Saturday 5 April 2008

Battery Balancer

I have been thinking about how best to balance my chain of batteries, all 50 of them.

As there are 50 units the price could be quite significant.

Ideally a microprocessor based system would be implemented with isolated sensors on each battery however at this time it is cost and time prohibitive.

There are 2 alternatives I have seen the first would place 15V Zener Diode in series with a light bulb or suitable resistor across the battery. When a battery exceeds 15 volts, the Zener and lamp would bypass current proportional to the over voltage. This would be the simplest to implement.

Another alternative is a more aggressive voltage clamp described here.

The clamp voltage will have to be above the maximum charge voltage from the chargers and the regenerative braking.

Under regen this is 15 volts. Under charge this is about 15 volts but will need to be measured on the charger.

Obviously this solution will only work under charge but should be sufficient for now.
I will test it on my self balancing scooter first.

4 comments:

Anonymous said...

You may be interested in the battery management system described on the KermitTheCar blog. Looks like it could be a cost effective solution.

Anonymous said...

Yes, I'm following the Kermit blog as well, but I'm afraid this is no diy concept, it is being used in another commercial project for e-bikes..

A said...

The BMS I am developing certainly has morphed beyond a DIY project, but we will be marketing it as a cost-effective system for DIYers.

Mal said...

Hi 'kermit'
I have been watching your blog also :) . Well done on your design. It is by far the best I have seen. All the best in commercialising it.

I am curious are you using PWM and a PID control loop for the shunt control on your balancer, or just switching on and off the resistor for a fixed period (several seconds). I am thinking about using PWM and adjusting the dut cycle according to the terminal voltage.
For the communication I will probable use an isolated 2 wire I2C bus. Fibre optic comms would be nice but might be cost prohibitive.

Mal.