Rupert, it powers up everything on the LV side of the board, the LTC chip and the microcontroller, etc, The micro is expecting to see an enable signal coming across the optocoupler at the isolation gap, so it's not clear what it might be doing to the four signal lines of the eeprom chip without the "enable" signal.

Made some good progress on circuit traces, hope to draw it up and post up schematic.

[kiev]

Here is a sketch of the power supply section. The EV-ECU controls a relay that sends the +12V to the CMUs.

The HV +5V is used to power up the CAN transceiver and the HV digital isolator; the LV +5V powers the microcontroller, the LTC, and the eeprom.

[coulomb]

Thanks a heap, Kenny, please keep 'em coming.

I find it strange that the lower voltage regulator has a voltage divider going to pin 1, and the upper one doesn't. Worth a double check?

[kiev]

More circuit drawings.

[kiev]

Thanks to Mário Fernandes in Portugal, we now have a CMU wiring color chart for the daisy chain harness in the pack.

Looking at the color of pin 3 wiring, which is the Input side for the Autonumber, i think CMU07 is the first module in the daisy chain, and pin 7 of CMU06 is the last Output back to the BMU. This is my assumption due to the different wire colors used. Hopefully Mike @Coulomb will find the section in code that sorts out this weird wiring and numbering sequence.

[kiev]

...
I find it strange that the lower voltage regulator has a voltage divider going to pin 1, and the upper one doesn't. Worth a double check?

i double checked and the threshold adjustment feature for pin 1 is not being used for the HV 5V regulator, it's tied to ground which is done to use the on-board threshold level. No delay cap either, so it's tied to ground.

Thanks for taking on the firmware--let me know if there is something else that you need traced out.

i'm thinking to put together an arduino eeprom reader to see what my 4-cell board contains, then i have a single 8-cell module that i can check.

[kiev]

[edit: ignore these values, found out the library code was defective]

i finally got back to writing some arduino code to read the eeprom data from the chip on one of my CMU boards. It was a 4-cell card from Martin's cell replacement adventure and video production. The chip is an S93C76A eeprom, 8k-bits organized as 512 words with 16-bits per word. Here is a picture of the lower 256 words, the upper 256 were exactly identical just add 0x100 to the address in the left column.