Battery Vehicle Society

Good point I'll have a look, lots of tweaking to do over the comming weeks. I have the decimal point issue almost sorted now so the numbers displayed make a bit more sense.

For those fiddling with the software on this project using the Picaxe system there is an Editor software update from the supplier today.

http://www.picaxeforum.co.uk/showthread.php?t=10455

I've downloaded and installed it, seems OK.

Been testing the slaves some more, in groups of 4 cells, charging with a lead acid charger and finishing with my dc power supply as the charger doesn't quite get there with the voltage. These cells were already charged but out of balance. The first set I charged one of the cells was very badly out and took about 12 hours to balance. I had been doing some crude balancing with a 5W bulb across the highest cell and left it connected to this cell overnight by accident , hence the out of balance condition.

Today I have just balanced another 4 cells in 1/2 hour using this method with just the on/off mode of charger control. In truth the charger never had to turn off and wait for the cells to bleed down. By the time the lowest cell had turned on the load resistor the highest cell hadn't reached the AbsMaxCellV value. I am pleased with this result and hopefully should see my charger today as well.

The charger is designed for lead acid but provides the CC-CV profile that these cells require at a slightly lower voltage. I will adjust the AbsMaxCellV to 3.65 volts and the MaxCellV to 3.55 volts. This should give me a 98% charge, but if it works it could mean that a standard lead acid charger designed for sealed lead acid cells could be used with these particular litium cells, which I think would be a very neat and useful trick.

Edit: The cells balanced to within 40mv.

GregsGarage wrote: The cells balanced to within 40mv.

That's about the limit of accuracy I expected from the design, I'm happy if they are within 50mv across the pack.

Just noticed a small problem with version 1_05 of the slave software. I am using this with the speed bumped up to 125khz, but the hysteresis added to V1_04 didn't get added to V1_05.

I would suggest that V1_06 (or the latest) have the hysteresis of V1_04 and run at 125kHz. This seems to be the slowest usable speed for analouge slaves. Noticed this working next to the pack while charging my last 4 cells as they were nearly finished.

Greg

Greg, sorry about that

V1.07 (Untested) below now contains more constants so you can adjust the Charger Cut in and out point seperately from the Load cut in and out points, both have hysteresis now as well.
So you now have five voltages to fiddle with.

Charger cut off voltage (Was AbsMaxCellV)
Charger cut in voltage

Load cut off voltage
Load cut in voltage (Was MaxCellV)

MinCellV is as before (unchanged)

www.solarvan.co.uk/bms/SlaveAnalogueOnO ... _v1_07.txt

Also a new Master update which is working quite well, display is neater and we have decimal points SOC calculation seems to be working.

www.solarvan.co.uk/bms/MASTER120908_V09.TXT

Pic of latest testing.

www.solarvan.co.uk/bms/MasterTest01.jpg

Master is updating rolling Soc and it appears to be working with simulated charge & discharge

Thanks for that Peter, will test and let you know.

OK I tweaked the SlaveAnalogueOnOff software a bit more for Greg.

All the settings inc low cell V now have hysteresis.

Greg try this, you now have six V settings to fiddle with.
I have set it for your spec to start with

CPU 125khz

Max Cell Opto On > 3.65V
Max Cell Opto Off < 3.60V

Load On > 3.55V
Load Off < 3.50V

Min Cell Opto On < 2.50V
Min Cell Opto Off > 2.60V

The cpu has a little bit more to do, so may need tweaking to 250khz.

http://www.solarvan.co.uk/bms/SlaveAnal ... _v1_08.txt

Quick load test on 5 of my new 40ah Highpower Lifepo4 cells.
Ambient temp 19C

40A (1C) maintained decent voltage around 3.20V

100A (2.5C) still managed to hold up just over 3.00V

BMS much improved now and returning the right values except under 100A load when it miss read voltages once load got above 75A, strange

I'll scope/debug it tomorrow and see if it is the Slaves or the Master getting the wrong end of the Cell V stick

Slaves didn't lock up or anything Master just read high

www.solarvan.co.uk/bms/LoadTest40A.jpg

Software updates today.

Master nearly ready to take control of charging

www.solarvan.co.uk/bms/SlaveDig140908_v09.txt
www.solarvan.co.uk/bms/Master150908_v18.txt

Edit Cancel the project state funeral, tell the Queen to stop blubbing.

It's a Master display error, caused by the integer only maths and modulo division, this loses the leading zero when voltage goes from 3.10v to 3.09v, the display shows it as 3.9v

I'll have to force it to display leading zero (0)

Cells are now sat on bench being monitored by BMS as they are charged back up after load test yesterday. I'll try and add the charger relay today so it can cut off when done.

Edit BMS Current Sensor now adjusted & working

www.solarvan.co.uk/bms/AllegroCurrentSensor.PDF

Edit Added charger relay to bench test setup and Batteries/Charger now under Master BMS control for first time

Quick youtube video tour and update for a change.

http://www.youtube.com/watch?v=gAaSTQ2cWrk

Peter, just a question about the Delay variable and its comments. In the master, Delay 2500 says that it's 1.25 seconds at 8Mhz. But in both Master and Slave, Delay = 10 says to delay 10 milliseconds. Does Delay 10 really delay only 5 milliseconds at 8Mhz?

Thanks.

Bill Dennis