20% capacity loss and bulging after 2.5 years of use

[Suijkerbuijk]

Very interesting. Should I expect the Grade B cells to now be stable for a while at their reduced capacities?

Yes it can be and the lower cell is the leader of the pack.
Than again stay nice in that 3-3.45 volts it will works for a longtime for about 80% of its rating

 

[Zwy]

Very interesting. Should I expect the Grade B cells to now be stable for a while at their reduced capacities?

I would charge/discharge a few times after charging to 3.60V. LFP relies on movement of lithium ions and if the pack had been out of balance for some time and not charged to 3.60V, then it is possible it might recover some capacity. Temperature is a huge factor on capacity tests and if you want to compare previous results, then temp needs to be the same.

Having said that, I did buy 8 280K Grade B cells from Amy to fill out a pack with 8 280N Grade A cells. Those Grade B's have never held balance, she sold them to me as not matched and it certainly shows. The Batrium passive balancing does balance them but I had to reduce bulk charging voltage as one is a severe runner and if charge rate from the mppt's is high, it could hit HV disconnect. I will be taking all 16 cells out and pairing those with a JK active balancer for occasional use in my work trailer. I told her those were the worst cells I had bought anywhere. The 280K V3 I have from her now in use hold balance thru the entire range of 2.50V to 3.50V. Very impressive.

 

[AshleyL]

Very interesting. Should I expect the Grade B cells to now be stable for a while at their reduced capacities?

Yes. Then the yearly capacity reduction will be at 2-3% (calendar life, whether you use it or not).
If you tend to cycle all the way down below 10% and recharging back to 100% many times in one day or everyday, you should opt for cell compression. The cycling data with compression at the datasheet exists for a reason, you know.

 

[Mariner62]

Yes. Then the yearly capacity reduction will be at 2-3% (calendar life, whether you use it or not).
If you tend to cycle all the way down below 10% and recharging back to 100% many times in one day or everyday, you should opt for cell compression. The cycling data with compression at the datasheet exists for a reason, you know.

Yeah I do cycle every day, down to 10% (of the latest measured capacity) and most days back up to 100%. I was aware that compression may extend the life ... but I opted to go with the SEPLOS rack which does not really compress the cells much at all.

I don't have any bulging cells at this stage.

 

[ocular]

Yeah I do cycle every day, down to 10% (of the latest measured capacity) and most days back up to 100%. I was aware that compression may extend the life ... but I opted to go with the SEPLOS rack which does not really compress the cells much at all.

I don't have any bulging cells at this stage.

@OP, haven't read the whole thread, I have great confidence in Amy at Luyuan, so I would be surprised if she knowingly sold you B Grade cells. I have only quickly glanced the first 2 pages. At 30C I would never go above 56V, preferably float at 55C. "Top balancing" is not required as the JKBMS does this. What total voltage do you push your pack to?

 

[bigbrovar]

Smartshunt would potentially be more accurate and closer to the true battery voltage versus what the multiplus sees. Multiplus soc depends on comms with victron sccs I believe.

No necessarily. I have the smart shunt but I use the JK Inverter BMS as my source of truth battery monitor. Difference between it and shunt is less than 2% and it has a much accurate voltage reading because I had the voltage calibrated using a precision multimeter.

 

[bigbrovar]

@bigbrovar

I have a similar capacity-loss situation with a lot of similarities to the OP. I purchased two sets of 16x 280Ah cells from Amy Wan (Luyuan) and installed them in Seplos Mason boxes with Seplos BMS. I commissioned the first set in November 2022 and the second set in March 2023.
The first set is showing a capacity loss of 15% after 2 years of operation with a total of 460 cycles. They are part of a Victron ESS system which was cycling daily from 10% to 100% SOC. Target cell voltage for 100% SOC is 3.45V.

The capacity loss was somewhat hidden initially since the SEPLOS BMS averages the SOC of the two battery packs and reports this to the Victron system on CAN. The bad pack that was losing capacity started to drop to lower than 10% while the (still) good pack would stay above 10% by the same margin, resulting in a reported SOC of 10%. It was not until the bad pack started to trigger Low SOC alarms on the BMS that I noticed the problem.

Somewhere along the way the BMS picked up that the capacity had changed and started reporting 240Ah as the "Total Capacity" (i.e. measured) rather than the rated 280Ah. The OP noticed the same thing although he had a different BMS.

I separated the two packs and after a lot of time spent top balancing the bad pack I conducted a full charge/discharge cycle shown in the attached graph. The pack was balanced within 5mV at 3.45V (Pack: 55.2V) and was discharged into normal house loads, varying a bit around 50A discharge.
The discharge knee is encountered when the SOC falls to about 15% and the battery voltage collapses below 3V before SOC falls to 12%.
The pack is recharged with 70-80A and the charge knee is encountered at about 97% SOC. My charging algorithm is designed to ensure no cell rises above 3.6V.

