Multiple cells failures and bad experiences from DIANKAISHOU Battery Store

[o0ragman0o]

Had 1 of 2 my sealed 12V LiFePO4 BMS's die last year, so I had to fork out to begin upgrading to a 48V system. I managed to land 16 320AH cells on last year's Ali Express Black Friday sale at DIANKAISHOU BatteryStore.

The first thing they did was demand shipping fees even though they'd advertise free shipping. There was about a week of arguing with them and holding to the advertised sale before they stopped telling me to cancel the order and agreed to ship it by 'standard shipping'. It was then another couple of weeks before they actually shipped.

When they finally arrived over 2 months later, 1 of the 16 cells was DOA, reading 0.2V. They sent a replacement which took another 5 weeks, during which time my 24V inverter completely shit itself for unknown reasons as I wasn't home at the time and I forked out for a 48V 6.2kW unit instead. I'm waiting for a XR04 48V battery box to put it all in order a month and a half ago.

So, just to get by, I've been running sans BMS for a couple of weeks on <1kW solar......yet I notice last night, the replacement cell has blow up! This battery has never seen full charge or more than a few amps of charge or discharge...so WTF?

That's 2/17 cell failures ~12% failure rate and I'm pretty pissed at them given that it's now 5 months and I'm still without a functional battery for at least another couple of months.

I don't want to change vendor as I want to keep the replacement cells matched as best I can but, seriously how can I trust the seller or any of these other cells???

 

[sunshine_eggo]

Aliexpress is a great way to get ripped off unless you're buying from a manufacturer's store front. Your description confirms you were dealing with a dishonest seller.

A strap around your cells is not at all an acceptable way to restrain them. You need a flat plate to distribute the load over the entire face of the cell and avoid the strap putting a concentrated load on the can edge. I'm not saying it's contributing, but it's important for a quality build.

Lastly, you have no idea what you got from these people. "sticking with them" makes ZERO sense. You have no idea if you'll get the same cells from them with a follow-on order. Did you capacity test 100% of your cells?

Punt. Find a reliable seller and pay what you have to... unless you want to continue on this current path.

 

[o0ragman0o]

Aliexpress is a great way to get ripped off unless you're buying from a manufacturer's store front. Your description confirms you were dealing with a dishonest seller.

A strap around your cells is not at all an acceptable way to restrain them. You need a flat plate to distribute the load over the entire face of the cell and avoid the strap putting a concentrated load on the can edge. I'm not saying it's contributing, but it's important for a quality build.

Lastly, you have no idea what you got from these people. "sticking with them" makes ZERO sense. You have no idea if you'll get the same cells from them with a follow-on order. Did you capacity test 100% of your cells?

Punt. Find a reliable seller and pay what you have to... unless you want to continue on this current path.

The strap is just temporary, just to get by until the battery box arrives, hopefully within the next two weeks. I'm kind of glad I caught it now instead of it blowing out under compression once it was in the box.

I'm completely off grid and entirely dependant upon it and so these were emergency purchases to which I don't really have the means to do better. Nor do I have a capacity tester or way to cycle them over the time frames needed to properly test them.

They were balanced from factory. The only thing I can imagine caused this failure, for my part, was that the cell itself was a replacement, so wan't balanced along with the original cells. However, it did arrive at the same voltage and none of the cells have been used before so it should have still be factory balanced. Even now it's still reading the same voltage as the rest of them so I can't see how it is anything but a factory defect.

Upon physical appearances, they seem genuine Grade A cells and the QR code report a manufacturing date Oct-24 and the failed cell is in the same series but a few 1000 after the other 16 cells. Don't know what kind of factory testing would detect potential out gassing though. Maybe there is such a test and maybe these rejects just get picked up by the secondary market

I notified them of the failure yesterday. They initially communicated but haven't yet responding today

 

[SparkyJJO]

Bet they aren't balanced because LFP has such a flat voltage they could be anywhere between 30% and 70% charged and be all the same voltage.

That said, the balloon one isn't at an excessive voltage so who knows. Suppose it could have simply been used some and discharged a few percent to drop back down to normal, so idk.

