Top Balancing "How to"

[Horsefly]

Two potential problems with the cheap Amazon power supplies that I've either heard about here on the forum or experienced myself:

1. Some of the power supplies tend to go up in voltage as the current gets low. I've never seen this, but a couple of posters were certain that they had not touched the voltage dial, but found the cells ended the top balance at a higher voltage. Obviously this can be disastrous.
2. I have had one power supply that turned into a load once it was turned off. I had been charging or top balancing (don't remember which) and then turned the supply off and disconnected it. A few minutes later I started to hook it back up (while it was still off) and I got a pretty big spark. Thinking it maybe was an inverter-like pre-charge thing, I disconnected it and immediately reconnected it. Again, a big spark. I then used a clamp meter and confirmed that there was around 4A flowing INTO the turned off power supply from the cells. Not good.

 

[EcoRV]

Hello i have built several 12V LIFEPO4 batteries in the past but in this instance in have 16 of the new 280K cells which have all completed the top balancing process. As their voltage varies at such a high state of charge i wanted to ask how long i should let them rest before matching them into 4 separate 12V groups.

Sorry if this has been asked previously,
Danny

 

[MisterSandals]

which have all completed the top balancing process.

You connected them all in parallel and charged them to 3.65V? (aka top balanced)

What cell voltages are you seeing/measuring? The shouldn't be very different.

What is your plan on choosing which cells will be matched with other cells to form each 12.8V battery?

 

[John Frum]

You connected them all in parallel and charged them to 3.65V? (aka top balanced)

What cell voltages are you seeing/measuring? The shouldn't be very different.

What is your plan on choosing which cells will be matched with other cells to form each 12.8V battery?

I think he is talking about separating them to see how each one settles individually.
Great idea!
Overnight is good, 48 hours better, I've read the factory does a 30 day self discharge test but only on the cells that pass all the shorter tests first.

 

[Bluedog225]

Subscribed

 

[MisterSandals]

I think he is talking about separating them to see how each one settles individually.

Yah, if they were in parallel then a decent settle period is good.
I was working on a theory that cells that initially (within the first hour) discharged quickest
were the lower capacity cells.
I had to put the cells into service in the greenhouse before doing enough tests and measurements unfortunately.

But, label your cells (I like using letters A, B, C... because cell positions are numbers) and take good notes. Take an early voltage reading after separating the cells. Then an hour later. Then the next day.

I'd love to hear if you have a correlation between quick initial self-discharge and lower capacity. Its just a theory i have.

 

[John Frum]

Yah, if they were in parallel then a decent settle period is good.

That is part of why I prefer to top off the cells individually.
Nothing magical happens by combining them in parallel and it may hide a stinker.

 

[MisterSandals]

combining them in parallel and it may hide a stinker.

Not questioning or doubting your approach, just curious as to how would you detect a stinker?

I am trying to figure this out, whether there are clues to capacity variations, short of a fill discharge capacity test.

 

[John Frum]

Not questioning or doubting your approach, just curious as to how would you detect a stinker?

I am trying to figure this out, whether there are clues to capacity variations, short of a fill discharge capacity test.

If your top off the cells individually and leave them to settle.
Good cells will quickly drop to ~3.45 volts within 12 hours and not drop much more over the balance of the month.
A real "stinker" or "lemon" or whatever you want to call it will drop through ~3.45 volts and keep dropping.
When its in parralel with 4, 8 or 16 other cells the effect is partially masked.
Also, most people don't wait very long to go to the next step in the commissioning process.

 

[time2roll]

So it is best to separate for the rest period?

 

[EcoRV]

Good conversation here. I took the 16 raw cells and built into 4 12v pacs with JDB BMS and charged till i hit a high voltage disconnect at 3.6. I then removed the BMS and reconfigured the busbars into a 4 cell parallel 3.2V configuration and top balanced those 4 cells with a power supply to 3.65v and immediately disassembled the 4 cells into a stand-alone config and placed them in a "Top Balanced waiting area". Then i just repeated the process 3 more times.

I guess i will wait till morning test / tag the cell with voltage readings and then place the closest matched cells into their respective 4 -12v sets.

 

[EcoRV]

I think he is talking about separating them to see how each one settles individually.
Great idea!
Overnight is good, 48 hours better, I've read the factory does a 30 day self discharge test but only on the cells that pass all the shorter tests first.

Yep i separated for the to settle individually. Then i will group them by the settled voltage and configure to 12v batteries

 

[EcoRV]

You connected them all in parallel and charged them to 3.65V? (aka top balanced)

What cell voltages are you seeing/measuring? The shouldn't be very different.

What is your plan on choosing which cells will be matched with other cells to form each 12.8V battery?

Thanks for the reply - i posted a better explanation above. Out of 16 cells I had 1 runner that hit 3.65 while the others in that 12v pac were at 3.3 and 3.4 - Thats why i am breaking the 12v packs down and then going to regroup them by fully charged / top balanced / rested voltage values.