I don't have any cell bulging. My cell temperatures rarely rise above 30degC. I have full records on VRM.

Similarities with OP:

1. Same cells;
2. Same supplier;
3. Same age;
4. Same architecture (two parallel packs);
5. Same capacity loss

I'm not convinced that using more than one SEPLOS battery with a Victron ESS system is a reliable solution without some tweaking. I'm interested to hear of others who have successfully operated such a system for 2+ years.

Your experience aligns very well with what I experienced. Interestingly so does the purchase timeline. I feel we got a bad batch or sold false goods. I paid premium price for the bank that went sideways. $200+ more than what others were offering for Grade A+ match and batch. Interestingly the grade B battery I paired it with (from another seller) is still working without issues. I just cut my losses and moved on.

 

[AntronX]

.

 

[sunshine_eggo]

.

That time of the month?

 

[rustyrotors]

i just saw this thread, after 2 years i have noticed some degradation on a few of my LF304 cells from docan energy. i first noticed a problem when my pack started getting a bit out of balance, due to cell voltage surge confusing the BMS balancer. i then recently took the pack apart for a 48V upgrade and noticed 3 of the cells bulging. i discharged the pack to 3.3V/cell before removing compression. the bulging only happened after compression was removed and cells were sitting out in my garage for a few weeks. i have not done any capacity tests, but just from watching the voltage over the past 2 years, it seems they now cycle about 10-15% more than they did initially. and i can see the weak cells surge/sag more than the others during charge/discharge. a bit dissapointing because ive been so paranoid about overcharging, i dont even use float charging to minimize time at high soc

my system is located outdoors, usually 30-40C in the summer
cells compressed hand tight with threaded rod and end plates
charged 40A peak, 0.13C to 3.5V absorption with no float
discharged 80-100A peak, 0.30C when running microwave
cycled 50-95% daily

im now still running the same cells, but with a bit of a mismatch of eight more LF280k v3 in series for a 48V system. same hand tight compression like before, but i lowered abs charge to 3.45V, and still no floa

 

[S Davis]

i just saw this thread, after 2 years i have noticed some degradation on a few of my LF304 cells from docan energy. i first noticed a problem when my pack started getting a bit out of balance, due to cell voltage surge confusing the BMS balancer. i then recently took the pack apart for a 48V upgrade and noticed 3 of the cells bulging. i discharged the pack to 3.3V/cell before removing compression. the bulging only happened after compression was removed and cells were sitting out in my garage for a few weeks. i have not done any capacity tests, but just from watching the voltage over the past 2 years, it seems they now cycle about 10-15% more than they did initially. and i can see the weak cells surge/sag more than the others during charge/discharge. a bit dissapointing because ive been so paranoid about overcharging, i dont even use float charging to minimize time at high soc

my system is located outdoors, usually 30-40C in the summer
cells compressed hand tight with threaded rod and end plates
charged 40A peak, 0.13C to 3.5V absorption with no float
discharged 80-100A peak, 0.30C when running microwave
cycled 50-95% daily

im now still running the same cells, but with a bit of a mismatch of eight more LF280k v3 in series for a 48V system. same hand tight compression like before, but i lowered abs charge to 3.45V, and still no float

How many cells in the stack?

 

[rustyrotors]

there were 8 cells total

actually i take back what i posted about capacity loss. after posting that yesterday i did a full discharge on the 16 cell pack (8x 304ah + 8x 280ah) after top balancing to measure capacity, and the 304s still have roughly 10% soc, 280s at 0%.. so it appears the 304s have not lost much if any capacity in 2 years. there is just some bulging, no capacity loss

 

[unsubscribed]

50% loss of capacity in less than four years. (!) ReBel Metal Case batteries.

Less than 300 cycles. Lightly used in a TOU scenario, 21 hours float, 3 hours Inverting, with a Mutliplus and nominal charge settings.

(Although knowing what I know now (thanks to @RCinFLA ) I have recently lowered Absorption to 14.2 V.)

Disassembled the worst one. I don't know how much bulging is allowable. But this one looks terrible to me.

The JBD BMS had been re-aligning capacity downward, showing 67% recently. (I didn't know it could do that.)

Phoenix, AZ in a garage, on the floor. Temperature in the garage can reach 100 F/38 C in August.

 

[sunshine_eggo]

50% loss of capacity in less than four years. (!) ReBel Metal Case batteries.

Less than 300 cycles. Lightly used in a TOU scenario, 21 hours float, 3 hours Inverting, with a Mutliplus and nominal charge settings.