 

[fatjay]

The strap is just temporary, just to get by until the battery box arrives, hopefully within the next two weeks. I'm kind of glad I caught it now instead of it blowing out under compression once it was in the box.

I'm afraid that's not how it works. Any cell when charged will expand. Even the first time. You cannot use them without adequate compression. Expansion is a fact, you cannot avoid it. The point of compression is to make the expansion consistent throughout the surface of the cell. I'm afraid it wasn't the cell's fault, but rather the user charging the cells without adequate compression.

 

[Whinny]

"So, just to get by, I've been running sans BMS for a couple of weeks"

I suspect that's what caused the cell failure....it would have failed even if compressed.

 

[AshleyL]

So......what is the model of the useless BMS?
That kind of swelling indicated overcharging way beyond 4v.

 

[o0ragman0o]

I'm afraid that's not how it works. Any cell when charged will expand. Even the first time. You cannot use them without adequate compression. Expansion is a fact, you cannot avoid it. The point of compression is to make the expansion consistent throughout the surface of the cell. I'm afraid it wasn't the cell's fault, but rather the user charging the cells without adequate compression.

Note taken on compression but this isn't just expansion from not being compressed, this is a completely failed cell which has neither been undercharged nor overcharged. Given that it's the only cell in the pack of 16, I don't see how it is user error.

 

[fatjay]

Note taken on compression but this isn't just expansion from not being compressed, this is a completely failed cell which has neither been undercharged nor overcharged. Given that it's the only cell in the pack of 16, I don't see how it is user error.

How do you know it hasn't been overcharged? Without a BMS that's designed to regulate and monitor these things?

 

[o0ragman0o]

"So, just to get by, I've been running sans BMS for a couple of weeks"

I suspect that's what caused the cell failure....it would have failed even if compressed.

Not unless the cells became highly unbalanced, which they weren't and couldn't have in the short time it had been used.

 

[o0ragman0o]

How do you know it hasn't been overcharged? Without a BMS that's designed to regulate and monitor these things?

The invert charge voltages were set down so it wouldn't overchage specifically because I wasn't running a BMS

 

[fatjay]

The invert charge voltages were set down so it wouldn't overchage specifically because I wasn't running a BMS

It's still a 48v pack charging at what, 54v? A cell's increased internal resistance could have caused this without a bms to regulate it.

 

[fatjay]

Was the pack top balanced?

 

[o0ragman0o]

Was the pack top balanced?

I've only got solar and the pack never even reached full charge before this happened. There's nothing about the installation that would explain its failure. LiFePO4 cells are normally pretty bullet proof so the only explanation is factory defect and given that it was a replacement for an original cell that arrived DOA, I have to assume the manufacturer must have a pretty high failure rate

 

[AshleyL]

The invert charge voltages were set down so it wouldn't overchage specifically because I wasn't running a BMS

...........16 cells.
Let say if you set 54v charging voltage which ideally each cell will be 3.375v.
Now let assume if there are 15 cells with 3.3v since you mentioned only 1 cell being replaced. You never mentioned if you properly top balanced this cell to match the rest....so......
15 x 3.3v = 49.5v

Assuming if the new cell has higher state of charge than the rest of 15 cells in usage before this....
54v - 49.5v = 4.5V, that single cell alone might had received up to 4.5V at 54V charge voltage during the end of pack charging voltage

Now you see why a BMS is important?

 

[740GLE]

"So, just to get by, I've been running sans BMS for a couple of weeks"

I suspect that's what caused the cell failure....it would have failed even if compressed.

Operator error strikes again!!

Blind over confidence does more harm than good.

 

[740GLE]

Note taken on compression but this isn't just expansion from not being compressed, this is a completely failed cell which has neither been undercharged nor overcharged. Given that it's the only cell in the pack of 16, I don't see how it is user error.

I feel for your current situation.

Right now you only know one thing for certain, the cell turned into a pillow and is no good.

You were operating the cells without a vital protection/safety system in place, at this point it’s a toss up if it was actually a bad cell or poor connection, but it is clear that cell saw voltage higher than it was ever designed to see.

