WARNING: Autopsy of a Chinese LifePo4 Battery

Zash

Zash

10 mW
Joined
Jun 28, 2015
Messages
20
I bought one of these in 2015:

48V 20AH consisting of 16 A123 style cells in series with a BMS.

Each cell should look like this:

a123%2020ah%20x2.jpg


HTB1GNR.HFXXXXb5XFXXq6xXFXXXP.jpg


Details here:
https://www.aliexpress.com/item/48V...WER-Portable-RC-battery-FREE/32328844192.html

It only ever gave a 4AH charge.

Recently, I opened it up to replace one of the 16s cells which was causing the issue and was shocked to find that:

1. All the cells were rejects; some of them had the bar codes and serial numbers blacked out with permanent marker.
https://i.imgur.com/RPUvRcl.jpg

Here is one without the serials blacked out but it has a new sticker covering the original serial and bar code. You can also see the state of the cell package:
https://i.imgur.com/7wDbVDc.jpg


2. There were punctures in some packs.

3. The specific cell needing replaced had half of the cell completely black and oxidised once I examined it closely by slicing it open.
https://i.imgur.com/2wBV0xu.jpg
https://i.imgur.com/l8XI39o.jpg
https://i.imgur.com/rwDeTBv.jpg







4. None of the terminals are soldered - they have been drilled and have 2 flimsy M3 philips head screws with M3 nuts holding each pair of tabs together, not making great contact. Atrocious.

https://i.imgur.com/yDXsL1v.jpg

https://i.imgur.com/ZWamdPG.jpg


BONUS:
With the replacement cell, I received a plug charger. For some reason, there was no output. After checking the mains and stripping he cable, I decided to open the plug. This is what I found inside:

https://i.imgur.com/JqPNmJS.jpg





It's too late to return the original battery pack; I have obviously been scammed.

A warning to everyone - Expect all 'sealed' batteries from China to be comprised of rejected cells and horrendous workmanship. Do not buy.

Next time, I am toying with the idea of buying 16 cells and building the battery myself after closely examining every A123 style package.

Regards,
-- Zash
Cambridge
UK
 
Sorry mate, your complaint is just a little too vague and incomplete for us to take seriously:

1. Your pack only ever gave 20% of claimed capacity, but you didn't complain?

2. Crossing out the serial number is likely to mean rejected as "A grade", but have you seen the specs of A123 system packs? Nominal IR new is 1.1mOhm. Even B grade A123 packs can be better than the A Grade stuff from other vendors

3. Mistreatment of LiFePo4 can result in the "oxidation" that you described. Leaving packs at low voltages for long times, charging at sub zero temps, etc, can cause separation and plating of metallic lithium. Since it was only one cell that failed, chances are it was of lesser spec than the other 15, but neither you nor we can guess at how was contributed by low quality, and how much by mistreatment (if any).

4. Dual M3 bolts are acceptable for a 30A load. Including washers would be a nice indication of quality, but their exclusions, I don't see as a criticism of quality.

Not saying you didn't get a faulty product. But there's a difference between a faulty product and a scam.

But you are right on one thing: Making your own is the best way to way to ensure quality. I've been making my own packs for a few years now. Every cell I buy gets:

1. A full discharge and recharge cycle to test capacity.
2. A load test at the claimed C rating to test Internal Resistance
3. Left for at least a week at full charge to test for self discharge.

If any cell is any different from the average (better or worse), it gets rejected. So I always buy 10-20% more cells than I actually need, and then make myself some "B-grade" packs for stationary energy storage out of the spares. (My dad has an off grid farm)
 
Those are all great points.

I shall try to go through them in order, hopefully to the satisfaction of most of you :)

I will be opening the battery pack again to balance the new battery cell/install monitoring, once cell monitors arrive from China. I'll take plenty of pictures then.

I'll then also edit the OP for posterity.