(Although knowing what I know now (thanks to @RCinFLA ) I have recently lowered Absorption to 14.2 V.)

Disassembled the worst one. I don't know how much bulging is allowable. But this one looks terrible to me.

The JBD BMS had been re-aligning capacity downward, showing 67% recently. (I didn't know it could do that.)

Phoenix, AZ in a garage, on the floor. Temperature in the garage can reach 100 F/38 C in August.

I would further reduce absorption to 13.8V since you have plenty of time to charge, and you can easily hit 100% @ 13.8V. Float at 13.5V.

How do you limit your garage to 100°F in Phoenix? My garage is often hotter than ambient, and we had an absurd number of days over 100°F this past summer.

What is between your cells?

 

[upnorthandpersonal]

21 hours float,

What voltage were you floating at?

 

[miran90]

i just saw this thread, after 2 years i have noticed some degradation on a few of my LF304 cells from docan energy. i first noticed a problem when my pack started getting a bit out of balance, due to cell voltage surge confusing the BMS balancer. i then recently took the pack apart for a 48V upgrade and noticed 3 of the cells bulging. i discharged the pack to 3.3V/cell before removing compression. the bulging only happened after compression was removed and cells were sitting out in my garage for a few weeks. i have not done any capacity tests, but just from watching the voltage over the past 2 years, it seems they now cycle about 10-15% more than they did initially. and i can see the weak cells surge/sag more than the others during charge/discharge. a bit dissapointing because ive been so paranoid about overcharging, i dont even use float charging to minimize time at high soc

my system is located outdoors, usually 30-40C in the summer
cells compressed hand tight with threaded rod and end plates
charged 40A peak, 0.13C to 3.5V absorption with no float
discharged 80-100A peak, 0.30C when running microwave
cycled 50-95% daily

im now still running the same cells, but with a bit of a mismatch of eight more LF280k v3 in series for a 48V system. same hand tight compression like before, but i lowered abs charge to 3.45V, and still no floa

You need to disconnect charging and let the cells voltage sit also do not keep them at 100%

 

[unsubscribed]

I would further reduce absorption to 13.8V since you have plenty of time to charge, and you can easily hit 100% @ 13.8V. Float at 13.5V.

How do you limit your garage to 100°F in Phoenix? My garage is often hotter than ambient, and we had an absurd number of days over 100°F this past summer.

What is between your cells?

There is some kind of hard (closed-cell?) black foam between the cells. (And on the sides of the battery box - but not at the mid-point.)

I don't actually know the max temp that the batteries reached in the summer. The garage is well insulated and the two car concrete floor has some thermal mass. The first summer that I placed the batteries, I occasionally went to the garage in the late afternoons and BT'd into the batteries to check their temp. I recall seeing maybe 37 C?. I don't recall ever seeing 40 C. After that I stopped checking it. (Nothing that I can do about it anyway.)

I'm thinking that this particular battery is toast. Yes?

I replaced the four ReBel's with six BattleBorns that have been sitting on an Emergency Cart for years. I agree that I should lower the Absorption charge even more. (Although, FWIW, I had set the Multiplus to stay in Absorption for only an hour each cycle.)

 

[unsubscribed]

What voltage were you floating at?

13.6 v.

 

[sunshine_eggo]

There is some kind of hard (closed-cell?) black foam between the cells. (And on the sides of the battery box - but not at the mid-point.)

Just making sure there's something between the cells.

I don't actually know the max temp that the batteries reached in the summer. The garage is well insulated and the two car concrete floor has some thermal mass. The first summer that I placed the batteries, I occasionally went to the garage in the late afternoons and BT'd into the batteries to check their temp. I recall seeing maybe 37 C?. I don't recall ever seeing 40 C. After that I stopped checking it. (Nothing that I can do about it anyway.)

Mine's uninsulated, so maybe that's the difference.

I'm thinking that this particular battery is toast. Yes?

Yep.

I replaced the four ReBel's with six BattleBorns that have been sitting on an Emergency Cart for years. I agree that I should lower the Absorption charge even more. (Although, FWIW, I had set the Multiplus to stay in Absorption for only an hour each cycle.)

Even that may be excessive especially in heat, and the only reason to run longer than 15-30 minutes at 14.4-14.6V is to allow additional balance time.

13.8V for 2 hours should do it. The lower voltage and lower current at peak voltage tend to maximize cycle life, and should presumably reduce charging stress when in a hotter enviroment.

13.6 v.

Too high. There is concern that this can result in low current overcharge, result in lithium plating and reduced capacity. 13.5V will allow the SoC to drop slightly from 100%. Experiment with lower.

Explore any option you might have for opportunistic charging at lower temperatures. WIth a 21 hour float window, it might make sense to only charge during the coolest part of the day. If you have a Cerbo, you can get very creative with the generator auto start feature to determine when and how to "ignore AC input."