What was the reason for buying from this vendor? Knowing they shipped you a dead cell make me question the rest of the cells quality, to then run the cells without a BMS is playing with fire in the powder room and being surprised when things go south.

 

[sunshine_eggo]

The strap is just temporary, just to get by until the battery box arrives, hopefully within the next two weeks. I'm kind of glad I caught it now instead of it blowing out under compression once it was in the box.

Just noticed... I'm not seeing any BMS leads on your picture. Did you attempt to charge this battery without a BMS?

If yes, I leave room for the possibility that you over-charged the cell(s) and are the the cause of the failure. You absolutely, positively do not run LFP without a BMS. Period. No exceptions.

I'm completely off grid and entirely dependant upon it and so these were emergency purchases to which I don't really have the means to do better.

I get that, but in this field, when you go for the bottom of the barrel, you tend to spend time in the barrel.

If part of your approach included not using a BMS, you have shot yourself in the face with a 12 gauge.

Nor do I have a capacity tester or way to cycle them over the time frames needed to properly test them.

There's functional testing and overkill. A single capacity test is typically a sufficient minimum.

They were balanced from factory.

99.999999% certain this is a false statement. NO factory balances them. Period. Yes. Many will tell you they do, but they all lie unless you're paying a notable premium for the cells.

The only thing I can imagine caused this failure, for my part, was that the cell itself was a replacement, so wan't balanced along with the original cells.However, it did arrive at the same voltage and none of the cells have been used before so it should have still be factory balanced. Even now it's still reading the same voltage as the rest of them so I can't see how it is anything but a factory defect.

Unfortunately, you may have used or cast-off cells that don't meet spec. Unless of course this is your fault because you're not using a BMS.

Upon physical appearances, they seem genuine Grade A cells and the QR code report a manufacturing date Oct-24 and the failed cell is in the same series but a few 1000 after the other 16 cells. Don't know what kind of factory testing would detect potential out gassing though. Maybe there is such a test and maybe these rejects just get picked up by the secondary market

So you've peeled off the black cover and confirmed there's no evidence of an obliterated QR code? It costs them pennies to put on a new QR code.

I notified them of the failure yesterday. They initially communicated but haven't yet responding today

Who is the seller?

 

[HRTKD]

I'm not sure if this needs to be moved into the "Up In Smoke" or "Vendor Review Corner" areas. I'm leaning towards "Up In Smoke".

 

[740GLE]

Meah id leave it as an example for someone looking at DIY batteries and an example why a BMS is needed.

 

[o0ragman0o]

I'm not sure if this needs to be moved into the "Up In Smoke" or "Vendor Review Corner" areas. I'm leaning towards "Up In Smoke".

I can find only one instance in the inverter data which records the pack at 55V, and I recall switching off the panels to draw it down. All other data show the pack not exceeding 54V and never below 53V.
The total PV power generated from the panels over the life of the pack is only 12kWh, less than the 16kWh of the pack itself.
Battery capacity has been stable at 80%
Load never reached 200W
Battery charge current never exceeded 18Amps
The failed cell was measured at (and still is) 3.3V as were all the other cells when I realised the failure. It was night time with light load, so may have relaxed it's voltage but has remained at that since disconnection

Given this was a replacement cell, there is only one event I can identify from the data in which it could have overcharged, but only if it was received from the factory with a significantly higher charge than the rest of the cells. This might be possible as the serial shows it younger by a few 1000 cells and may not have self discharged as much as the rest of the pack before use

So, to run some numbers.....
The failed cell would have needed to be charged >4.2V for such a failure.
55V (max seen in data) - 4.2 = 50.8V for remaining cells
50.8 / 15 = 3.39V (average for remaining cells)

None of the cells were measured >3.39V after the failure meaning the failed cell would have had to be much greater than 4.2V.

