diy solar

diy solar

Warning of under-capacity batteries purchased on Alibaba

DeeK

New Member
Joined
Nov 30, 2020
Messages
74
I posted this on another location on this site specific to the vendor in question but thought it might be of genera;l interest to anyone currently sourcing batteries. If it's bad form to post in two places, please let me know and I will try to remove one. I was recently made aware of this forum and joined so that I could post my experience with 64 LFP cells I purchased from Simon Lin of Dongguan Huanhuan Energy Ltd.n AliBaba. I hope to spare others from similar disappointment.

I purchased 64 200Ah CALB cells in November 2019, arrived January 6th, 2020. They were connected in parallel and balanced overnight (not charged) and put into service in a 4P16S, 800Ah configuration January 7th.

Cells have been cycled lightly, typical usage cycle is between 85% and 96% SOC. Twice they were cycled down to between 50 and 60% SOC. Charge rate never exceeded 50A and was usually between 35A and 40A (<0.05C). Absorb is 3.45 or less for less than a half hour. The only time they have been above that until the capacity tests was for a one-time, full-bank, parallel top-end balance at 3.65V.

During a recent snowstorm and 3-4 days with limited recharge, they reached their low-end cutoff (3.065V, 49.0V bank). The cells were all within 10mV balance at this time thanks to a 10A active balancer.

At cutoff they had produced 644Ah. There is perhaps 6-7% capacity below 3.065V and the cells began this cycle at 97-98% charged. Therefore, 644Ah should represent roughly 92% of the capacity which indicated conservatively a full capacity around 700Ah, well below 800Ah.

I purchased two ZKETech EBC-A10H testers to evaluate the capacity of my cells. My tests are done at 20 degrees Celsius as follows:


  • Charge cell to 3.65V, cutoff charging when charge current falls to 1A, immediately begin capacity test.
  • Capacity test – 10A discharge (0.05 x 200Ah), 2.6V cutoff
  • Recharge – charge at 0.001C (1.6A) to 3.1V, then 20A charge to normal.
I have now tested 19 of the 64 cells and the mean is 164.72 Ah standard deviation of 4.13. Assuming less than 1% capacity below 2.6V and an initial charge to 100% that means these cells are about 166 Ah cells, 83% of stated capacity. Even adjusting a couple percent for testing below standard temperature (20 vs 25), that’s still significantly below spec.

In hindsight, I should not have purchased these cells until I had the ability to capacity test them. Since the cells are usually cycled so lightly, I was not aware there was an issue until the deep discharge and shutdown. If there is something I am missing or have done incorrectly, please let me know. These cells were advertised as new, 1.5-year old batteries which shouldn’t be a problem for LiFePO4 cells. They were 3.29V when I received them and reasonably consistent across the set. The only explanation I have been able to come up with is that these cells had been used heavily before I received them. I have exchanged quite a few messages with Simon Lin since discovering this and have provided him test data and details but he has offered no explanation or resolution.

Please let me know if there could be something else that is causing them to perform so far below specification.
 
Thank you for going to the effort of actually testing your cells rather than just extrapolating based on assumptions. In addition to an average/stdev, the total range would be useful.

Note that a single hour of absorption at 3.45V is not likely to get you to a high state of charge.

Did you confirm both current and voltage via a second source, such as a clamp meter or shunt for current and a separate voltmeter?

I'm not familiar with the unit, but it appears to use separate voltage sense wires, which is likely to ensure voltage drop through the current carrying leads is a non-issue. For me, cheap testers require secondary verification of accuracy.

Also, while there may be a reluctance to run the cells through their full range, it is necessary to do so to establish their actual performance vs. specification. Assumptions need to be eliminated. The fact that you're well below 0.2C means your results are going to be optimistic, which furter validates your concerns.

Overall, your testing is adequate, and the above are simply proposed refinements. On subsequent tests, I'd try to verify current and voltage values via an alternate source to validate the unit is reporting correlated results.

Based on your description of usage, I wouldn't expect significant degradation beyond 4-5% assuming a full year of daily operation. Additionally, for cells stored at about 50% SoC (based on your as-received voltage), I would not expect more than 1-2% degradation from 1.5 years of storage.

IMHO, if your numbers are accurate, and you've NEVER charged these cells below freezing, these cells almost certainly did not meet specification at the time of purchase. Making allowances for the possible cycle wear and capacity loss due to storage, this deficit was significant. If the cells were covered by a warranty, they are likely eligible for replacement depending on the terms of the warranty.
 
Thank you for going to the effort of actually testing your cells rather than just extrapolating based on assumptions. In addition to an average/stdev, the total range would be useful.

Note that a single hour of absorption at 3.45V is not likely to get you to a high state of charge.

Did you confirm both current and voltage via a second source, such as a clamp meter or shunt for current and a separate voltmeter?

I'm not familiar with the unit, but it appears to use separate voltage sense wires, which is likely to ensure voltage drop through the current carrying leads is a non-issue. For me, cheap testers require secondary verification of accuracy.

