Mysteriously premature pack cutout - Diagnosis?

[swbluto]

Here are the relevant(I hope) details:

24-48V 20-amp analog pedal-first controller
Crystalyte 407 rear wheel
48V 10AH ping 1.0 LiFePO4
Battery's BMS temperature <110 degrees Fahrenheit
Battery's temperature <100 Fahrenheit
Ambient temperature ~= 89 degrees Fahrenheit
"Cycle Analyst" shows that 7-7.1 AH have been used.

I'm double-checking the last figure by clocking the 2-amp charger's charging time. It seems to be pretty predictable when it's going to be done when I know the amount used according to the Cycle analyst. The 7.1 AH was used over three separate rides over a period of a week, so I'm wondering if I reset the Analyst accidentally, thus my double-checking but I doubt I reset it.

The history of the Battery pack is that it was actually recently purchased used from an Endless Sphere member who purportedly used it as part of a parallel 2 48V 10AH ping lifepo4 set-up(20 AH, effectively) for 1500 thousand miles on a 40amp controller in a slightly hilly area so it's guessed to have been charged approximately 50 cycles, according to him. His posts here seem to corroborate the mileage, so I don't doubt it.

Looking at Cycle analyst to check what voltage/amp conditions it was cutting out at, I found that A-Max varied between 21-38 amps with V-min varying between 43.1-43.7v as I tested it 4 separate times to cutout(I had to manually plug/unplug the battery to reset it.).

Does the voltage/amps shown correspond to whatever ping's voltage cutout and/or current cutout is? If not, what might be the problem?

 

[chessir]

Here are the relevant(I hope) details:

24-48V analog pedal-first controller
Crystalyte 407 rear wheel
10AH ping 1.0 LiFePO4
Battery's BMS temperature <110 degrees Fahrenheit
Battery's temperature <100 Fahrenheit



I happened to notice all my temperatures going up as soon as I connected my Doc Watson between the batteries and the controller.Before I connected the monitoring unit I never had any temperature elavations in the wiring or the batteries.
Just thought I would mention this observation.
I wondered what happened to your walmart bike? I happenend to have a 07 model which quit on me too. I am still trying to figure out what the problem is. Motor and batteries are OK . I replaced the controller but still no luck. Are your walmart bike problems similar?

 

[pwbset]

The 7.1 AH was used over three separate rides over a period of a week

I have a pack that should, in theory, provide me 6ah as advertised. I've been averaging 4.6-4.8ah actual useable capacity before LVC cutoff, but I ride it very hard each morning to climb the 1,600ft I go up to get to work (avg 35-45wh/mi). Do you ride hard? That could effect your useable capacity by quite a bit... or so I've noticed.

 

[swbluto]

I'm usually using between 10-15 amps(1-1.5C) with short periods of 15-20 amps(1.5C-2C) with microshort periods of 20-30 amps(2-3) that the controller reins in during starts and going uphill with an average of somewhere between 15-20Wh/mile. I suppose I'm riding it hard but I don't understand how that'd affect LiFePO4's usable capacity by that much(reduce its usable capacity from 10 to 7 being a 30% drop in capacity?). If "riding it hard" reduced the usable amount, then wouldn't this be Lithium's version of the "Peukert effect"?

This battery type(48V10AH ping 1.0) was actually tested to have a lifespan of 10.5 AH when purchased new(By someone on here according to some post) and it seems the lifespan increases by <10-15% during the first 100-200 uses as it seems lithium batteries do, so it'd seem surprising to have a 4AH drop from ~11AH(expected) to 7Ah.

 

[swbluto]

I wondered what happened to your walmart bike? I happenend to have a 07 model which quit on me too. I am still trying to figure out what the problem is. Motor and batteries are OK . I replaced the controller but still no luck. Are your walmart bike problems similar?

Ahh. The walmart bike was a used regular dual suspension(MGX-DXR mongoose) that I put a Crystalyte 407 Rear wheel on. The right pedal stripped the threads off the crank and that's when it "fell apart". I didn't really want to repair it since upgrading/replacing parts would've seemed to be as costly as simply buying a new, better one in the long-run.

I did, though, have an Electric S-500 scooter, also produced by Currie Technologies. It mysteriously quit on me but I just returned it to Wal-mart for a refund.

 

[swbluto]

It's confirmed. The battery took about 3 hours and 40 minutes which predicts a discharged amount of 7.3333 amp-hours if 2 amps was the actual current absorbed by the battery from the 2-amp charger. Accounting for resistive losses, this seems to confirm the 7.1 AH discharged amount. Anyways, can anyone help? Is there a better forum category to post this? I'm sure there's someone who'd have some idea from the information I've noted. Perhaps there's an unbalanced cell and the whole pack is cutting out due to that one low cell(which might be inferred by the possibility 43 volts isn't the cutoff of a 48 volt ping battery, but I don't know what the LVC is.). Though, if that we're the problem, how would it be fixed?

 

[dogman dan]

A similar problem was reported which turned out to be a disconnected bms wire in the pack, so If you haven't looked, take a peek to see if everything seems connected up. The other thing it could be is a charger that shuts off early. Do you unplug it right away? or leave it hooked up till morning? If possible try charging it with another charger, even an sla one and see what happens. For you to get less than 10 ah out of the battery per charge would be normal, but not a lot less, like with sla's. The previous owners use shouldn't have toasted it yet, though it's possible. Any clue how many amps his ride drew? My Ping 1.0 has now gone about 1200 miles and is showing no signs of less range, but I don't have a huge motor or controller. Undercharging makes the most sense to me. It's not as obvious when it happens with lifpo4 as it would be with sla's.