So, an overcharge event is 'possible' but I'm still not that convinced.
The cell's voltage is still stable (3.3V) and comparable to the rest. So there appears no significant self discharge that would drop it from the excessive overcharged voltage that would have been needed to fail, so it should still read higher than the other cells if overcharge was the case

The 'no compression' can be ruled out as compression wouldn't have stopped this kind of failure and likely blown the vent instead. In my understanding, compression is there to limit electrode expansion and degradation over time and charge cycles, not contain gas generating failures such as this.

The history is that the cell was a replacement for a DOA cell that arrived with only 0.22V SoC. The seller claims they are tested at factory before shipping so the discharge likely happened in the >2 months it took to arrive. That's obviously indicative of internal short and points to a manufacturing failure of the electrodes and raising the likelihood of the replacement also being a factory defect.


Current state of saga is that once I posted seller an image of the cell and voltage they responded with some confusing back and forth saying they already sent a replacement but when I asked for clarification of this failure or the previous one, they asked if I want them to replace all 16. Things got very confusing and their only response was I need to 'be patient'. When pressed again for clarification of whether they sent another replacement, they called me anxious, but eventually said they have sent a replacement. I've asked numerous times for a tracking number which they still haven't provided. I think this behaviour goes beyond normal 'lost in translation' and they are simply poor communicators

The seller is DIANKAISHOU BATTERY Store on AliExpress
There is also a DIANKAISHOU Factory Store which has the same seller ID. I've spoken with them also trying to source the same model cell but they kept saying theirs were different, "not compatible", even though the QR code they sent showed the same factory, model and all. Again things got confusing.

The failed cell:
LifFePO4 320Ah
QR code: OHACBHALBOOL3DEA80026864

I've tried it against Global Power QR decoder but it doesn't know actual factory.

 

[AshleyL]

I can find only one instance in the inverter data which records the pack at 55V, and I recall switching off the panels to draw it down. All other data show the pack not exceeding 54V and never below 53V.
The total PV power generated from the panels over the life of the pack is only 12kWh, less than the 16kWh of the pack itself.
Battery capacity has been stable at 80%
Load never reached 200W
Battery charge current never exceeded 18Amps
The failed cell was measured at (and still is) 3.3V as were all the other cells when I realised the failure. It was night time with light load, so may have relaxed it's voltage but has remained at that since disconnection

Cause LFP cell voltage quickly settle down to 3.2V or 3.3V from maximum safe 3.65v once the charging stopped.

Do you ever take a look on LFP charge and discharge voltage curve? I really hope you did.

Between 10% to 95% state of charge, LFP cell will maintain a consistent nominal 3.2V ( or 3.3V depending on the multimeter accuracy ).

I am not sure where you get that 80% capacity either. This is LFP, not lead acid. You cannot determine the state of charge via voltage for LFP cell.
If you calculated that 80% based on the 12kwh/16kwh, you are in deep trouble. You cannot be confident/certain that you actually get 314Ah cells in the first place. There were cases where scammers re-lasered 280Ah cell to be sold as 314Ah. Plus you don't know the actual state of charge of all those cells when you received it.

Plus, do you know that lithium based battery cannot be shipped with state of charge exceeding 30%?

https://www.dhl.com/discover/en-global/logistics-advice/import-export-advice/how-to-ship-batteries

Where do you get that 80%? Seriously?
To begin with, not all real 314Ah cells has actual 314Ah capacity.

There is no such thing as "come balanced from factory".

Remember, there is no grade A or B nonsense. There is only cell that passed certification/inspection/verification at the factory or rejected cells. These factories do not sell passed cell to the public, these cells are sold for EV manufacturers. So, what those factories do with rejected cells? They auction them to earn back some of the production cost. After all, these rejected cells might be usable for home energy storage. Understand that these rejected cells might have dangerous defects as well.

Majority of those dealers/sellers at aliexpress or alibaba are re-selling rejected cells that they got from the auction.

Given this was a replacement cell, there is only one event I can identify from the data in which it could have overcharged, but only if it was received from the factory with a significantly higher charge than the rest of the cells. This might be possible as the serial shows it younger by a few 1000 cells and may not have self discharged as much as the rest of the pack before use

So, to run some numbers.....
The failed cell would have needed to be charged >4.2V for such a failure.
55V (max seen in data) - 4.2 = 50.8V for remaining cells
50.8 / 15 = 3.39V (average for remaining cells)

None of the cells were measured >3.39V after the failure meaning the failed cell would have had to be much greater than 4.2V.