1. I was an excited fool, trying to get my DIY electric bike up and running cost effectively. Some cost effective shortcuts worked out great, others caused problems where solutions needed to be evolved over time. For the first summer, I just didn't have the bike working for any extended length of time to be able to test the battery capacity. I suspected a problem, but hoped for the best, thinking the batteries were out of balance. The fool is me. This is the 3rd summer where I can now test the battery extensively with a meter, as the bike is performing well; the range, however, remains unchanged. This indicates that the battery pack was defective upon delivery.

2. The packs actually don't have any identifying logos on them. I'll take a pic. I assumed they would be Grade A packs, A123 or otherwise.

3. There was no mistreatment I can assure you. I am a poor man and treat all my equipment with the utmost respect. In fact, one of the first things I did was add padding to the inside of the battery to dampen the shock exerted if hitting against the case. I thoroughly researched the optimum charging methods, what charge to store over winter etc, all before the purchase. If it's any indication, the battery pack at fault was second to bottom. Presumably, if there was any physical damage then either the top or the bottom packs would be most likely to be affected.

4. Perhaps, if done correctly using a combination of flat + star/spring washers/retaining compound for retention. The problem is that the nuts are not tight, and without washers on either side; the surface contact area of the M3 head and nut isn't adequate for securing 2 foil tabs together. To tighten them puts strain on the paper thin tabs as the nut rotates into security and mangles the tab. The added problem is that a nut will not screw all the way up to the head of an M3 phillips head. This means that even if tight, there is still a gap between the head and the nut. The paper thin tabs at these locations are, then, obviously cannot be compressed against each other. Additionally, one can easily unscrew the nuts with fingers at some locations. The Obvious problem is that vibrations on the road could easily undo these, causing a short etc.

Pics to follow...
 
Hold the battery upside down so if you drop washers and nuts they drop to the ground not inside and in between the battery cells. I bolted my A123 together with aluminum and plastic blocks I use that 90 amps 7000 Watts over a thousand Cycles so far.
 
Hi Zash.

Watt's it called? "store over winter"? (Hehe... from an urban guy that bikes year `round.)

:wink:
L
 
Oh wow. That really is shonky construction. I take it back. That is more likely to be a scam than a faulty product.

I also noticed that the cell with the serial number had a 66Ah in it. I was going to google the whole thing to see if that meant it was originally a 66Ah cell, but was so poor in quality, they "downrated" it 20Ah and it could have been even so poor to as only give 4Ah... But it seems the file has been removed from your Imgur account between clicking on it the first time, and going back for it...

Sorry to hear you went through that, mate.

If you want my advice for a budget but reliable pack, wait for the 16Ah or 20Ah Multistar packs to go on sale, and buy 2 x 6S packs to make a 44v pack (They seem to go on sale at least once a quarter). The cell quality is brilliant for the price. I've got a 36S version running in my motorcycle. I draw up to 130Amps for about 3 seconds (About 8C) and drain it to 3.5v per cell (about 95% flat) once a month, and for the last 9 months, the cells have not needed balancing. Just get yourself a cheap 6S balancer, and check once every couple months to make sure you get the same performance

20Ah: https://hobbyking.com/en_us/multistar-high-capacity-6s-20000mah-multi-rotor-lipo-pack.html?wrh_pdp=3
16Ah: https://hobbyking.com/en_us/multistar-high-capacity-6s-16000mah-multi-rotor-lipo-pack.html?wrh_pdp=3
 
Thanks for the information.

I have now "repaired" the pack with a new cell as best I can, and re-balanced it. So far I have managed 7Ah out of it, with the pack still showing 51.8V.

This links should work now, if not before.

I do not believe a 66Ah cell exists in this size ( 9.5*140*240mm). Likely, it is a random sticker to cover a blemish or another sticker similar to the sticker above it. That was nicely spotted - I didn't notice it. The other 2 stickers seem to show the date of manufacture, as 2012 12 25.