Also, while there may be a reluctance to run the cells through their full range, it is necessary to do so to establish their actual performance vs. specification. Assumptions need to be eliminated. The fact that you're well below 0.2C means your results are going to be optimistic, which furter validates your concerns.

Overall, your testing is adequate, and the above are simply proposed refinements. On subsequent tests, I'd try to verify current and voltage values via an alternate source to validate the unit is reporting correlated results.

Based on your description of usage, I wouldn't expect significant degradation beyond 4-5% assuming a full year of daily operation. Additionally, for cells stored at about 50% SoC (based on your as-received voltage), I would not expect more than 1-2% degradation from 1.5 years of storage.

IMHO, if your numbers are accurate, and you've NEVER charged these cells below freezing, these cells almost certainly did not meet specification at the time of purchase. Making allowances for the possible cycle wear and capacity loss due to storage, this deficit was significant. If the cells were covered by a warranty, they are likely eligible for replacement depending on the terms of the warranty.
I greatly appreciate your comments. First of all, yes, I constantly test the tester with two other clamp meters and several volt meters and have verified that the rate of discharge and voltage readings are very accurate. And, yes, it does have separate Voltage and Current leads.
In terms of charging when freezing, no, my system is in a passive-solar, well-insulated shed that I built specific for housing the solar system. It's typically between 60 and 70 even in winter, rarely drops below 50F, in there and I have a thermostatic heater and a radiant oil heater I use occassionally as needed. The cells have never fallen below 45F even for a short time.
I wondered if my cells were being charged to upper 90s% at 3.45V and short absorb but I think the information from the low V shutdown and the individual detailed capacity tests I am running now tend to agree with each other and point to cells that are between 80 and 85% of their stated capacity and also seems to indicate that the cells were at somewhere above 95% before the major discharge event.
I am continuing with the testing until I get all the cells done. My hope is that with true capacity results, I can do a better job of ensuring that each group of 4 cells is as balanced as possible to maximize the energy harvest from this diminished battery. I'll be happy to provide the full set here when I get the completed results and am happy to provide discharge curves as well. In answer to your question about range, so far High Capacity is 172.4Ah and low is 160.44Ah. It will be interesting to see the distribution when done. I am kind of curious if there are two sets of cells that indicate two different histories.
Thanks again
 
"During a recent snowstorm and 3-4 days with limited recharge..."
What is the low voltage cut-off for recharging these batteries, and what was the battery temperature during this time?
EDIT: Answered battery temperature question while I was posting this.

Also, thanks for the comprehensive write-up on these.
 
What is the low voltage cut-off for recharging these batteries, and what was the battery temperature during this time?
EDIT: Answered battery temperature question while I was posting this.

Also, thanks for the comprehensive write-up on these.
Right now, the only low cutoff I have is my inverter and the highest low cutoff it offers is 48.8V (3.05V cell) which is what triggered. There's a small V difference between the inverter and reality though and I noted that the cell voltage was actually 3.065V at the lowest and very consistent across all cells, which was a huge relief! I am in process of rigging my Victron BMV712 to a contactor/relay and will be able to set a slightly higher LVCO of 3.1V/cell.
 
Thank you for going to the effort of actually testing your cells rather than just extrapolating based on assumptions. In addition to an average/stdev, the total range would be useful.

Note that a single hour of absorption at 3.45V is not likely to get you to a high state of charge.

Did you confirm both current and voltage via a second source, such as a clamp meter or shunt for current and a separate voltmeter?

I'm not familiar with the unit, but it appears to use separate voltage sense wires, which is likely to ensure voltage drop through the current carrying leads is a non-issue. For me, cheap testers require secondary verification of accuracy.

Also, while there may be a reluctance to run the cells through their full range, it is necessary to do so to establish their actual performance vs. specification. Assumptions need to be eliminated. The fact that you're well below 0.2C means your results are going to be optimistic, which furter validates your concerns.

Overall, your testing is adequate, and the above are simply proposed refinements. On subsequent tests, I'd try to verify current and voltage values via an alternate source to validate the unit is reporting correlated results.

Based on your description of usage, I wouldn't expect significant degradation beyond 4-5% assuming a full year of daily operation. Additionally, for cells stored at about 50% SoC (based on your as-received voltage), I would not expect more than 1-2% degradation from 1.5 years of storage.

IMHO, if your numbers are accurate, and you've NEVER charged these cells below freezing, these cells almost certainly did not meet specification at the time of purchase. Making allowances for the possible cycle wear and capacity loss due to storage, this deficit was significant. If the cells were covered by a warranty, they are likely eligible for replacement depending on the terms of the warranty.
I thought I'd add that having two identical capacity testers is another good internal check. If one were much different than the other, I should be able to see consistently higher or lower results from one tester which I don't seem to be getting. Also, I had to run an early test over because I botched it right at the end and didn't save the data. I intentionally ran it on the other tester and got the same results to within 0.2Ah.
 