So, an overcharge event is 'possible' but I'm still not that convinced.
The cell's voltage is still stable (3.3V) and comparable to the rest. So there appears no significant self discharge that would drop it from the excessive overcharged voltage that would have been needed to fail, so it should still read higher than the other cells if overcharge was the case

It is an overcharged situation. The earlier you accept it, the better. Next time, get a battery management system (BMS)

 

[740GLE]

Spending that much on cells (even from a discount seller) and being so excited that you have to assemble the battery without waiting for the BMS to arrive is a tough pill to swallow.

Again a BMS won’t fix a bad cell from the factory but it’d absolutely pinpoint the troubled cell before it turned into a pillow.

Thanks for posting the sellers name, I can’t say it’s been mentioned before, but alas there are so many resellers in AliExpress they probably change their name more often than I change my underwear.

 

[sunshine_eggo]

I can find only one instance in the inverter data which records the pack at 55V, and I recall switching off the panels to draw it down. All other data show the pack not exceeding 54V and never below 53V.

55V is more than enough to over volt a cell without BMS protection.

The total PV power generated from the panels over the life of the pack is only 12kWh, less than the 16kWh of the pack itself.
Battery capacity has been stable at 80%
Load never reached 200W
Battery charge current never exceeded 18Amps

The above are all irrelevant.

The failed cell was measured at (and still is) 3.3V as were all the other cells when I realised the failure. It was night time with light load, so may have relaxed it's voltage but has remained at that since disconnection

If you overcharged/over-voltage it, and then placed a load on it, the voltage would be identical to the other cells. This is how LFP works.

Given this was a replacement cell, there is only one event I can identify from the data in which it could have overcharged, but only if it was received from the factory with a significantly higher charge than the rest of the cells. This might be possible as the serial shows it younger by a few 1000 cells and may not have self discharged as much as the rest of the pack before use

It's not uncommon to see significant variations of cells from the same "manufacturer". Even just a few % difference in SoC is enough to over-volt a cell.

So, to run some numbers.....
The failed cell would have needed to be charged >4.2V for such a failure.
55V (max seen in data) - 4.2 = 50.8V for remaining cells
50.8 / 15 = 3.39V (average for remaining cells)

Very realistic scenario above. Not at all odd for an imbalanced and unmanaged system to very easily attain those values.

None of the cells were measured >3.39V after the failure meaning the failed cell would have had to be much greater than 4.2V.

Completely incorrect conclusion. You have indicated you loaded after charging. Loading would nearly immediately pull all cells below 3.40V.

LFP charging is similar to lead acid in that you charge to an elevated voltage, and then voltage settles. Once loaded, the voltages drop to below 3.40V very very very quickly. LFP is NOT like 3.7V lithium chemistry like NCA or NMC. Those cells have a very high voltage to SoC correlation, and they experience only minimal settling when low current charge is removed.

So, an overcharge event is 'possible' but I'm still not that convinced.

No data presented support the conclusion that it didn't occur.

The cell's voltage is still stable (3.3V) and comparable to the rest. So there appears no significant self discharge that would drop it from the excessive overcharged voltage that would have been needed to fail, so it should still read higher than the other cells if overcharge was the case

The above conclusion is not correct. The overcharge releases gases, which swell the case. The electrolyte is damaged, but the cell does not become completely non-functional. Without a cell vent event, I would not expect resting voltage to be significantly different than non-swelled cells, but reliability/performance is compromised.

The 'no compression' can be ruled out as compression wouldn't have stopped this kind of failure and likely blown the vent instead. In my understanding, compression is there to limit electrode expansion and degradation over time and charge cycles, not contain gas generating failures such as this.

You are correct that compression would not have saved a cell subjected to overcharge.

