Top Balance Questions

[xc2]

I recently got 4 new 280AH batteries from Ailexpress and have been using since March. The pack would go out of balance at high state of charge. Here is what the BMS would say:
1 - 3.608
2- 3.366
3 - 3.364
4 - 3.368
Total = 13.7
While the BMS would try to balance, it never really happened.

Based on Will's instruction in "How to Top Balance LiFePO4 Battery Cells Quickly" on Youtube, I bought a power supply with the hopes of getting this pack right. The batteries are currently in parallel and the power supply is set at 3.6 volts. It's been taking 2.5 - 3.0 amps and is currently at 3.8 volts. I've had this on the charger for 40 hours now, so I've added about 110AH.

Since these batteries are in parallel right now does this mean the capacity is 4x 280AH or 1120AH? I started this exercise at about 80% soc. Do I have charge 224 AH to bring this pack to 3.5 volts? Is it okay to leave these on the power supply for another 40 hours to achieve 3.5v?

 

[Sojourner1]

Parallel- voltage stays the same, ah is combined for a total.

Series- voltage combines for a total, ah remains the same.

I might be wrong but I thought that when using a bench power supply once you set the max voltage it holds there and the amps will decrease as the batteries become full.

 

[upnorthandpersonal]

the power supply is set at 3.6 volts. It's been taking 2.5 - 3.0 amps and is currently at 3.8 volts.

It can't be at 3.8V. Is that a typo? That's not good - should not be more than 3.65.

 

[xc2]

yes, typo - I mean 3.38

 

[xc2]

It's been taking 2.5 - 3.0 amps and is currently at 3.8 volts.

Sorry! I mean 3.38 volts. It's taken 40 hours @ 2.5 amps to move from 3.36 volts.

 

[upnorthandpersonal]

yes, typo - I mean 3.38

Ah, good

Yes, it will take a while since you have an 1120Ah 'cell' essentially. However once you get near to the end 3.4 to 3.5V things go rather quickly and you should see the current drop rather fast. See curve below:

 

[xc2]

I might be wrong but I thought that when using a bench power supply once you set the max voltage it holds there and the amps will decrease as the batteries become full.

Thanks - this appears to be the case. If I increase the voltage, the amps will increase too. I'm new to this and trying to learn. I increased the voltage by .1 or .2 just for a second to see what happens.

 

[Sojourner1]

If I recall from one of Wills videos he increased the voltage but was there to watch then decreased after it started to climb. This can be dangerous if you allow yourself to get side tracked with something else.

 

[upnorthandpersonal]

Don't play with the voltage. 3.6 (even 3.55) is more than enough. You just increase the risk by increasing the voltage. Make sure you measure the voltage at no load, hook it up, and be patient I'm doing this with 32 cells individually. That's just the price you pay for doing a DIY pack properly.

 

[xc2]

Right - I have this set at 3.6, but I think I will reduce to 3.55 when I can't keep an eye on it. I took it up to 3.7 - 3.8 on Sunday for less than 10 seconds. What good is a new toy of you can't twist all of the knobs!

 

[upnorthandpersonal]

Do make sure of the following: once you see the current drop, and it's done, disconnect it. It is possible to overcharge LiFePO4 cells by keeping the power on all the time. For reference:

https://files.ev-power.eu/inc/_auto88/_info/Doc/GWL-Power-Cell-Damage-OverCharge.pdf

 

[Ampster]

What good is a new toy of you can't twist all of the knobs!

Seems like I remember that was the thinking of the technicians at Chernoybol. Only problem was that their knobs were hooked up to something bigger than one cell.

 

[Turd Furguson]

I just did this with 8x150ah cells, so about the same capacity. It takes forever, and them some.

Set your bench charger to 3.65v. and walk away for a few days. Amps will drop to zero. It feels magical after staring at them eternally at 3.45v. No intervention required.

If you want to speed things up, connect them in a 4s configuration and monitor until the first cell hits 3.65. Then swap to the 4p configuration to top them off.

Do it right and then never again.

 

[Will Prowse]

Power supply is the cheapest way to do it. I think it's best to charge up all the cells in 4S configuration first, then attach in parallel, then trickle that last bit of absorption with the power supply as I mentioned in the video.

If you are charging the cells from 50% SOC, it will take weeks haha! Much easier to use cheap 12v charger and cheap common port bms to get to high soc first, then balance them second.

 

[Will Prowse]

Another option is to parallel connect a few of these: https://www.aliexpress.com/item/328...earchweb0_0,searchweb201602_,searchweb201603_

But yeah, cheapest method is the one I just mentioned above.

 

[MisterSandals]

I'm doing this with 32 cells individually.

I am not sure why you do this individually, it has to take more time charging one at a time (and a lot more effort). Maybe you like to tick off progress as you get thru them?

I think that top balancing is a process that teaches the cells to charge similarly, while connected in parallel.