As this pack will be close to my body on my pedal bike raising safety concerns, and the windings/controller being custom set at factory for a 48V pack, I am going to try and salvage as much performance as I can from this LifePO4 pack instead of immediately buying a new pack or switching to LiPo. I concede that I will probably never attain 20AH for the stated 1800 charges, but if I can get half of each figure, I can at least be satisfied that the money wasn't a complete waste. Towards that end, I have ordered a CellLog 8M style monitors to let me know the health of each cell as time passes.

https://www.youtube.com/watch?v=hSCaZBwgpFs

Folk here seem very knowledgeable. I have 4 questions:
I have used a hot melt glue gun to try and seal a punctured cell. Voltmeter shows that it is 0.01V below other cells at 25% discharge.

Question 1: As we know, oxidation takes place when the cell is in direct contact with air i.e. when it is punctured. Does this permanently destroy the cell once a puncture occurs i.e air is a catalyst for oxidation; or will the oxidation stop once the puncture is sealed and the contaminating molecules have been depleted after being utilised in the chemical reaction, leaving a lower capacity but still use-able cell?

Question 2: Perhaps specifically @Sunder: You state that you reject cells which have higher than standard capacity. Why is this? My (perhaps flawed) understanding was that a higher capacity cell in series with other standard capacity cells would not matter much realistically, although the BMS would indeed continually try to balance the pack by draining that particular cell.

Question 3: Do the affordable passive BMS circuits on the likes of aliexpress:
a. continuously and perpetually drain cells to try and equalise the battery pack,
b. or do they stop draining cells once they are within a range of other cells and then stop,
c. or only balance the battery pack while charging and then disengage any bleeding once the charger is unplugged?

Question 4: I have always been interesting in an active (but affordable) BMS which instead of draining the cells to the lowest capacity cell, would charge the lowest capacity cell using power from the higher capacity cells. Is such a system available which is affordable? My searches turn up little information.

I would appreciate the insight.
 
Obviously i cannot see all that you can with this pack, but i have to say that from what i can see, you are wasting your time and effort trying to repair it. Sealing punctured cells is a risky, short term bodge .!
I would second Sunders suggestion and get some of those Multistar packs ..much more reliable and well tested on here by many users.
Maybe you can keep/ reuse that nice battery case as a little compensation
 
If you're not using close to 100% of the cell, a cell with a higher capacity shouldn't matter that much. But if you regularly do, then especially on lower cell count packs, two or three over-sized cell in a 12S battery can allow the other cells to be over-discharged, while tricking the BMS to thinking the cut off voltage hasn't been reached yet.

In really good BMSes, this is not a problem. The one in my eMotorcycle does per cell voltage monitoring, and any one cell going to LVC cuts the entire pack. But the cheapie on my eBike relies on whole pack voltage to cut. If I have one cell at 4.2, and the rest at 3.5v, (Extreme exaggeration), then the BMS will think the pack voltage is still okay.

Cells can continue to work after they have been punctured, but as hillhater said, not really recommended. I'm not too familiar with the chemistry of LiFePo4, as the little I have read, is about the chemical reaction under ideal situations, not under compromised cells.

As for how BMSes work, the majority of them from eBay are simple - if a cell is over the preset limit, turn the energy into heat at the rate specified in the BMS. Slightly better ones will kill the charge circuit while it does that. Even better ones will balance at any voltage, as long as you're over XX%. The top ones which I have read about, but never seen, will take from the highest cell to give to the lowest ones. (as you described). If you do find one at a reasonable cost, let me know.
 
Thanks for the input fellas.

I have now "repaired" the pack with a new cell as best I can, and re-balanced it. So far I have managed 7Ah out of it in a day's worth of cycling (or lack thereof ;) ), and it's still going - more capacity drain testing tomorrow. LiFe cells are intrinsically 'safe' when damaged, so I'm confident in not having much more to lose after trying the seal on the found puncture. Worst case in the short term, I need to replace another cell for £25 from China before calling it a day.

