Top balancing with custom lower cost power supply

[Bibb]

Hi

My idea is to not buying a benchtop power supply I'll never use again after top balancing process.
Instead, I'd like to use:

1. old computer power supply that gives me 12V (AC to DC)
2. Boost converter (step-up to 14.6V) for pre-charging step (max output: 20A)
3. Buck converter (step-down to 3.65V) for top balancing step (max output: 20A)

Total cost: $25 and reusable converters!

Good or bad idea? Any suggestions?

 

[Supervstech]

Hi

My idea is to not buying a benchtop power supply I'll never use again after top balancing process.
Instead, I'd like to use:

1. old computer power supply that gives me 12V (AC to DC)
2. Boost converter for pre-charging step (max output: 20A)
3. Buck converter for top balancing step (max output: 9A)

Total cost: $20 and reusable converters!

Good or bad idea? Any suggestions?

I don’t get anything on the buck converter link, but the plan seems solid.
I use the 1200W buck for charging my 48v pack off my 14V small bank and it works fine, it is VERY poorly designed and melts the solder joints when over 8a draw for more than a few minutes… but it does lock the voltage well.
I use a Drock programmable buck converter for top balancing, but it only does about 5 a…

 

[Bibb]

melts the solder joints when over 8a draw for more than a few minutes…

Thank you, I have to pay attention to wires. But connectors on converters are not so big

 

[RCinFLA]

I use the buck converters that claim 20 amps but don't adjust current limit on them above 15 amps output or they get too hot. For a 12v computer supply as input it takes 5 amps from 12v supply to get 3.65v @ 15 amps output. You can feed multiple of these buck switchers from a single high current 12v computer supply. Make sure you set the 3.65v voltage limit first with no load on buck switcher.

Just bought four more of them a month ago for $28 delivered. Geekcreit® DC 6-40V To 1.2-36V 300W 20A. They are LM25116 based synchronous buck switcher with external NCE8290 MOSFET's on heatsinks. Very good conversion efficiency.

You can run four of them on a 30v 10 amp power supply for four 3.65v @ 15A sources. Just break your cells into four groups. Don't parallel buck switcher outputs together.

 

[Bibb]

You can run four of them on a 30v 10 amp power supply for four 3.65v @ 15A sources. Just break your cells into four groups. Don't parallel buck switcher outputs together.

My battery is a 4s.
So I could buy 4 bucks (1 by cell) and pre-charge+top balance each cell without putting them in series and parallel?
I read in FilterGuy's tutorial this method is possible:

Some people like to top balance by charging each cell individually to the target balance voltage. The arguments for doing this is that it can be done without disassembling the battery (Less chance of striping terminal threads). They also feel that doing it this way is safer (less likely to mess up and over charge, but I do not understand that argument.
I personally prefer to parallel top balance because I know for sure all the cells get charged to the exact same voltage.

Or maybe just pre-charging each cell with 4 bucks, then parallel them for top-balancing with 1 buck?
If that's what you meant, the boost converter is useless... More and more cost savings!

 

[mikefitz]

My battery is a 4s.

My suggestion is to build the battery with the cells in series with a BMS connected ( a BMS that reports cell voltages) and charge with your usual charging process, ( solar or AC charger), set to 13.8 or 14 volts. If one cell runs ahead of the others, bleed off charge on that cell using a car headlight bulb for a few seconds. Repeat the process on the 'high' cell varying the 'bleed' connection time as required. With reasonable quality cells its possible to balance within a few mV.

This process can be carried out with the cells in the final containment arrangement, ( box, taped together , compression rods), used to prevent cell swelling.

If you still feel its necessary to charge to 3.65 volts with the cells in parallel, then with the cells still in the containment situation, make up temporary cables to replace the buss bars and wire the cells in parallel. Attach the supply, since the cells are almost fully charged at this stage it will not take long.

Mike

 

[Bibb]

My suggestion is to build the battery with the cells in series with a BMS connected ( a BMS that reports cell voltages) and charge with your usual charging process, ( solar or AC charger), set to 13.8 or 14 volts. If one cell runs ahead of the others, bleed off charge on that cell using a car headlight bulb for a few seconds. Repeat the process on the 'high' cell varying the 'bleed' connection time as required. With reasonable quality cells its possible to balance within a few mV.

Your method is probably the cheapest I couldn't imagine!
I should constantly keep an eye on cells voltage. I wonder how long the process would last. I suppose it's hard to estimate using solar charger.

 

[mikefitz]

I wonder how long the process would last.

Should not be too difficult to calculate. Usually the cells are shipped at 50% capacity. Any solar controller worth having or the BMS will report charge current and voltage so you can follow the charge process.
Having the series setup and the BMS in place prevents any cell for rising to too high a voltage. If playing 'human ballancer' with the cell bleed gets too hectic, disable the charger (or reduce the charge voltage) for a short time to get things under control.

Mike

 

[RCinFLA]

My battery is a 4s.
So I could buy 4 bucks (1 by cell) and pre-charge+top balance each cell without putting them in series and parallel?
I read in FilterGuy's tutorial this method is possible:

Or maybe just pre-charging each cell with 4 bucks, then parallel them for top-balancing with 1 buck?
If that's what you meant, the boost converter is useless... More and more cost savings!

The buck DC-DC converters are not input to output isolated. If you run four buck switchers from same input supply the batteries on each buck switcher output need to be isolated, not series connected in a battery array.

What I meant is divide cells into four groups. If four cells, one on each buck switcher. If eight then two parallel cells on each buck switcher. If 16 cells then four on each buck switchers. The more cells on each switcher the longer the top balancing time based on distribution of 15 amps output on each buck converter. But still provides more charging current then just using the the power supply with 5A or 10A max output.

For example if you have 16 cells, with four on each 15A output buck switcher then with each of four 280 AH cells starting at about 50% state of charge will take 0.5 * 280AH * four cells / 15A = 37.33 hours.

If you just used a 10 amp power supply with 16 paralleled 280 AH cell starting at 50% SOC would take 0.5 * 280A * 16 / 10 A = 224 hours.