The history is that the cell was a replacement for a DOA cell that arrived with only 0.22V SoC. The seller claims they are tested at factory before shipping so the discharge likely happened in the >2 months it took to arrive. That's obviously indicative of internal short and points to a manufacturing failure of the electrodes and raising the likelihood of the replacement also being a factory defect.

The above may be true, but since you are running without a BMS, you have no proof either way. For whatever reason, you chose to do it completely wrong, and you want to assign blame to everything but your bad decision(s).

Current state of saga is that once I posted seller an image of the cell and voltage they responded with some confusing back and forth saying they already sent a replacement but when I asked for clarification of this failure or the previous one, they asked if I want them to replace all 16. Things got very confusing and their only response was I need to 'be patient'. When pressed again for clarification of whether they sent another replacement, they called me anxious, but eventually said they have sent a replacement. I've asked numerous times for a tracking number which they still haven't provided. I think this behaviour goes beyond normal 'lost in translation' and they are simply poor communicators

The seller is DIANKAISHOU BATTERY Store on AliExpress
There is also a DIANKAISHOU Factory Store which has the same seller ID. I've spoken with them also trying to source the same model cell but they kept saying theirs were different, "not compatible", even though the QR code they sent showed the same factory, model and all. Again things got confusing.

The failed cell:
LifFePO4 320Ah
QR code: OHACBHALBOOL3DEA80026864

I've tried it against Global Power QR decoder but it doesn't know actual factory.

You cheaped out on a garbage factory/seller, and you chose to use questionable cells with NO BMS. There aren't many major mistakes left for you to make. This is what setting your money on fire looks like. Sometimes doing it cheap is just a lesson on how not to do it.

Good luck.

 

[o0ragman0o]

55V is more than enough to over volt a cell without BMS protection.



The above are all irrelevant.



If you overcharged/over-voltage it, and then placed a load on it, the voltage would be identical to the other cells. This is how LFP works.



It's not uncommon to see significant variations of cells from the same "manufacturer". Even just a few % difference in SoC is enough to over-volt a cell.



Very realistic scenario above. Not at all odd for an imbalanced and unmanaged system to very easily attain those values.



Completely incorrect conclusion. You have indicated you loaded after charging. Loading would nearly immediately pull all cells below 3.40V.

LFP charging is similar to lead acid in that you charge to an elevated voltage, and then voltage settles. Once loaded, the voltages drop to below 3.40V very very very quickly. LFP is NOT like 3.7V lithium chemistry like NCA or NMC. Those cells have a very high voltage to SoC correlation, and they experience only minimal settling when low current charge is removed.



No data presented support the conclusion that it didn't occur.



The above conclusion is not correct. The overcharge releases gases, which swell the case. The electrolyte is damaged, but the cell does not become completely non-functional. Without a cell vent event, I would not expect resting voltage to be significantly different than non-swelled cells, but reliability/performance is compromised.



You are correct that compression would not have saved a cell subjected to overcharge.



The above may be true, but since you are running without a BMS, you have no proof either way. For whatever reason, you chose to do it completely wrong, and you want to assign blame to everything but your bad decision(s).



You cheaped out on a garbage factory/seller, and you chose to use questionable cells with NO BMS. There aren't many major mistakes left for you to make. This is what setting your money on fire looks like. Sometimes doing it cheap is just a lesson on how not to do it.

Good luck.

Thanks for the point responses. I'll accept this as the best and most likely explaination of the failure. What you've said has been pretty convincing and I appreciate the time and consideration

For those asking about where I got the 80% SoC number, it's taken from the inverter which is set to LeFePO4 so I figure it has value more than worthless.

As for people saying I made a bad 'choice'. I'm not in a position to make any choices at all. I live remote with very limited income and am entirely dependant upon an off-grid system. My last batteries failed last november and my last inverter died in February. My 'choices' have been more a case of survival. To save up for better, which I had been doing, to wait months and months, means I'm simply without any power at all. I wasn't 'cheaping out,' it was more than I could afford.
I know quite well that free balling a battery like this isn't ideal. However fate and failures have forced my hand and there's little else I can do but struggle on.