Uneven battery charging and can't get to 100%

[elizabethii]

I have 3 EG43000EHV-48 inverters in split-phase backed up by 6 EG4 LifePower4 batteries. All of the batteries were purchased within 6 months of each other and commissioned at the same time. When charging the batteries, 1 battery appears to hit 100% while the others are around 70% charged. The DC bus voltage is at 56.2 volts, and the inverters promptly clip the power coming from the MPPTs because they're showing as charged. Ultimately, I'm leaning toward this being a BMS calibration issue, but I'm hoping for some guidance on the matter.

Below are screenshots of the pack info once the charging stops. This has been annoying me for quite a while. This morning, I tried turning the circuit breaker off on pack 3 to see if without it, the remaining batteries would get above where they're getting, but no luck. My thinking is if pack 3 were charging quicker and not discharging as deeply, without it, the bus voltage would be lower and allow the rest of them to fully charge, but that doesn't appear to be the case.



The batteries are in a rack and addressed in the order they're in the rack, top to bottom. Both positive and negative are fed into the bus bars from the top, which I know is not optimal, but if it were contributing to the uneven charge/discharge, I would expect packs 1 and 2 to be exhibiting the symptoms, not the middle. To get the rich details on the batteries through solar assistant, the inverters are in USE mode instead of the EG4-specific RS-485 protocol. Float is set to 54 and absorption is set to 56.2. I tried to run the inverters in the EG4 BMS protocol for a day to see if that fixed it, and the results appeared to be the same.

 

[Zapper77]

Cells look good.

Easiest way to calibrate the SOC is drain them down to 20%'ish and recharge. If you have a small load that only pulls 1.5 amps or less from each battery your SOC will fluctuate.

You may try updating the battery firmware. I updated to v3.26 last night and the SOC popped to 100% after hitting 56v. (I should check the bms logs from today...just to make sure I'm making an accurate statement).

 

[Zapper77]

Battery BMS ver 3.26 Log shows that the SOC went from 95% to 100% in just over two minutes.

@Markus_SignatureSolar maybe that was "the" fluke. Notice when the 200 popped up.

 

[Quattrohead]

Your float might be a little low, try 55v.

 

[Zapper77]

Maybe the magic number is 57 or 56.5v

Just check another

 

[Zapper77]

I'll add...your cells are balanced...so you can push the bulk up to 56.5 and see if the SOC resets to 100. If that doesn't work try 57v...I'd like to know if the BMS has a specific voltage where it will reset SOC to 100.

Edit to add. It's obvious all of your batteries are full in the images above.

 

[elizabethii]

I bumped float to 55 and absorption to 57, and we'll see what happens. My original voltages were set per the EG4 manual for the batteries because I didn't know what would happen if I exceeded them. This is my first foray into a battery that's not lead acid.

https://signaturesolar.com/content/documents/EG4/EG4-Lifepower4%2048V-24V%20Manual-2.3.1.pdf

 

[TomC4306]

I'll add...your cells are balanced...so you can push the bulk up to 56.5 and see if the SOC resets to 100. If that doesn't work try 57v...I'd like to know if the BMS has a specific voltage where it will reset SOC to 100.

Edit to add. It's obvious all of your batteries are full in the images above.

Except pack 3. 140mv cell delta.

bumped float to 55 and absorption to 57,

For five of the packs that is fine however it will just exacerbate the problem with pack 3. The imbalance in pack 3 is probably causing the BMS to turn it off for high voltage disconnect.

And how does one reach more than 100% state of charge?

The rest of the packs are enviably balanced.

I'd turn off pack 3 and cycle the remaining five for a day or two to see if they don't reset to 100% together. Then troubleshoot pack 3.

 

[Zapper77]

Except pack 3. 140mv cell delta.

Yep, I missed that one.

The higher float should help balance that pack.

 

[elizabethii]

I've been a few days now, and unfortunately, the 5 balanced packs haven't reset their SoC scales, and pack 3 doesn't appear to have balanced itself. Though, when not fully charged, pack 3's voltages no longer show the 100mv delta when it's fully charged.

Yesterday I tried to apply the BMS firmware update EG4 recently released to pack 3, and had no luck. I never could get it to pull board info, even after building a new cable that connects to the other RS-485 pins in the connector. I'm sure it's something I'm doing wrong, but I don't know what.

 

[Zapper77]

I've been a few days now, and unfortunately, the 5 balanced packs haven't reset their SoC scales, and pack 3 doesn't appear to have balanced itself. Though, when not fully charged, pack 3's voltages no longer show the 100mv delta when it's fully charged.

Yesterday I tried to apply the BMS firmware update EG4 recently released to pack 3, and had no luck. I never could get it to pull board info, even after building a new cable that connects to the other RS-485 pins in the connector. I'm sure it's something I'm doing wrong, but I don't know what.

What is your bulk voltage?

 

[TomC4306]

Float is set to 54 and absorption is set to 56.2

These batteries are fully charged. Physics determines this.
According to SA, all packs are at 56.2 volts. Just above 3.5v per cell. That's full.

Your problem is a BMS reporting issue on SOC.

