diy solar

diy solar

Top Balancing "How to"

Although you don't have a picture to prove 3.62V each.
Presumably there was some divergence in voltages as shown down around 3.1V?
It seems like cell #1 has only 50% of expected capacity, but make sure cells really are top-balanced.
Sometimes after the act of parallel top balance people report settling to unequal voltage (happened due to poor contact), can't be sure at the voltages you show.
Some people report capacity a bit below spec on new cells, but nothing like 50%. It isn't badly bulged or some obvious difference, is it?

Oh, discharge rate: When cell hit 2.5V, how much current was being drawn? Contact resistance causes voltage drop below actual cell voltage. But that often shows up as elevated voltage on adjacent cell, so less likely.

Double check all cells with DMM.
Firstly, when i got these cells, i put them in series and charged them. I was surprised that the cell 1 triggered BMS on 3.65v when the 3 others were 3.43v.
So i charged them separately to 3.62v (not in parallel), didn't dismantle series. Cell1 was 99% already full. All the 3 cells took about 50Ah (5 hours charging with 10A per cell). Finally, every cells had same voltage and normally the same capacity.
6 hours later, i put a load, small load of 3.5A. About 13 hours later, BMS triggered on cell 1 at 2.5v when the others were 3.26v

There is no bulging, no difference.
 
I agree with others it's most likely a connection problem. Disconnect the BMS and check your busbar connections. While possible, I would be surprised if the cell is bad. Did you check the resting voltages after you did your second pass? An active balancer will not help in this case.
Yes, they were all on same voltage.
 
Firstly, when i got these cells, i put them in series and charged them. I was surprised that the cell 1 triggered BMS on 3.65v when the 3 others were 3.43v.
So i charged them separately to 3.62v (not in parallel), didn't dismantle series. Cell1 was 99% already full. All the 3 cells took about 50Ah (5 hours charging with 10A per cell). Finally, every cells had same voltage and normally the same capacity.
6 hours later, i put a load, small load of 3.5A. About 13 hours later, BMS triggered on cell 1 at 2.5v when the others were 3.26v

There is no bulging, no difference.

It is sounding like that one cell has half the capacity of the rest.
There is a off chance high contact resistance caused false 3.65V reading in BMS. Doesn't seem likely since its voltage has been close to the others when charged.
Also, chance of contact resistance affecting low reading.

Can you poke DMM probes into cell terminals themselves, confirm cell voltages when full and when empty?
Recharge pack, then separately top off the poor performing cell? Maybe, it didn't actually get fully charged during that initial series charge.

Otherwise, sounds like you need one replacement cell.
 
It is sounding like that one cell has half the capacity of the rest.
There is a off chance high contact resistance caused false 3.65V reading in BMS. Doesn't seem likely since its voltage has been close to the others when charged.
Also, chance of contact resistance affecting low reading.

Can you poke DMM probes into cell terminals themselves, confirm cell voltages when full and when empty?
Recharge pack, then separately top off the poor performing cell? Maybe, it didn't actually get fully charged during that initial series charge.

Otherwise, sounds like you need one replacement cell.
I was also monitoring the cells with a DMM on charging and discharging.
The poor cell couldn't take anything from my bench power supply when i plugged in. It means that initially it was fully charged in series when the BMS trigged.

I will check the connections but i don't know what could go wrong.
The only thing to do is dismantle, charge cell 1 and discharge to see it capacity. Right ?
These cells are "new", well not grade A, but not supposed to be half capacity.
 
Last edited:
Right, if you can test that cell stand alone and measure its capacity, that is the most direct confirmation.

Seems pretty difficult for a cell to have half the capacity, given similar amount of material in its construction. Something like bad charging (e.g. while frozen), but not much else. Hope you can find a replacement, maybe get a refund for the bad one.
 
Right, if you can test that cell stand alone and measure its capacity, that is the most direct confirmation.

Seems pretty difficult for a cell to have half the capacity, given similar amount of material in its construction. Something like bad charging (e.g. while frozen), but not much else. Hope you can find a replacement, maybe get a refund for the bad one.
Only one day after i connected it back, i see a small oxydation on the Neg terminal on that cell, where the B- and small negative are connected.
What is the reason ? Is it due because the washer is in aluminium and the lug in copper ? Oxydation is between them.
 
Visible oxidation, might be some contamination.

Aluminum & copper do have more corrosion issue. Better if one is tin plated.

Bare aluminum is almost never used for electrical, except for wire which gets special treatment.
All aluminum UL listed parts I've seen are tin plated.
It appears aluminum cell terminals were meant to be welded.

Scrub oxide off aluminum surfaces, and use a corrosion inhibitor.

"Washer"?
There shouldn't be a washer sandwiched between current-carrying components. Only under a nut or bolt outside the conductors.

Have a picture of the assembly?
 