And with self discharge of the surface charge down to around 3.35v for my cells, i would expect them to be somewhat dissimilar after some settle for many days and some for only hours.

Certainly a few philosophies regarding top balancing in parallel vs individual cell charging.

I see Will just chimed in. Regarding charging in series first and finishing in parallel, i would think a top balance thru most of the charge range would be best. The longer the training process, the better the result of balancing.

 

[upnorthandpersonal]

@MisterSandals Check this document for one: https://files.ev-power.eu/inc/_auto88/_info/Doc/GWL-Power-Cell-Damage-OverCharge.pdf

Of course, that's just one reference. I also have a clear charge log of each cell this way. If anything goes wrong later with a particular cell, I have a capacity log, charge log, etc. per cell. Yes, takes a bit more time, but you can do this in parallel - you just need a bunch of power supplies.

top balancing is a process that teaches the cells to charge similarly,

I disagree with this. Top balancing is just making sure each cell is at a known and equal state of charge. The reason you can't do this at 3.4V for example is because the curve is too flat and even small discrepancies mean large capacity differences. If you're at 3.6V, even somewhat large discrepancies are negligible. To put that into perspective (and I'll use bottom for a change), the amount of power in the battery cell between 2.0V and 2.5V is negligible. The amount of power in a battery cell between 3V and 3.5V is over 90% (even 95%) of the cell. Just 0.5V difference between both.

 

[Ampster]

Power supply is the cheapest way to do it. I think it's best to charge up all the cells in 4S configuration first, then attach in parallel, then trickle that last bit of absorption with the power supply as I mentioned in the video.

If you are charging the cells from 50% SOC, it will take weeks haha! Much easier to use cheap 12v charger and cheap common port bms to get to high soc first, then balance them second.

Weeks? It would be months for my 32 cells.
Good advice. I have a 24 volt 60 Amp power supply ordered from Alibaba to power my capacity tester. I have also ordered 32 of those 280 Ahr cells. I can hook the cells temporarily in 4P8S and charge them to 3.5 or 3.55v if my power supply can go to 28 volts. Then I can do the final Constant Voltage (absorb) finish in parallel with my smaller 15 Amp power supply set to 3.6 volts.

Another option is to parallel connect a few of these: ......

To do that successfully, it is important that they be isolated power supplies or they don't work together well in parallel. Most good ones are isolated but it is worth checking to be sure.

 

[Ampster]

I think that top balancing is a process that teaches the cells to charge similarly

I thought charging memory was a NimH concept. LFPs are too dumb to remember.

 

[xc2]

Do make sure of the following: once you see the current drop, and it's done, disconnect it.

Current drop to what? My cheap Amazon power supply can go to 10 amps, but it's only saying 2. I think it was charging at over 3 amps when I started on Sunday.

 

[upnorthandpersonal]

It'll go to pretty much zero really quick, similar to this:

 

[Ampster]

Current drop to what? My cheap Amazon power supply can go to 10 amps, but it's only saying 2. I think it was charging at over 3 amps when I started on Sunday.

I really doesn't matter if you stop now at 2.1, 1.0 or .5 Amps. Look at the discharge curve that @upnorthandpersonal posted above and it is pretty steep once you get to 3.5 volts. A few Amps might be only a few Watthours of capacity. But according to the chart it is not long to get to zero. It is your time versus how much accuracy you want. He is doing them all individually so I understand why he would go to zero. In my case I may go to zero or a little above because with 32 cells and a 15 Amp power supply. I can only pump in 0.5 Amp per cell anyway. I am just going to make sure I am at the steep part of the curve.

 

[MisterSandals]

Check this document for one:

Thats one of the best articles i have seen, thanks.
I see why you are avoiding a long slow top balancing in parallel: a huge set of cells and a smallish power supply keeps them in the overcharge zone for a long time.

I also agree with the day to day charge philosophy of lower charge cutoff. I have been using 13.6v (3.4v) for a few weeks and find it works well in 2 aspects. First, i see no diminished capacity and secondly, the balance seems a LOT better (i am not using a BMS just yet).

Thanks, good discussion. My view is somewhat changed. I still like the idea of top balancing in parallel most. BUT, having sufficient charge current, specifically for very large groups of large cells, that may not be attainable to keep the charge time short.

 

[MisterSandals]

I thought charging memory was a NimH concept. LFPs are too dumb to remember.

Maybe.
I had quite a few chemistry classes and have a hard time reasoning with my thoughts that there is some “learning” during the top balance process.

Maybe this is the thing i learn today?

 

[Ampster]

I had quite a few chemistry classes and have a hard time reasoning with my thoughts that there is some “learning” during the top balance process.

Chemistry class? That was almost 60 years ago in high school. I forgot so much I couldn't even help my daughter through her high school chemistry 18 years ago.