 

[EastTexCowboy]

I fought with a similar situation for weeks, trying to charge separately, cycling the whole batch down to 20% (I have 12 of them), and a few other suggestions. Upping my float and bulk to 56v and 57v resolved most of it. Then I added a Victron smart shunt based on recommendations from several here. Once I got the shunt settings tweaked I haven't had any problems. SA reports SOC based on the shunt. I think the BMS data is unreliable for this. I've checked them via the BMS test software a couple of times since switching to the shunt and they're usually within 2% of each other. I do have one that lags a little lower but I figure I'll run it a year and see what it looks like then. I think part of the problem is that when the weather is temperate I only drain them down to 75% or so. Right now with this heat wave and drought I'm cycling down to 35-40%. I still get back to 100% most days but it doesn't stay there long.

 

[elizabethii]

What is your bulk voltage?

Bulk is set to 57v

 

[Zapper77]

Bulk is set to 57v

Do the batteries make it to 57v?

 

[sunshine_eggo]

I've been a few days now, and unfortunately, the 5 balanced packs haven't reset their SoC scales, and pack 3 doesn't appear to have balanced itself.

More patience is required. The balancing current is VERY low. It may take 3-4 days to adjust the balance by only 1Ah (1%).

Though, when not fully charged, pack 3's voltages no longer show the 100mv delta when it's fully charged.

Even with imbalanced cells, voltages in the operating range (3.1-3.4V) are usually very very close.

Raise float to 55.2V and let it run for a week.

 

[Zwy]

I fought with a similar situation for weeks, trying to charge separately, cycling the whole batch down to 20% (I have 12 of them), and a few other suggestions. Upping my float and bulk to 56v and 57v resolved most of it. Then I added a Victron smart shunt based on recommendations from several here. Once I got the shunt settings tweaked I haven't had any problems. SA reports SOC based on the shunt. I think the BMS data is unreliable for this. I've checked them via the BMS test software a couple of times since switching to the shunt and they're usually within 2% of each other. I do have one that lags a little lower but I figure I'll run it a year and see what it looks like then. I think part of the problem is that when the weather is temperate I only drain them down to 75% or so. Right now with this heat wave and drought I'm cycling down to 35-40%. I still get back to 100% most days but it doesn't stay there long.

What I have found for any battery with a BMS that has cell balancing is to look at the target cell voltage where you want balance to end. Let's say it is decided the cells should have an end charge of 3.50V per cell while charging. On a 48V battery with 16S, then absorption charge needs to 3.50V x 16= 56V minimum. If 3.55V per cell is desired, then 3.55V x 16= 56.8V minimum.

If the absorption charge is not set minimally at the sum of the cell voltages, it will never achieve cell balance. It can't because some cells may be slightly higher and some slightly lower. The charger only reads the sum of cell voltages.

Another aspect is half of cells could be at 3.45V while half at 3.55V. The absorption charge may be on a timer and will cut charging before pack balance is achieved. This can take several charge cycles to achieve balance. On some BMS's, such as the Batrium, the cell balance voltage can be custom set, thus allowing balancing to occur earlier than the preset value of 3.50V. Balancing will not occur until a cell exceeds 3.50V using the preset on the Batrium. If a custom setting is available, then it can be set to begin occurring just after 3.4V, usually around 3.45V. Then decide at what voltage cells are desired to be charged to, then multiply it by the number of cells. That is the minimum target absorption voltage and the pack needs to reach that point to get cell balance.

 

[elizabethii]

Do the batteries make it to 57v?

Yes, they hit 57 during bulk, then drop to 55.2 for float.

 

[Zapper77]

Yes, they hit 57 during bulk, then drop to 55.2 for float.

But do they actually hit 57v or do they have a cell OV prior to 57v and shut down charging?

 

[Zwy]

What I have found for any battery with a BMS that has cell balancing is to look at the target cell voltage where you want balance to end. Let's say it is decided the cells should have an end charge of 3.50V per cell while charging. On a 48V battery with 16S, then absorption charge needs to 3.50V x 16= 56V minimum. If 3.55V per cell is desired, then 3.55V x 16= 56.8V minimum.

If the absorption charge is not set minimally at the sum of the cell voltages, it will never achieve cell balance. It can't because some cells may be slightly higher and some slightly lower. The charger only reads the sum of cell voltages.

Another aspect is half of cells could be at 3.45V while half at 3.55V. The absorption charge may be on a timer and will cut charging before pack balance is achieved. This can take several charge cycles to achieve balance. On some BMS's, such as the Batrium, the cell balance voltage can be custom set, thus allowing balancing to occur earlier than the preset value of 3.50V. Balancing will not occur until a cell exceeds 3.50V using the preset on the Batrium. If a custom setting is available, then it can be set to begin occurring just after 3.4V, usually around 3.45V. Then decide at what voltage cells are desired to be charged to, then multiply it by the number of cells. That is the minimum target absorption voltage and the pack needs to reach that point to get cell balance.

Here is an example, the pack voltage is set at 56.2V and balance was set first for exceeding 3.49V to begin balance, then 3.50V. I've had 2 cells that have been slightly lower for several months, usually about 4mv, cells were 32 and 54. I've been working on getting this pack in balance the last few days by changing settings. The pack voltage needs to be just slightly higher than the sum of 16 cells. This allows the lower voltage cells to finally work higher in voltage. Will these stay in balance? Hard to say, I know when I'm using the well for several hours to water the garden and trees that it tends to make the cells bounce all over. I didn't water today so it presented the opportunity to get a better balance even with the mini split and one window air running. Heat index here is at about 100F today and tomorrow.

The cells from 33 to 48 are a mix of 280N and 280K, the cells from 49 to 64 are LF304 Grade B.