Page 1 of 1

14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Fri Sep 14, 2012 5:15 am
by retepsnikrep
Just a heads up.

I have just finished a schematic for a special slave board to fit onto four OEM A123 28S3P pouch cell packs I bought recently. These fine 60Ah packs are about 100V each fully charged and comprise 20Ah A123 pouch cells in 3 parallel 28 series groups.

http://endless-sphere.com/forums/downlo ... a17fe71ff3

They have a nice JST part number S22B-AIT2G-2FS connector and OEM BMS board on the end.
Unfortunately using the OEM BMS was impossible as nothing is known about it.

However is has two nice connectors on the packs which are split into two 14 cell halves. It seemed a shame and risky to hack that off :shock: so I have sourced the sockets from JST based on the pcb info and designed a 14 cell 12F1822 schematic to suit it.

The schematic is currently being laid out for pcb production to be the exact size and shape as OEM pcb, same mounting points and socket positions etc. It will be a drop in replacement for the OEM BMS board.

The Slave's themselves are similiar to all our previous designs but feature a much reduced component count and are based on the 12F1822 pic which has an internal voltage reference.

The OEM A123 Packs have four NTC thermistors mounted on them so each BMS board has the first slave also looking after the output from two of these sensors. Each pack has two 14 cell slave boards so the first slave on each board looks after the two sensors at it's end of the pack.

The Master will be similiar to earlier designs and use the same comms with three additional commands.

1) Report NTC 1 Reading.
2) Report NTC 2 Reading.
3) Report 12F1822 Internal Temp reading


These temp commands only need to be sent every 10-30 seconds or so, so won't have much impact on the scanning.

All slaves will run the same code, the slaves that arent connected to NTC's will report garbage but that doesnt matter as we know which slaves are monitoring the NTC's.

Some info for you

1) Discussion about the A123 OEM packs.
http://endless-sphere.com/forums/viewto ... a17fe71ff3

2) A123 BMS Connector Pinout
https://docs.google.com/spreadsheet/ccc ... mk2aXBaWnc

3) Connector PDF
https://www.jst.co.uk/productSeries.php?pid=15175

4) Picture of three of my packs.
http://endless-sphere.com/forums/downlo ... a17fe71ff3

5) OEM BMS Board
http://endless-sphere.com/forums/downlo ... a17fe71ff3

6) New 14 Cell Slave Schematic.
http://www.solarvan.co.uk/A123/12F1822_ ... 130912.jpg

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Fri Sep 14, 2012 10:12 am
by GregsGarage
Looks good, are you going surface mount or through hole? What is the overall size of the original bms board and is there one at each end of the pack? Can't wait to see the pcbs.

Greg

P.S. if you are going surface mount I'll send you some 12f1822s

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Fri Sep 14, 2012 10:25 am
by retepsnikrep
Yes there is one at each end of pack.

Size is very close to the 16 cell slaves we did ages ago.

I'm using through hole parts.

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Fri Sep 14, 2012 10:52 am
by GregsGarage
How about putting another 2 slaves on the board, making it 16 cells? You've got spare pins on J1. It would be more work on the layout but then the board could be used for more than just these A123 packs. Build it with 14 slaves for the a123 and 16 cells for other packs.

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Fri Sep 14, 2012 11:19 am
by retepsnikrep
For this project I'm going for 14 slaves it's for a specific purpose. But point taken.

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Thu Sep 27, 2012 2:58 pm
by retepsnikrep
Sent my files to China as well the other day comming back on 9th October.

My board is same size shape and connector position as OEM BMS and fits into OEM bms position on cell packs and plugs straight into battery blocks using same elusive and expensive connector!

First slave also monitors the two NTC thermistors in the end of each pack.

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Sat Nov 10, 2012 11:31 am
by retepsnikrep
Well after a few minor glitches and issues now got 4 boards built and two up and running on my bench test setup.

I tried to use the chip internal pull up for one comms channel :oops: that did not work as it was far too weak :roll: so I had to hack on a 4.7k resistor to the pcb. Not an issue and easy to fit one of those 0.125w ones.

Need to tweak the voltage calculations and i'm also going to use Darell Taylors PBPRO oversampling routine to get 12 bit resolution with the 10 bit adc.

This is pretty much my old circuit but with a greatly reduced component count and using the 12F1822 chips's own internal voltage reference to calculate the cell voltage.

Note this test setup is also using the minimal 12F683 mini master.

The boards fit perfectly onto the A123 blocks and the cells connections just plug straight in :D

The lcd picture shows the number of cells (28) and pack Voltage on the first line.
On the second line we have the number and voltage of highest V and lowest V cell.

Re: 14 Cell 12F1822 Slave Board for A123 28S3P Packs

Posted: Sat Nov 17, 2012 6:48 am
by retepsnikrep
Now finished 8 boards which is enough for my 4 x 28S3P A123 oem blocks.

I will run one pair in series in my Insight test car to give a 60A capacity ~180V Pack
Ditto in a new HCH1 Hybrid Honda Civic I will be buying this weekend with luck.