 

[unsubscribed]

Tnx for your input. As I think about it more now, the battery that fared the worst was the one closest to an inlet (vent) to outside air. (To allow air to enter when the exhaust fan is on.)

Still, I saw where Will purposely left a LiFePO4 battery out in the hot Las Vega Sun, probably seeing 130 - 140 F internally some days, and it only recently degraded. And even then, it degenerated only to 80%. Whereas I'm at 50%. (And his battery sat unused most of the time too... opening an electric gate.)

I would have set the Multiplus to a shorter absorb time. But, unless I'm missing something, mine can only be set to a minimum of 1 hour. (Latest firmware.)

When we were all buying batteries in 2021, we were told that the BMS would take care of everything. Seems like maybe a BMS should take temperature into account as they used to with Lead Acid? So I didn't sweat voltage settings. (No pun intended. But I'll take it.)

 

[sunshine_eggo]

I would have set the Multiplus to a shorter absorb time. But, unless I'm missing something, mine can only be set to a minimum of 1 hour. (Latest firmware.)

13.8V, 2 hours.

When we were all buying batteries in 2021, we were told that the BMS would take care of everything.

No we weren't.

Seems like maybe a BMS should take temperature into account as they used to with Lead Acid? So I didn't sweat voltage settings. (No pun intended. But I'll take it.)

It does. It terminates charge or discharge based on its settings. BMSs are on/off switches, not regulators.

Lead acid temperature compensation doesn't save lead acid batteries from extreme conditions, it just minimizes damage. A lead acid battery at 100° with proper temp comp will still only last 50% as long as a battery operated at 77°F.

 

[upnorthandpersonal]

13.6 v.

Too high. You can overcharge by keeping it at this voltage. You should be at 3.375V or lower (lower especially if you keep it there for long times). I have a test planned in the lab to show this at some point, but in the past we have seen this issue with telecoms equipment that sits idle and floats at too high voltage.

When we were all buying batteries in 2021, we were told that the BMS would take care of everything. Seems like maybe a BMS should take temperature into account as they used to with Lead Acid?

Temperature is the biggest killer of LFP. I think this has been repeated over and over here on this forum, and is also made clear in the cycle life comparison in the cell datasheet. Put a BMS over temperature charge protection at 35C at the highest. You shouldn't charge above that if you want long cell cycle life.

 

[Koyaanisqatsi]

The thought that stuck in my mind, reading as much as I could of this thread:

I've read a few things lately on the forums that tell me some inverters are backfeeding noise onto the DC bus, which may be at a high voltage, but doesn't show up in monitoring tools. Which, in turn, indicates it's event-triggered and much too short of a duration for voltage monitoring to detect. Something like this could be caused by (large?) loads dropping on the AC side of the inverter (or worse, kicking voltage back at the inverter), and causing a transient that is sent back to the DC inputs. These kinds of voltage spikes could easily cause a pack to swell prematurely. Or in the case of a BMS, burn out the TVS diodes. It would only require a spike like this once every few days to degrade a battery pack in a couple years.

A quality inverter will have large capacitors on the DC input that would absorb such an event. But I wonder about these cheaper all-MOSFET units. This is a great argument for installing large (super)capacitors on the DC bus bars. I was considering this, but the concensus seemed to indicate it's not worth the effort. It is if you have an inverter that's making the DC bus noisy. It should also help with surge capacity on the inverters.

 

[Brucey]

Some place midnite spds on their batteries/power bus, tho traditional thought is the batteries act as a big buffer (perhaps more so with lead acid and less so with complex bms in lfp).

 

[Koyaanisqatsi]

tho traditional thought is the batteries act as a big buffer (perhaps more so with lead acid and less so with complex bms in lfp).

This is valid for lead-acid, but not at all for lithium batteries. On LA batts, the voltage can't do much harm. The lead plates absorb the energy and the voltage causes a brief off-gassing, which is nothing out of the ordinary for LA batteries. In a lithium cell, the effect is similar, but Li cells can't deal with off-gassing very well, and the cell swells up. LA can handle a pretty high voltage and duration surge without any long-term damage because they are just lead plates with liquid around them. Li can't convert the energy to chemistry fast enough and the spike voltage can damage the cells.

I'm thinking it might be prudent for us to start looking at a snubber circuit on the DC bus as a normal thing, with these cheaper inverters on the market. The MS SPDs are probably good - I haven't looked at them. But something needs to be installed that assures the DC bus does not have any noise above around 60V.

This is when I wish I had the money to buy a high-speed voltage meter than can catch such a fast spike and graph it. Would be super-helpful for identifying the cause of the problem.