Only one day after i connected it back, i see a small oxydation on the Neg terminal on that cell, where the B- and small negative are connected.
What is the reason ? Is it due because the washer is in aluminium and the lug in copper ? Oxydation is between them.
It's called galvanic corrosion. It's a well known problem when connecting aluminum and copper, two common (and cheap) products to prevent it after you clean the terminal and busbar are Noalox and Ox-Gard.
 
Last edited:
Visible oxidation, might be some contamination.

Aluminum & copper do have more corrosion issue. Better if one is tin plated.

Bare aluminum is almost never used for electrical, except for wire which gets special treatment.
All aluminum UL listed parts I've seen are tin plated.
It appears aluminum cell terminals were meant to be welded.

Scrub oxide off aluminum surfaces, and use a corrosion inhibitor.

"Washer"?
There shouldn't be a washer sandwiched between current-carrying components. Only under a nut or bolt outside the conductors.

Have a picture of the assembly?
It is under a bolt of course.
I can't screw it if don't put a washer.
I need two washers with a bus bar.
 

Attachments

  • 216746012_310696560796061_362005509832606542_n.jpg
    216746012_310696560796061_362005509832606542_n.jpg
    314.6 KB · Views: 13
Last edited:
It is under a bolt of course.
I can't screw if if don't put a washer.
I need two washers with a bus bar.

Washer under screw head is not as problem. Even corrosion between that washer and ring terminal isn't really a problem, not in electrical path.
Copper ring terminal to aluminum cell terminal is the potential problem, and even if no visible corrosion, native oxide can be high resistance.
But not necessarily related to cell behaving like low capacity.
 
A shorter bolt may be in order also. The lugs need to be firmly held down to the terminal. But not so firmly that you strip the threads on the terminal, of course. Still, with the use of the washer you may be contacting more of the threads than if you used the next shorter bolt and no washer.
 
So I decided to ditch the included wires, swapped to 10awg ring terminals. Using extra studs/bolts to connect to the +/- of the 120ah LiFePo4 pack. Set my voltage to 3.65 (verified with multi meter before connecting), cranked my current to max. Connected Positive then Negative to my pack on SAME end vs diagonally opposed. After connecting, my PS switched to CC vs CV mode pushing 10, but voltage read 3.51-3.81 swapping back and forth. My cells were reading 3.352 on all four. Sometimes it would switch back to CV and voltage then steadied at 3.8. Bare in mind...I didn't touch ANY dials to cause it to dance around. Is this normal? Should I trust that it will come back to 3.65 the close the cells get there?

I read the document, watched Will's video (two disparities; Will connects positive then negative, Will connects positive and negative on same cell, not diagonally opposed) and the "set it and forget it" freaks me out for first time. I don't want to run these cells up if voltage says 3.8... Am I overthinking this?

Pics below I did NOTHING to the PS.
 

Attachments

  • 3.51CC.jpg
    3.51CC.jpg
    235.8 KB · Views: 18
  • 3.8_set3.65.jpg
    3.8_set3.65.jpg
    140.9 KB · Views: 18
Measure the voltage on the terminals of that power supply? Might want to turn the current down a bit, too, in case that stabilizes things.
 
Measure the voltage on the terminals of that power supply? Might want to turn the current down a bit, too, in case that stabilizes things.
I ended up saying eff it...bought the PS that is now on Will's site and got a good faith return in case these dials are messed up. Luckily 1 day shipping.
 
The voltage displayed by these cheap power supplies is virtually meaningless. Do you have a good multimeter? If not, get one. If so:
  1. With everything disconnected from the power supply, set the voltage looking only at the multimeter. Ignore the PS display.
  2. Periodically check the voltage on the cells, using the good multimeter (sounds like you have been doing this).
  3. When the PS says 3.8V and you really haven't touched anything, check it with the multimeter. That is, leave everything connected like it was, and just measure the voltage across the two PS terminals.
Chances are you'll see that you just can't trust the display on these power supplies.
 
So I decided to ditch the included wires, swapped to 10awg ring terminals. Using extra studs/bolts to connect to the +/- of the 120ah LiFePo4 pack. Set my voltage to 3.65 (verified with multi meter before connecting), cranked my current to max. Connected Positive then Negative to my pack on SAME end vs diagonally opposed. After connecting, my PS switched to CC vs CV mode pushing 10, but voltage read 3.51-3.81 swapping back and forth. My cells were reading 3.352 on all four. Sometimes it would switch back to CV and voltage then steadied at 3.8. Bare in mind...I didn't touch ANY dials to cause it to dance around. Is this normal? Should I trust that it will come back to 3.65 the close the cells get there?

I read the document, watched Will's video (two disparities; Will connects positive then negative, Will connects positive and negative on same cell, not diagonally opposed) and the "set it and forget it" freaks me out for first time. I don't want to run these cells up if voltage says 3.8... Am I overthinking this?

Pics below I did NOTHING to the PS.
That indicates a bad connection, possibly internal to the supply.
 
Back
Top