Unfortunately, I see no advantage of spending significantly more money on another pack which is also more dangerous and out of designed spec for my bike, even if it were to give me 50% more capacity. It doesn't make sense on a pedal bike used for leisure where a day's cycling only requires ~7Ah. My original 20Ah decision was taking into account the loss of capacity over many years of use, so foresight has at least not failed me in that particular regard (my calculations were for 10Ah of use).

Indeed, the BMS takes input from each cell and cuts the pack off when a single cell reaches 2.4V, so I guess I should be quite safe potentially having capacity compromised cells in the short term. These BMS's are available for ~$30 for 16S configurations, which I feel is quite affordable.

I did a search after my last post and came across this promising 8S stackable BMS, which actively charges the lower batteries with the higher voltage ones at up to 4A:

https://www.aliexpress.com/item/4-2...lgo_pvid=4898ae89-6786-4d30-8517-d0b897c6955c

Unfortunately, it does not indicate whether it is designed for one-off usage, or whether it can be permanently attached to a pack. If the latter, I would seriously consider this purchase if and when I build a new battery pack.

I'll keep this thread updated with more experience from the battery pack. Hopefully my loss and experience will mean that others will gain.

Kind regards.
 
Zash said:
LiFe cells are intrinsically 'safe' when damaged, so I'm confident in not having much more to lose after trying the seal on the found puncture.
Click to expand...
... You may want to review that opinion with a few Fisker (A123 powered) and BYD owners in China .!

.....Unfortunately, I see no advantage of spending significantly more money on another pack which is also more dangerous and out of designed spec for my bike, even if it were to give me 50% more capacity. It doesn't make sense on a pedal bike used for leisure where a day's cycling only requires ~7Ah. My original 20Ah decision was taking into account the loss of capacity over many years of use, so foresight has at least not failed me in that particular regard (my calculations were for 10Ah of use).....
Click to expand...
All lithium batteries are dangerous including Lifepo4, LiPo, pouch, prismatic, cylindrical, etc etc.
Its the assembly, maintenance, use , and users, that create the danger.
If you only need 10Ah, then only buy that capacity, They are available at lower prices
$200 (or less at sale time) would get you a complete new 12 s 10Ah set.
.......and only weigh 3 kg !....1/3 the weight of your 7Ah unit.
 
Zash said:
I did a search after my last post and came across this promising 8S stackable BMS, which actively charges the lower batteries with the higher voltage ones at up to 4A:

https://www.aliexpress.com/item/4-2...lgo_pvid=4898ae89-6786-4d30-8517-d0b897c6955c

Unfortunately, it does not indicate whether it is designed for one-off usage, or whether it can be permanently attached to a pack. If the latter, I would seriously consider this purchase if and when I build a new battery pack.
Click to expand...

That's a great find! It's a bit cheaper than I expected, but it also isn't a full BMS, as it says in the description - No over or under voltage cut off, and presumably no current limiting.

It says 8-24, so I wonder if you can daisy chain up to 3 of them together - and if you can, why not more than 3? (As you can see in my Sig, my biggest battery is 36S)

It also does say you can leave it on 24 x 7, so you can say it's permanently attached, but as it also says, it's not a BMS, so if you're not manually watching your charge/discharge voltages, it won't give you full protection
 
Zash said:
Question 1: As we know, oxidation takes place when the cell is in direct contact with air i.e. when it is punctured. Does this permanently destroy the cell once a puncture occurs i.e air is a catalyst for oxidation; or will the oxidation stop once the puncture is sealed and the contaminating molecules have been depleted after being utilised in the chemical reaction, leaving a lower capacity but still use-able cell?
Click to expand...
We don't know exactly what happens as it hasnt' been studied in detail here on ES. Would require experimentation and destructive cell-disassembly and chemical checking within the cell layers/etc, with multiple cells over their lifetime, starting with the same puncture and conditions on each one.

