Adding cell balancing to BQ25703A

[ali_asadzadeh]

Hi,
BQ25703A is a relatively new part from TI, it has almost everything I needed, but it lacks cell balancing capability, since I would have a Cortex M in my design, I tough maybe I could use spare ADC and IO for performing Cell balancing on the batteries too with the MCU, Do you suggest a better solution? or do you recommend any better Idea?

The system is 4S battery, measuring each cell voltage with a simple resistor divider and connecting it to an ADC channel of MCU and by that measuring each cell voltage and turning on/off each cell MOSFET along it's series resistor, the only problem remains is Driving MOSFET's cheaply!

[ali_asadzadeh]

Can we use DW01 for cell balancing?

[Daixiwen]

The system is 4S battery, measuring each cell voltage with a simple resistor divider and connecting it to an ADC channel of MCU

Just be careful that you don't introduce unbalancing with that solution. If the "simple resistor divider" draws different currents from each mid point between the cells, it will discharge the cells at a slightly different rate. The current is usually negligible, but you can notice it if you keep the battery in storage for a long time.
It should work. Remember to turn the MOSFets off when you do the voltage measurement, and do preferably the balancing during the end of charge. This is where the voltage difference between the cells, if they are not balanced, will be the highest.

[NiHaoMike]

4S balancer boards are cheap. Just get one.

[ali_asadzadeh]

Thanks, do you have a link or schematic for them? It's not a one of job, so buying the modules is not an option.

[paulca]

There are very few options to do this correctly.  Balance charging LiPos is tricky.   It becomes even more tricky if you want to pull a load while charging, aka BMS.

Probably the simplest solution is to charge the pack as is, eg. 16.8V, put a resistive load switched with MosFets on the balance leads to each cell, but you need to measure individual cell voltages very accurately and switch the mosfets to discharge the high cell while at the same time lowering the charge current to the pack to something that allows the resistive load to keep up.  This solution is what is used in LiPo chargers, but they are typically short cycle devices, you connect them, charge the pack for under an hour and disconnect them.  This is important as resistive load balancing will discharge the pack if left connected for a prolonged period.

If you don't need high current output, you can charge individual cells via the balance leads only, but the voltages required are not easy to produce accurately.  If for example you want to put 4.2V to Cell 3, then you need to add that 4.2V to the voltage produced at Cell 2's + end.  So you need to switch in, say, 8.2V to Cell 3 - and 12.4V to Cell 3 +.

Very often people making custom projects just give up and switch to putting the cells into parallel config giving 3.3-4.2V and use a boost converter to get the higher voltage required.  Which is probably the better option for a homebrew BMS.