Parallel 2x16s 48v LiFePO4 inbalance

[matt1309]

Hi Folks,

I have just added a second 48v 16s LiFePO4 pack to be battery cabinet. (Originally purchased the first 16s last year). Both packs "Non EV grade" which may be the reason for the behavior I'm seeing. Both JK BMS but new pack is 200a the older is 150a.

I added the new pack and full charged the system however the new pack was still drawing around 800w when the tail current limit of solar system kicked in and switched to float. The old pack was 100% (0 amps going into old pack).

Now whenever I'm charging/discharging the packs there's a voltage difference of about 0.1 to 0.2v, with the newer pack being lower voltage in both a charge and discharge scenario (with less amps to correspond).

If it was the other way around I'd have said it was due to aging of older pack. Could it be that the new pack is of lesser quality/higher internal resistance. Or am I overthinking it and I just need to have a longer absorption phase to ensure both packs are at the same SoC. (in case it matters to folk I set absorption at 55.2, 3.45v per cell, conservative I know...)

(I know the obvious answer is just charge the pack fully with the tail current disabled and find out but I'm having a very wet summer so it's been a few days and I was unsure if there was a quick way to confirm/deny my guesses).

I've done the basics of checking for any dodgy connections but all seems to be solid/no obvious hot spots. The new pack doesn't seem to tank to ridiculously low voltage when discharging/charging just 0.1 to 0.2v behind the old pack, which gave me hope that it wasn't just a dodgy pack.

Either way I'm stuck with the pack so worse case scenario the old pack will just be degraded faster if the new pack is worse.

 

[sunshine_eggo]

If it was the other way around I'd have said it was due to aging of older pack. Could it be that the new pack is of lesser quality/higher internal resistance. Or am I overthinking it and I just need to have a longer absorption phase to ensure both packs are at the same SoC. (in case it matters to folk I set absorption at 55.2, 3.45v per cell, conservative I know...)

Yes. 3.45 may need a long absorption time to get two imbalanced backs balanced. Have you tried charging to 3.55V?

(I know the obvious answer is just charge the pack fully with the tail current disabled and find out but I'm having a very wet summer so it's been a few days and I was unsure if there was a quick way to confirm/deny my guesses).

Good.

I've done the basics of checking for any dodgy connections but all seems to be solid/no obvious hot spots. The new pack doesn't seem to tank to ridiculously low voltage when discharging/charging just 0.1 to 0.2v behind the old pack, which gave me hope that it wasn't just a dodgy pack.

confirm that you are connected "diagonally", i.e., your main leads are connected to the (+) of one battery and the (-) of the other battery.

Either way I'm stuck with the pack so worse case scenario the old pack will just be degraded faster if the new pack is worse.

Are these voltage discrepancies based on BMS reported values or based on measurements with the same voltmeter?

 

[matt1309]

Thanks for the reply.

I'll give longer balance a go next sunny day (Based in UK so could be awhile...), hopefully that's the culprit. Rather than the new pack being notable worse than the old.

I'm pretty sure I've checked it with volt meter as well but I'll double check in the morning as most of my time has been spent looking at BMS data. (However I have confirmed amps with CT clamp are lesser so I suspect difference in BMS measurements are not the cause).

Appreciate the diagonal advice however Both packs are connected to a central bus bars with equal length runs to each pack.

I also thought I'd seen particular cells dipping in BMS if I had some bad cells? Unless maybe they're all of lesser quality?

 

[sunshine_eggo]

Thanks for the reply.

I'll give longer balance a go next sunny day (Based in UK so could be awhile...), hopefully that's the culprit. Rather than the new pack being notable worse than the old.

Until you run both up into the area where you KNOW they are fully charged and at the same SoC relative to one another, I wouldn't worry.

I'm pretty sure I've checked it with volt meter as well but I'll double check in the morning as most of my time has been spent looking at BMS data. (However I have confirmed amps with CT clamp are lesser so I suspect difference in BMS measurements are not the cause).

By default, I don't trust any measurement until I correlate it with something.

Appreciate the diagonal advice however Both packs are connected to a central bus bars with equal length runs to each pack.

Even with bus bars, diagonal connection is preferred as well.

I also thought I'd seen particular cells dipping in BMS if I had some bad cells? Unless maybe they're all of lesser quality?

Again, until you run both up into the area where you KNOW they are fully charged and at the same SoC relative to one another, I wouldn't worry.

 

[matt1309]

Until you run both up into the area where you KNOW they are fully charged and at the same SoC relative to one another, I wouldn't worry.



By default, I don't trust any measurement until I correlate it with something.



Even with bus bars, diagonal connection is preferred as well.

View attachment 161284



Again, until you run both up into the area where you KNOW they are fully charged and at the same SoC relative to one another, I wouldn't worry.

Thanks for the help! In case anyone stumbles upon this later. @sunshine_eggo's advice was spot on.

Random aside: My battery shelf is essentially very oversized busbars and the second option in the above diagrams. However the original issue I had was solved by a combination of things.

TLDR: Needed longer top balances and old BMS needed calibrating.

At first I let the cells full top balance again. Setting absorption from adaptive (tail current was stopping it too early). To a fixed 6 hours, as a one off to force both packs to get fully saturated. This definitely helped as the old pack was fully saturated and the new pack was still drawing about 1000w. It took bloody ages.... but eventually got down to the new pack drawing 100w which I felt was close enough.

Although I'm convinced this definitely helped I was still seeing a large deviation between the two BMS's data, they are different models so i suspected maybe different internal resistances. However I double checked with a ct clamp and volt meter. Which solved the second issue (likely the bigger culprit), the voltage reading on old pack BMS was out by 0.1v, after calibration this was much better. As i mentioned before i'd already confirmed with ct clamp that the one pack was drawing more than the other, however I tried again to check if that was also an issue where the BMS current readings also needed calibration.

Turned out the ct clamp i had wasn't properly fitting around the wire and the readings where jumping all over the place (I've really oversized the cable as well as busbars in my cabinet). Upon using a larger CT clamp the current readings are as expected.