Why test all received cells?

[dyihans]

I have just ordered 16 280Ah LiFePO4 cells and an EBC-A40L tester (+BMS and other things).
I see that many test all their cells, so they can say "all the cells were in the X to Y Ah range".
Testing all cells will take a lot of time.
What about this:
1 Do the top balancing (not only the EBC-A40L, I will probably help with other power supplies to speed it up at the beginning)
2 Build the battery bank
3 Put the wanted test load (probably not critical) on the whole bank via an inverter or whatever
4 Let the loading continue until you have located the weakest cell (first to reach the knee, 2.9V ?)
5 Disconnect that cell from the bank and test it
6 If it tests to 280Ah or more, why would I waste time to test all the other cells?

 

[JoeHam]

You will find various levels of persnicketiness on the forum, there is no absolute right or wrong.

It’s your money, your choice.

I think your plan is solid and if you continue to monitor individual cell voltages via your BMS I don’t see any problems at all.

When my cells are behaving I check in on the cell level voltage once a month at most.

 

[wattmatters]

4 Let the loading continue until you have located the weakest cell

There may be more than one reason for a cell to be first to that voltage than it being "the weakest".

Whether or not you test none, some or all cells is personal choice I'd have thought.

 

[time2roll]

How about just build the pack.
Charge it up and check balance. Only boost low cells as needed. Charge the needed cells one at a time while still assembled.
Put the battery in service and see if capacity performs as expected.

 

[brb58]

I bought 10 280AH Eve cells for my 2 12V batteries so that I could have 2 spares. I top balanced all of them and then did a load test on every cell. All were 280 or 281AH. I know I have 2 good cells leftover.

 

[RCinFLA]

You obviously want to get the capacity you paid for, but the other very important factor is matching of cells. Poorly matched cells will require balancing more often and the cells will age at different rates.

You should load the cells with 0.2 C(A) to 0.4 C(A) to get a good test on matching. EBC-A40L tester, at 40 amps max is okay but on the low end of load test range for 280 AH cells (0.2 CA is 56 amps load)

I suggest you program the EBC-A40L tester with test segments to periodically unload the cell during discharge for a 5 minute rest period every so often along the discharge test run to allow cell voltage to recover to open circuit voltage, then pick up the 40 amp discharge load again. The amount of terminal voltage slump you see in the first 120 to 180 seconds after reapplying 40 amp load is a good test for matching. Ideally you want all the cells to exhibit the same matching of terminal voltage slump along the SOC on discharge curve. At 40 amp load you should only see about a 40 mV terminal voltage slump after two or three minutes of 40 amp load for a new 280 AH cell. If you get significantly more voltage slump you likely have used cells. Over the useful life of cell the voltage slump will increase 3x or more at given load current. This aging terminal voltage slump may limit their use for moderate to high load current from cell.

The tester accounts for the zero load gaps in the discharge test profile in its AH cumulative tally.

Make sure you use the EBC-A40L tester remote cell voltage sensing wires directly to terminals of cell to avoid the voltage drop on the load current cables. Also temp of cells matters. Try to do testing with all cells close to 20-25 degs C temp range. Below 15 degs C you will not get full AH yield from cell.

 

[time2roll]

You obviously want to get the capacity you paid for, but the other very important factor is matching of cells. Poorly matched cells will require balancing more often and the cells will age at different rates.

At what point does a person set those poorly matched cells aside and order more hoping for a better match?
Most people buy just enough cells and live with it.

 

[MrThisIsME]

If ordering from the Asian market, testing is done to ensure honesty. While others mentioned it is a hit or miss, what if you spent 45 or 90 days building your battery and find 1/2 are bad. Good luck. Most that test each cell do it immediately upon arrival that way return time frame is in effect or credit card reversal.

 

[Browneye]

My cells have been in service for nearly a year...still have not 'tested' them. They just work.
Build the pack and go.

 

[sunshine_eggo]

I have just ordered 16 280Ah LiFePO4 cells and an EBC-A40L tester (+BMS and other things).
I see that many test all their cells, so they can say "all the cells were in the X to Y Ah range".
Testing all cells will take a lot of time.
What about this:

0 read and follow the guide. You'll understand why once you read it.

1 Do the top balancing (not only the EBC-A40L, I will probably help with other power supplies to speed it up at the beginning)
2 Build the battery bank
3 Put the wanted test load (probably not critical) on the whole bank via an inverter or whatever
4 Let the loading continue until you have located the weakest cell (first to reach the knee, 2.9V ?)
5 Disconnect that cell from the bank and test it
6 If it tests to 280Ah or more, why would I waste time to test all the other cells?

 

[Nobodybusiness]

At what point does a person set those poorly matched cells aside and order more hoping for a better match?
Most people buy just enough cells and live with it.

Or order an active balancer to try and remediate.

 

[sunshine_eggo]

Or order an active balancer to try and remediate.

The performance of active balancers can be very misleading. When you see 2A, 4A, 6A or whatever, that's typically based on a full 1V difference between adjacent cells, which you'll never see on LFP. Yes, way better than 30-70mA passive, but it's not typically a game changer.

 

[Nobodybusiness]

The performance of active balancers can be very misleading. When you see 2A, 4A, 6A or whatever, that's typically based on a full 1V difference between adjacent cells, which you'll never see on LFP. Yes, way better than 30-70mA passive, but it's not typically a game changer.

I can only tell you what i see. The alternative is to to buy another battery and hope you get it and it’s not a runner and full capacity.

The runners always seem to shut off the BMS every few seconds and they gallop to whatever your high voltage is set at or down to the bottom under load.

I have corrected runners with 5a Heltech but only at rest and almost 100% full charge.
During charging it helps but still shuts off BMS just slightly less.

 

[RCinFLA]

At what point does a person set those poorly matched cells aside and order more hoping for a better match?
Most people buy just enough cells and live with it.

As the saying goes, 'it's amazing what a hungry man will eat.'

I would prefer not to have more than about 25% difference is cell voltage slump for same load current in a series connected stack.
The closer the better. The greater the mismatch, the more balancing time required to keep them inline.

Cell voltage slump is logarithmic with current. For very light cell current load there is 5 to 10 millivolts of slump. To get good indication of cell condition you need to have enough load current to get up to at least 40 mV of terminal voltage slump from no-load current, rested cell voltage.

The 5 to 10 millivolts of voltage slump for very small cell current is also the reason why just passively paralleling cells will not get them very well balanced. You will end up with 10-20 mV difference in cells' rested OCV which is still quite a bit of SOC imbalance.