 

[Just John]

Hello i have built several 12V LIFEPO4 batteries in the past but in this instance in have 16 of the new 280K cells which have all completed the top balancing process. As their voltage varies at such a high state of charge i wanted to ask how long i should let them rest before matching them into 4 separate 12V groups.

Sorry if this has been asked previously,
Danny

If they are fully charged, just a couple hours, you don't need to hold them at 100% full unless you are testing self discharge. Put them in a pack and capacity test otherwise, which also separates the cells. Putting them in a pack would also allow you to test self discharge. When you top balance, they should settle within about 4 hours to about 3.5v. Any significantly weaker cell should show up in the first few cycles.

 

[Dedou]

How close should the current drop when top balancing? I have 4x 90Ah cells as a patient. On the first try, I charged the cells in parallel at 3.60v (8A charge) and stopped charging when the current had dropped to 0.020A then I waited 1.5 hours, switched to series and did a capacity test with the BMS. The capacity test ended already at 12V even though the lowest voltage was set to 10V.

I suspect the reason was "cell different volt protect" this parameter my daly bms won't let change, it's always 0.26V. Bms did not report any error code. After 1.5h of rest, the diff volt was 0.063V, at the beginning of the test 0.008V and at the end of the test 0.274V (lowest cell 2.868V).

Now I recharging the cells in parallel, the current has dropped to 0.016A and doesn't seem to be dropping from it. Will I stop charging already, or will I still let it charge?
How many differences in current should there be when the cells have rested?
Do I let the cells rest in parallel, or series? I am going to do the next capacity test with JBD BMS. I use the DL24P capasity tester.

 

[Just John]

How close should the current drop when top balancing? I have 4x 90Ah cells as a patient. On the first try, I charged the cells in parallel at 3.60v (8A charge) and stopped charging when the current had dropped to 0.020A then I waited 1.5 hours, switched to series and did a capacity test with the BMS. The capacity test ended already at 12V even though the lowest voltage was set to 10V.

I suspect the reason was "cell different volt protect" this parameter my daly bms won't let change, it's always 0.26V. Bms did not report any error code. After 1.5h of rest, the diff volt was 0.063V, at the beginning of the test 0.008V and at the end of the test 0.274V (lowest cell 2.868V).

Now I recharging the cells in parallel, the current has dropped to 0.016A and doesn't seem to be dropping from it. Will I stop charging already, or will I still let it charge?
How many differences in current should there be when the cells have rested?
Do I let the cells rest in parallel, or series? I am going to do the next capacity test with JBD BMS. I use the DL24P capasity tester.

I would let it drop to under one amp, as long as after an hour their resting voltage is above 3.5v you can be certain they were fully saturated.

Probably one of your cells is significantly less capacity than the others and is triggering low voltage disconnect.
Did any cell reach 2.5v (with a cell at 2.8v, I suspect it did under load and you just missed it)? If this is the case, it won't matter which BMS you use, as long as it works.
You can let cells rest in either parallel or in series, manufacturers usually specify a one hour rest after a full charge for capacity testing purposes.

I would have just charged them back up while still in series, it's much faster that way. Just be sure to use a BMS.

Two resources you should check out:

Top Balancing LiFePo4 Cells using a low cost benchtop power supply.

To get the paper, click on the orange at the top of this page. This is a tutorial with detailed steps on how I top-balance my LiFePO4 Cells using a low-cost benchtop power supply. To get the tutorial, click on the orange "Download" button...

diysolarforum.com

Explanation for Beginners of Top and Bottom Balance

This is a beginners explanation of top and bottom balancing cells that should help you understand what it is all about. To get the document, click on the orange button at the top of the page. NOTE: This is not intended to be a tutorial of how...

diysolarforum.com

 

[Dedou]

Thanks for the answer.

Yes I charged back up with the bms and then the switch in parallel to load the top balance. it is likely that they have lower capasity because they are Liitokala cells, however i thought it would be more than 55ah what i tested.

I didn't think the current would be lower when discharging the cells, i.e. the 2.8v cell after the test could drip during the test 2.5v

 

[Just John]

I didn't think the current would be lower when discharging the cells, i.e. the 2.8v cell after the test could drip during the test 2.5v

Yes, that is exactly what happens, as soon as the load is disconnected, they will jump back up to 2.7 to 2.8v. Same thing happens when charging, but in the opposite direction. I call it the bounce effect.

 

[noenegdod]

Yes, that is exactly what happens, as soon as the load is disconnected, they will jump back up to 2.7 to 2.8v. Same thing happens when charging, but in the opposite direction. I call it the bounce effect.

This is similar to the ground water depression that occurs when drawing water out of a well. Water level drops while removing water and when removal stops, the level recovers.

 

[Dedou]

Yes it is so that 2 cells have a lower capacity, another test done and still a capacity is 55ah. Cell voltage after test. 2.747 + 2.713 + 3.079 + 3.044.

Next I do a test for each cell, after that I try to get a refund for the price, probably useless what I’ve read, I still try.