At a guess, it will depend on the end products of the primary chemical reactions, and their reactability with the rest of the stuff in the battery, and whether any of the compounds have further reactions of any kind (including combustion) due to current flowing thru them or voltage across them.


Question 2: Perhaps specifically @Sunder: You state that you reject cells which have higher than standard capacity. Why is this?
Click to expand...
Any cell that has one characteristic that is sufficiently different from a "standard" may also have other characteristics that are that far outside the standard. Sufficient deviation could mean a chemical or construction problem in the cell (or batch); the manufacturer might know but we won't. Consequences of that problem could be minor or serious...but it's hard to guarantee something if it isnt' the same as all the other ones.
 
Fiskar Karma waste stream cells are still turning up it seems. Hadn't seen any for a year or so.
 
I dont see anything ..(other than a picture posted by the OP.)...to suggest these are A123 cells.
The specs..1.5 C max discharge ??...are no where near genuine A123 pouch performance
Infact the original product add is from "K-Power" and it claims they make all their own cells....but i dont necessarily believe that either !
More likely just another shonky asian vendor selling reject crap to gulible foreign buyers who they know wont bother with warranty returns. :roll:
 
If it's too good to be true, it probably is.

Here is a discharge graph of my 10ah '10C' lipo pack on about 1 amp of load. It's a 5 cell pack, and was balanced before this test. each color represents one of the 5 cells' voltage.

leadershobby_bigdischarge2.gif


I got stupid and fell for some of this crap myself.
Here is an article i wrote detailing how screwed i got:
http://neptronix.org/wordpress/dont-buy-unknown-no-name-chinese-batteries/
 
Appreciate the input guys. Mostly all very insightful!

The repaired battery reached 10Ah before cutting out. Voltmeter showed that the problem is the newly installed battery pack. As you know, it is a 48V 20Ah 16S pack with a BMS measuring/balancing individual cells. The drain on my pedal bike is ~ 0.2C.

I have a hypothesis regarding why it might not actually be a bad cell. Would someone kindly chime in to either confirm or refute my conjecture:

The following is a rough timeline:

1. The new pack was at 3.25V upon reciept, and was drained down to 2.4V before the BMS cut the power - The rest of the 15 cells were reading 3.3V at this stage.

2. My single cell LiFePO4 charger was DOA, I am in the process of getting a refund/replacement. So to bring up the charge of the single cell to the rest of the 15 cells which were at 3.3V, I used a universal 800mA multi voltage selector plug set to 3V. This was again of shoddy construction - the output voltage was actually at 6V but I paid close attention to ensure that the resting voltage of the charging cell came up to 3.3V and stopped it immediately thereafter (~70% full charge). To the best of my knowledge, the battery voltage never reached above 4.2V (cell breakdown voltage) during charge.

3. Assuming that the pack was now balanced, I put it back together and started test-riding. The recharged cell gave me 4Ah when it dropped down to 2.4V cut-off before the BMS cut the power. The rest of the pack sits at 10Ah total, and retains a charge of 3.25V.

Now; my conjecture is that high 6V charge voltage brought up the cell voltage to 3.3V mentioned in timeline point 2, much quicker than normal, without the voltage actually being an indicator of what the real charge on the battery was, i.e. I falsely believed that the pack was balanced at 3.3V when in fact the cell charged at higher voltage was only masquerading as being at the same charge as the rest of the cells at 3.3V which were previously charged with a proper LiFePO4 charger at appropriate CC/CV voltages.

Further evidence: the current flow into the battery from the plug was ~300mA but the pack was at apparent 70% charge after only ~9 hours. This should only have charged the cell to ~15% of the 20Ah spec; or at max, double that if charging at ~double the voltage (6v).

Is my conjecture potentially correct? Or am I deluding myself into thinking there isn't a problem when in fact there is.

