Stack-able Low Cost Balancing BMS

parabellum said:
okashira said:
The neptune draws power from the top cell. Total fail
I interpreted "top cell" actually as powering from pack terminals, as ground is common usually and 3V will not be best voltage to work with. :) But we can ask again, more specifically.

Glad to hear it meant "full pack,"!
Might have to get a few of these to play with.
 
The 02micro oz890 based units (bmsb smart) keep on coming so must be available?
The 02mirco site main page offers 4 main catagories of business, BMU is one of them, where they seem to offer a number of solutions http://www.o2micro.com/products/battery-mgmnt.html

The oz890 is listed along side the oz8906 stackable
 
Ken at O2micro got back to me. He thought I may mean the OZ8902 which is disscontinued with no replacement planned. I have sent him a highlighted copy of their product list and an idea of what we would like. I have also asked after the oz890 though at $20 I imagine bmsb have it covered, you just need to stack them as we have seen done, and tidy up the boards and programming. New fets. Physically stack them and put the leds back they omit now.

The real manufacturing would be on the programming side of things. It's already well covered by other people working on them.

I would put up some funding to ease this along. $100 or something.

Edit: Perhaps I could buy myself the programming side of things, and you could use it for 6 months. It's $200 leaving you buying $40-$60 worth of boards you could always sell on.
 
Well they didn't find an order of 100 chips worthwhile. I Have gone back asking what sort of minimum order size they would like to see. Perhaps 1000USD ?

No reply as yet, but has this thread gone cold? I'm happy to cover this with the right people and targets in place. I know the customers are out there, but I can't do this project alone.
 
Just a WAG here, but you probably need to order more like $50-100k worth as a minimum to get their attention, if not more. Even then its only because at that level they see you as being serious enough that you'll want to recoup the cost and order more.
 
What the DIY community needs first is:
-A balancer/basic monitor.
That's it.
Why?

A. HVC can be handled by using a decent charger (duh) and (perhaps) warning light / buzzer on the balancer board.
B. LVC cut can be handled by a decent controller and a (perhaps) warning light / buzzer on the balancer board.
C. Short circuit, over-current, etc can be handled by a fuse or breaker and good pack construction.
D. Protection against low temperature charging can be handled by the user. ie DON'T CHARGE WHEN IT' TOO COLD.
--D.1 Or a temperature controlled switch
--D.2 Or, again, (perhaps) warning light / buzzer to reflect a cold pack condition (with a note on the light that says "don't charge or disable regen")
E. Protection against overheating can be handled by pack sizing and a temperature gauge


The only thing that a DiYer really needs to have done by a computer is making sure the pack stays balanced. Everything else falls into place.

The balancer MUST have ALL of the following:
1. Be smart enough to balance at the right times, ie, at 4.00V or higher, always balance may be good enough with quality batteries.
2. Accurate enough to achieve good balance. Cells should be within 0.004V or less if checked anytime unloaded above 80% SoC.
3. Low quiescent current draw so it can remain connected at all times even if the pack is stored at 5% SoC for a year. 100μA or less.
4. Programmable, so that balance voltage, HVC warning, LVC warning can be set by the pack builder or the user
5. Low balance current - 50mA is fine - so that the board can be fully potted and sealed within the pack without worry of overheating.
6. Powered by full pack - balancer board circuity (i.e. micro-controller, FET drivers, etc.) shall be powered by the full pack voltage and stepped down with high efficiency buck converters, AND be low power with a very low power sleep mode that automatically wakes up when appropriate (> 4V/cell, active discharge, etc) - 200μA or less sleep current
7. LOW COST - target $4 a cell or less is fine! $40 for a 10s, $80 for a 20s.


From there, such a product can be further developed to add more features. CANBUS, contactor controller, current sensing, SoC calculations, cool displays, fancy marketing.
BUT DON'T ADD ANY OF THOSE FEATURES UNTIL YOU HAVE A SELLING PRODUCT WITH **ALL** OF THE ABOVE FEATURES AND REQUIREMENTS FIRST.
FURTHERMORE, NONE OF THOSE FEATURES SHALL BE ADDED IF THEY DO ANYTHING THAT COMPROMISE THE BALANCER REQUIREMENTS 1. - 7.

That's my opinion. :D
 
Hvc can't be handled by the charger, as it can't see the individual cells. An aging pack will likely go out of balance beyond a 50ma balance currents control if pushed to it's final moments. Letting a cell past it's respective hvc requirement
Likewise lvc, a controller can't see individual cells. To try and make it work people avoid using the last third of their pack. A wasteful exercise for both capacity money and lightness.
As the bms must control the hvc lvc levels, a fuse won't do as it won't be quick enough to protect the fets in many circumstances.


You have it bang on, a simple unit people won't get themselves in trouble with is needed. The first time somebody gets it wrong, is the last time you will sell a board unfortunately.

I still like the smarts. $20 and does everything upto 13s and 60amps with additional cooling. They do need work though, the design was rushed.
 
The BMS would still do cell-level LVC and HVC, just not pack-level.

I believe the balance current would be when charging, not when discharging. A pack that goes significantly out of balance on each cycle is really a job for the trash rather than a BMS IMO.
 
Keen. Whats the best way to get involved if you have zero electrical engineering skills? :lol:

Either way, I'll buy a few once they are tested and working.
 
Back
Top