Old BMS General Thread

[retepsnikrep]

Ethan ,Yes I concur about using it with non terminal cells, good point.

I don't think I have uploaded layout files yet. The layout has not changed since page 11 but I have altered some odd component values (resistors down from 1k to 330R in some places). So I have now uploaded the pcb design files for those who want to tinker with pcb etc.

The zip file contains the latest software/schematics and pcb design files.

The usual disclaimer applies, and an acknowledgment of the source would be nice if you use a modified version for your own ends.

By downloading the files you accept the disclaimer below.

"The BMS modules, software and documentation carry no warranty or guarantee of any kind! They are used at your own risk, and I make no claims as to their suitability for a particular function. Prospective users must evaluate the system before using it, and no liability will be entertained by myself in any shape or form whatsoever. The modules and software have been produced at low cost for the benefit of the EV & electronic community. The software is available free via the internet. Users may modify or adapt the system as they see fit. If you are not fully competent to work on potentially lethal battery systems and high voltages, then do not experiment with or use this system. Be aware that vehicle modifications can lead to invalidated insurance and warranty issues. You the end user remain fully liable for any modifications made to your vehicle. Peter Perkins 2008"

www.solarvan.co.uk/bms/Li-IonBMSProject090908.zip 3.5mb

[dillond666]

Got my first slave assembled
Those via things were holes and although they had electrical continuity I soldered them through anyway, is this correct?

http://www.dillond1.pwp.blueyonder.co.uk/pcbfull.png

[arnolde]

If they were full-featured vias from the PCB manufacturer, then you didnt need to solder them, but it won't hurt either.

I'm now working on a multi-slave (min 14 units per board) PCB using SMD parts (I'm a miniature freak ), but I need to check a lot of datasheets while replacing parts to make sure I get everything right. I was happy to even find an SMD 3W/15Ohm resistor I'll post my results as soon as I have some, maybe by tomorrow.

I should think that at 200 slaves, it might start to get worthwhile to look into automated component placement & soldering. But even if not, these SMD parts are not that hard to solder with a normal small soldering iron and 0.5mm solder.

Ethan

[retepsnikrep]

Ethan. The 08M (Slave) and the 28X1 (Master) Pic's (Picaxes) are available in smd versions FYI, I did think about that as a build option, but I wanted to make it easier for the home constructor, so that laser eye surgery was not also on the parts list. Also were you able to download and use/load my pcb files?

Derek. Nice boards matey. Did you go for a load/balancing resistor? if so which one is it on your board? I used a 3w 15R part from mine to disipate the 250ma balancing/bypass load. It does/could get pretty hot. Cancel that I see you have an external balancing resistor and used an on board mosfet to control it. How much current can that handle?
Which version of my software are you going to use/modify or you going to write your own? Please post your results etc.


Peter

[arnolde]

Peter,

yes, I have the specs for the SMD PICs. There are even libraries for my layout program which include them. I also found a dual SMD optocoupler (8 legs) type ILD217T from Vishay which will do nicely. The resistors and capacitors are no problem, I'll make them all size 1206 for easy soldering. Curiously the LM385 zener diode is only available in an SOIC8 package (8 legs, only 2 used) - maybe for cooling?

I have painters in the house today, I hope I can concentrate enough to finish the SMD layout.

regards,
Ethan

[dillond666]

Peter,

The mosfet I am using is a 30a 60v logic level npn jobbie.
The load resistor connects on the screw terminals. I am undecided about what to use but I think I will need to bypass more current than your application as I am guessing that agms will go out of balance more than lovely lithium.
3ohm would bypass about 5amps, that'll do for starters I think. I might wind up my own resistors
from an old toaster element or something, I just haven't decided
I'll modify the digital slave software you wrote, should require minimal butchering I think. Once I get into it I can get a bit fancy maybe.
Glad mention was made of hysteresis on the load balancing circuit, I was thinking about that because the old way of doing it with op amps needed a bit of tickling with the feedback for this very purpose. It also makes me appreciate how the picaxe opens up electronics for the newbs like me, very useful devices!

Ethan,

Thanks for the info about vias. I had no idea they even existed till I played with eagle pcb.

Derek

[arnolde]

I should think you can do with much smaller balancing resistors. You dont need to completely rebalance the pack in a few hours during daily charging... it's fine if the balance is gradually achieved in a week or so, and for that 250mA should be enough.

Hmmm, BC337 is not available in SMD, I'll use this one (FMMT619TA) instead, should be a suitable replacement? http://docs-europe.electrocomponents.co ... 9c30e1.pdf

[retepsnikrep]

Sorry Ethan, I got your name the wrong way round I have edited my post

Hmm, looks like this thread has now reached critcal mass and will be self sustaining thanks to Greg, Derek & Ethan's recent contributions

[GregsGarage]

It must be my turn then to continue self sustaining this thread. I have finished my 24 slave boards.

Yesterday I connected 4 of them to the 12 volt lead acid charger I have been using temporarily. The slaves where connected to my management module and this in turn shut off the ac power to the charger when any of the cells reached the AbsMaxV of 3.8 volts and turned back on when the resistor had bled off 0.1 volt. It is very slow doing it this way, but also one of the cells is much more out of balance than the others. Started at 9am and by 5pm it was making progress, but still a long way to go. Was planning on leaving it on overnight but then my charger decided that the battery was fully charged so it shut off and said "I'm finished". It finishes at 14.4 volts which is 3.6 volts per cell.

Now that I have finished my boards I have connected my variable dc power supply to the 4 cells and set it so that the highest cell stays at 3.75 volts. I will see how long it takes for the rest to come up, probably be a while because the current is around 250ma, basically all the current is going through the load resistor. This should simulate the proportional control software that Peter has written. I will investigate it once my charger arrives from China (Just checked, it has arrived at Heathrow )

Now just have to finish my battery boxes and fit it all in to the Fiat.

Edit: After 3 hours 2 cells are now blinking the "load" led, one has the load led on constant, and the last one is still at 3.38 volts.

Another edit: After 6 hours still the same, the last cell will eventually get there but at 1/4 amp it will take a while. In normal use cells wouldn't be this far out of balance, but I am board now so have connected my power supply to the low cell. Overall these boards seem to work well. The load resistor was on continuously for 6 hors for 2 days, not a hiccup in sight.

[arnolde]

Gee, I just realized that even if I can get the production price per slave down to 5 EUR each, 200 slaves will still cost me over 1000 EUR

I think maybe I should reconsider using a purely passive approach...