We've done a lot with A grade CALB cells going all the way back to 2010, and they consistently test at 107% - 112% of their rated capacity depending upon the batch. We can extract this charging to 3.45V and allowing current to taper. Pushing to 3.65V gains a gnats hair in capacity and is not really even worth discussing. CALB SE cells (blue case), more so than Winston or Sinopoly, don't seem to like staying at high voltages and can lose some capacity when held at high SoC.. This does seem to level out after the first big loss. That said I think you may have been swindled with used or B grade or reject grade cells, which is not at all unusual these days.
 
Right now, the only low cutoff I have is my inverter and the highest low cutoff it offers is 48.8V (3.05V cell) which is what triggered. There's a small V difference between the inverter and reality though and I noted that the cell voltage was actually 3.065V at the lowest and very consistent across all cells, which was a huge relief! I am in process of rigging my Victron BMV712 to a contactor/relay and will be able to set a slightly higher LVCO of 3.1V/cell.

Concerning LVCO:

While it may be beneficial to think in terms of reducing the depth of discharge via an increased LVCO, is it really necessary?

Personally, I'd rather dip into my last 5-10% of capacity on a very infrequent basis rather than deal with a hard shutdown. If the system is left unattended on a regular basis, or there is a potential for irregular use (visitors that have no clue about limited resources), I can certainly see the value, but the occasional discharge to 2.5V/cell will not appreciably affect the life of the cells.
 
Concerning LVCO:

While it may be beneficial to think in terms of reducing the depth of discharge via an increased LVCO, is it really necessary?

Personally, I'd rather dip into my last 5-10% of capacity on a very infrequent basis rather than deal with a hard shutdown. If the system is left unattended on a regular basis, or there is a potential for irregular use (visitors that have no clue about limited resources), I can certainly see the value, but the occasional discharge to 2.5V/cell will not appreciably affect the life of the cells.
that's great to know. i laughed about your parenthetical statement because that is exactly what happened, house/cat sitter who is a wonderful person but not in tune to living off grid in combo with an snow on panels and unusually long time without sun in normally sunny new mexico; literally the perfect storm. sometime i would like to follow up with you on your earlier comment that 3.45 for an hour was not likely a high state of charge. without benefit of a real capacity tester, i have been trying to determine charge profile that will get me 95-97% capactiy and with my slow charge rates, it's seemed by watching all the indicators i can that my cells are indeed close to full and even starting to get into that dangerous runaway zone if i try to push them much beyond 3.45 or for a long absorb. that also seems to jive with most of the charge curves i've reviewed. is there something specific that makes you conclude that this is not the case? i guess now i could take them through the normal routine then pull offline and do a real capacity test and check for sure. thanks again for your help.
 
that's great to know. i laughed about your parenthetical statement because that is exactly what happened, house/cat sitter who is a wonderful person but not in tune to living off grid in combo with an snow on panels and unusually long time without sun in normally sunny new mexico; literally the perfect storm. sometime i would like to follow up with you on your earlier comment that 3.45 for an hour was not likely a high state of charge. without benefit of a real capacity tester, i have been trying to determine charge profile that will get me 95-97% capactiy and with my slow charge rates, it's seemed by watching all the indicators i can that my cells are indeed close to full and even starting to get into that dangerous runaway zone if i try to push them much beyond 3.45 or for a long absorb. that also seems to jive with most of the charge curves i've reviewed. is there something specific that makes you conclude that this is not the case? i guess now i could take them through the normal routine then pull offline and do a real capacity test and check for sure. thanks again for your help.

I am my wife's power company. I work to make things seem seamless and convenient. She takes daily naps when we're at the off-grid compound, so I make a lot of tweaks while she's napping. Last weekend was a 1.5hr generator run due to a bit of a late arrival (late start on the electric water heater) and deployment of a heated throw that made her so happy. :) Plus, I needed to test how it worked as it was the first time using it to recharge the bank vs. powering the 5th wheel directly.

Looking again at your actual numbers, 800Ah rated capacity, you are likely getting to a higher state of charge given your lower charge current. I didn't have both numbers in my head when I stated that.

I recently charged several 40A CALB cells to 3.4V via a bit more than 0.3C and allowed them to taper to 1A before terminating. The data was logged and I can essentially estimate SoC at 0.XC at 3.4V.

Charges terminating at 0.1C @ 3.4V were 80-88% SoC. .05C terminations at 3.40V were consistently 92-94% SoC. These charges took approximately 6+ hours from empty.

Since your charges are closer to 0.05C, and you are terminating at 3.45V with a 30 minute absorption, you are likely 95%+. I would consider cutting the absorption to 5 minutes or specify a termination current.
 
Does this label look like a legit CALB 200Ah cell? The first characters on all of them are P8215 which makes me wonder if they were supposed to be 215Ah when new.
1607375663121.jpeg
 
Back
Top