I'll leave the pack to charge and balance over the next few days on a proper LiFePO4 charger and then test the pack again. Fun! :lol: :roll:
 
3.3v on LiFePo4 can be 10% full, or 90% full. If I told you the battery was at 3.3v on this graph (well, actually 3.4v on this graph - probably because they were measuring at the charger, not the cell), could you tell me how full it was?

charge.jpg


Anywhere from 20 to 80% would be a good guess, and you could hardly say anyone was wrong for guessing down to 10%, or up to ~88%.

You need to get the cell to at least 3.5v *resting* voltage to be fairly sure it's full. This is also why many manufacturers say the charge termination should be 3.65, 3.7, 3.8. Older Thunderskies used to recommend charging to 4.2! (This is because unlike LiPo, LiFePo4 has minimal chance of thermal run-away, and does not get substantially shortened by overcharging).
 
Yeah, that is one of the graphs I had studied, but I was more fascinated by this one:
life-discharge-series-h.png


As I described in my previous post, it is indeed the "resting voltage" at the cell that was being checked, not the charging voltage (charge was switched off and the battery left to stabilise for a few minutes), and indeed modern charge indicators do reference voltage of the cell to determine how much charge is remaining in the cell. This is all basic stuff.

It is accepted fairly well that if a pack is charged with a specific charger, then one can track the voltage of all the cells in the pack. This voltage will be a good indicator of the remaining charge. For example, if a cell has 3.3V and 50% remaining charge, then another cell at 3.3v in that pack a that time will also have 50% charge.

Further, if one knows the charge remaining in a cell, and measures its voltage to at least 2 decimal points, one can bring up the resting voltage of another cell to that level using the same charger and process used to charge the first cell, and expect to be around the same charge level. Again, basic stuff.

My situation is a little different.

15 cells have been charged with one charger. 1 cell has been charged with another charger at double the voltage. The question is, does the charging voltage change the indication of remaining capacity?

The above graph shows that indeed yes it does, there is a difference. This would mean that one cannot rely on the voltage of the 15s cells to compare to the single cell to determine charge level because the reference 15s cells were not charged under the same conditions.

The pack is currently under charge by the full pack LiFePO4 charger, and will be tested over the coming days for capacity on the new cell; it will take some time for the BMS to balance such an out of balance cell to the other 15s levels. I am confident that my hypothesis will be held true, and the new cell will function well. The other 15s cells, however... I'm not going to hold my breath. I can currently (no pun intended) get 10Ah out of them, so it would seem that the original purchase will not be a total loss.

I'll report back once the tests have been concluded :wink:


By the way, would you folks be kind enough to recommend a good 20Ah capacity single cell LiFePO4 capable charger? It would be nice if it was affordable.
 
....... The question is, does the charging voltage change the indication of remaining capacity?

The above graph shows that indeed yes it does, there is a difference. ........
Click to expand...
No...that graph is simply showing that cells charged TO a lower voltage will have less Ah to discharge than cells charged TO a higher voltage .....assuming the same RATE of charge (amps) is used in both cases. Which mwans a proper CC/CV charge cycle being used.
Two identical cells, or the same capacity , condition, and age, with the same voltage (precisely measured) will have the same discharge capacity available.
But.... If you go and attach some random charger of unknown spec, then you can expect random results.
Chargers are not rated by Ah capacity, just their max charge amps and voltage setting giving a max power (Watts) rating
Decent chargers will have adjustments for both charge rate and voltage as well ad different charge profiles for different cell chemistry and the number od cells being charged.
There are many $25 chargers that have all of these features, and if you pay more you get better quality and power rating.
Look at some of the charger options from Hobby king on line...their "Reaktor" series are good value.
 
You must log in or register to reply here.

Similar threads

D
How to handle series connection in a moderately large 18650 battery build?
Replies
17
Views
2,452
David01
D
thunderheart
Queen Battery QB26650-2300 LiFePO4 cell's test + comparison with A123 Systems ANR26650M1B
Replies
10
Views
1,659
thunderheart
thunderheart
Back
Top