Peculiar system balancing suggestions.

[Keith C]

So I have eight of those monster BYD batteries I'm wrestling with. I need to get them fully charged so I capacity test them, so I can evenly pair them into my 48V system. I need to get them into balance to capacity test them. These batteries are very resistant to a lot of the typical solutions. Taking them apart is a huge painful task I'm not really considering. Because taking them apart is not an option I'm forced to use the original cell leads that are going to the original mystery BMSes. I've abandoned the original (oem) BMSes in place and tapped out of the cell wiring harness the 8 (9) required wires leading to each cell (they're in with 8 individual temp sensors). I've spliced the cell wires into the harnesses for the the standard [OverKillSolar 8s BMS 100a LifePo4 with M6 Threaded Terminals] Style BMSes.

Firing up the BMSes I see the cells all within about 5mV. Putting a lab supply on them they placidly start charging (at about 1A) as I feel my way along. They stay matched until one takes the lead and swoops up to 3.55V where I have the BMS kick off on "Cell Protection". So there I sit with 7 cells at 3.4V and one at 3.55. Dead in the water. The imbecilic balancing provided stops if charging stops instead of "balancing" charging-or-not a "high cell protection" cell. IF the dang BMS balanced the "Cell Protection" cells the entire battery would eventually be ratcheted up to all cells being top balanced. But alas, the system just sits there stalled as the battery is disconnected.

What I'm doing now is unplugging the cell cable and jacking in a resistor shunt across the offending cell which drains it a bit. I have to guess how long to discharge the cell to get it in line with the rest of the herd. It's a total crapshoot! Sometimes the same cell comes back to "cell protect" sometimes a different cell becomes the top-dog.

I suspect this process will wear out the connectors a couple of times and it will take about 3 years to complete doing 8 of these monsters.

Does anyone have any suggestions? Perhaps a passive balancer I could plug in that would automatically take down the high cell so I don't have to play wack-a-mole. (That would be a mole of electrons in this case.)

A limitation is that I'm afraid of much more than 0.5A of balancing current because these batteries use thin circuit board traces to get to the cells.

I welcome any suggestions - out of the box even.

​

 

[codfish]

Turn the charging voltage down till the current equals or less than the balance current. Then any runner cells will be balancing and laggard will be charging. Might take awhile to get there, but with the laggards at 3.4 you are getting close.

 

[GXMnow]

You can combine both ideas.

Reduce the charge current to about 200 ma (0.2 amps) AND add a resistor across the high cell. If the shut off is 3.55 volts, then choose a resistor value and charge current combination that hits 3.54 volts. 0.2 amps x 3.5 volts = .7 watts, so a 1 watt resistor will run warm but hold up. Ideally that would be 17.7 ohms. 22 ohms is a standard value. 3.54 volts / 22 ohms = 0.161 amps of charge current. That will not allow the cell with the resistor to go any higher, while all the other cells keep getting the 0.161 amps. Parallel two 22 ohm 1 watt resistors, and you can up the charging current to 0.322 amps. Do you trust the balance leads for that much? I would actually put the resistor as close to the cell as possible to ensure lead resistance does not add to the cell voltage. And you can err a little lower on the current to dial in the cell voltage you want. 22 ohms x 0.161 amps = 3.542 volts x 0.161 amps again, gets the resistor watts to 0.57 watts. Stick with the 1 watt rating. Adding a third 22 ohm resistor and you can up the current to 0.483 amps. Ideally, you want to connect the resistor to each cell as it reaches the desired set voltage. The resistor ends up taking the charge current so that cell stops charging.

On my 360 amp hour Li NMC battery bank, I used two 8 ohm resistors on the highest two cells and 0.5 amps of charge current. That pulled all the cells to 4.0 volts. I had to move the resistors a few times over 2 days to get that pack perfectly balanced. I was very happy when I got my second batch of cells, they were near perfectly balanced when I got them. Only problem was, they were charged higher than I ever bring my baery bank, so I had to pull them down a bit before I could add them into the system.

If you are not in a hurry at all, and don't want to baby sit this, you could just put a 22 ohm resistor every cell, and set the charge current at the 160 milliamps. This will take days though. That would pull every cell to 3.52 volts. But hand test the resistors. Even 1% tolerance resistors could be 1% high to 1% low. 21.78 to 22.22 ohms. That gives cell voltages from 3.485 to 3.553. Buy twice as many resistors as you need, and hand pick the closest matching ones. Put the ones that read a little lower on the highest cell voltage. If you do all of the cells like this, it will be VERY SLOW! The actual current going in or out of any cell is tiny. And when the cell hits the voltage of the resistor drop, that cell drops to zero current and will just hang there. If you go with the 2 or 3 resistors in parallel across each cell, use double or triple the current as above, and it will go a little faster, but still days. If you are going to do the resistor on every cell, I also recommend connecting the string of resistors across the charger first, and measuring all of the voltages. Each resistor should have your desired cell voltage across it. If you misread a resistor color code, it will show up real quick. Doing some quick math, you could go down to 5 ohm 5 watt resistors. (They are typically 5% tolerance, so you will absolutely need to hand match those) but then you can up the charge current to 0.7 amps for 3.5 volts per cell. Keep in mind 16 of these in series at 0.7 amps it going to be throwing away over 39 watts into heat.

 

[Keith C]

Great ideas!!! Freaking outstanding.

Cod; I would be so happy if the dumb BMS would allow that but it appears to forbid balancing at anything less than about 400mA of charging current and I suspect its balancing is around 100mA so it can't actually help balancing at all since it still allows the high cell(s) to still gain charge. Very frustrating that the BMS blocks every possible way of actually allowing ANY form of balancing. Image when a normal 20~30A charge is happening. You'd maybe get 3 minutes of 100mA balancing against 30A. Sheesh.

GXMnow; This is a great idea! Let me get this straight:

Set up an ohm's law based bleed resistor that equals the desired cell voltage at a specific desired low current.

Shunt cells as they near the desired resistor/cell voltage. Eventually they all will get shunted.

Charge the entire battery at the current used to calculate the cell voltage resistor shunts. (XmA)

As each cell reaches the calculated fixed resistor balancer voltage the shunt will essentially suspend further charging of that cell.

Do I understand the game plan?

​

 

[codfish]

I'm a big fan of the capacitive active balancers ( ebay or amazon). I have some less than Grade A, Eve 280's and after expanding the harness to include the active balancers everyone stays happy. (Just disconnect them if battery will be unused for months.) The batteries would self discharge over winter and get way out of balance. Re-top balancing every spring was not desirable. These handle it fine as long as the first charge of the year just barely gets in to the "knee".

 

[GXMnow]

As each cell reaches the calculated fixed resistor balancer voltage the shunt will essentially suspend further charging of that cell.

Do I understand the game plan?

Yup, that is the whole idea.