 

[swbluto]

I'll try slicing open the yellow "shrink wrap"(or whatever it is) to see if there's anything that's obviously loose with the BMS's connections there. Anyways, it seems that undercharging would make sense since it seems likely the cut-off voltage is somewhere around 43-44 volts according to various sources for LiFePO4 in general which is where my minimum voltage seemed to be hanging around but, then again, it seems like my charging voltage after settling is around 53-54 volts so that seems about right?? I'll also try measuring the leads on the BMS to measure the strings' individual voltages after completely discharging later tonight and after charging, assuming this problem recurs. Is there anything else I should check? I'm thinking the issue might not become obvious to me with what I'm planning on checking and my lack of experience with dissecting battery packs and battery packs in general.

I don't really know what the last guy was running at, but I'm inferring he periodically took it to about 2-2.5C considering the hills in his area and the 40-amp controller+20AH_combo but considering he had a geared BMC Hub from EVtech, it seems likely that the motor didn't draw above 1C while "coasting" near top speed(or any constant speed under) which probably made up a large part of the battery's life thus far considering the <20,000 population of his town(a small town) and what I infer to be an average 15 mile ride from his guesstimations. Although, I think the controller might actually be the 35 amp one since he calls it a "36-72V 40A c'lite" controller and it seems a 35 amp one is the closest one that ebikes.ca stocks. Also, I calculated 13 wh/mi. for a 52 mile ride from one of his posts here which would seem to suggest he didn't really use it "heavily", even more so that it was a 20ah combination. Also, he later stated he used 18.9 AH from the total trip. Assuming the worst, than that'd suggest this battery had at least 8 Ahs(More likely 9-9.5AH) in his use.

Lemme think. Is this pack made with 2 5-ah 48V strings? Then if one is disconnected, I should expect around 5-6 AH, right? If it then is a disconnection between one of the cells in the strings, however that'd work out, then I might expect an abnormally low voltage for the amount of energy drawn... so I guess that might make sense.

 

[dogman dan]

Look at the ends of the bms wires in the battery itself too . Its pretty easy to untape just the top, and remove a cardboard insulator. 2.5 c is a pretty good hit on a ping. They will be much happier at 1 c. If the previous owner had a big motor and ran it without a bms, that would explain a lot, but that wasn't the case. For all I know, maybe I get less than I think out of my ping too, I just know it goes far enough, and hasn't degraded from what it did new. My 36v 20 ah takes about 8-10 hours to charge with a 2 amp charger when it's about 3/4 drained. One thing though, I allways harp on and on and on about buy the 20 ah size. I think the 10 ah is too small for anything bigger than an Izip or a 300 watt scooter. I think the 10 ah batteries will die earlier than the 20's when run on similar bikes. I expect I pull 2c some too climbing hills, but when on flatish ground, much less. But still, If your battery is really down to 70% that is hundreds of cycles earlier than expected.

 

[swbluto]

Yeah, it's much less than I expected as well.

I just measured the voltages from the wires. I really couldn't tell how exactly the individual cells are set-up, but I saw nifty individual wires coming from the battery pack with one big negative wire so I measured the voltage between the black wire and the individual small wires as shown in the picture.

Here are the tabulated results for the voltage readings. The first enumerated column contains the voltage difference between wire #(Whatever the row is) and the black wire that's apparently the negative wire. The second enumerated column contains the voltage difference between wire and the wire previous to it(Which I'm guessing are "cell units", 10AH ones: Probably 2P1S for each wire with what he seems to use are 5Ah3V cells.): so, I'd hastily guess the second column with numbers is the voltage of that "cell unit" from that wire #.

It seems like the 10th row/10th wire shows 3.05 volts which is about .2 volts below the average which, I'm guessing, is the problem. What do you think would cause that? Anyways, at this point, it doesn't seem like there are any obviously broken wires or connections. I haven't busted into the pouches yet as I'm trying to hold that off as much as possible given what complications that I think could arise from pulling duct-tape off whatever it's attached to inside.

Right now the battery is charging and I'll come back with those measurements later, maybe 4 hours or so(probably tomorrow).

EDIT: After an hour of charging, I checked all the voltages between wire (N) and wire (N-1) while the battery was charging and they all seem to be at 3.36 volts. FURTHER EDIT: After completing charging, the battery oscillated between "charging" and "done" with the charger, with the charging typically lasting about a minute and the time between different chargings was about 5-20 seconds but after 5-6 chargings, it seems like the charging time decreased to 15 seconds and I suspect it'll keep decreasing(I don't know if this is normal). It also seemed that my DC voltage readings from my voltometer were hay-wire so I'll try again later.

Ok, I just got done with my testing. I tested the individual wires just as I did before and here are the spreadsheets. The first image is was taken right after it was charged and the voltages weren't jumping around.

And here are the voltages after it was drained about .05 AH in an effort to remove the interface/surface/superficial charge that seems to come right after charging. The later voltages where the numerical precision reduces the numbers down to 3.3 and 3.4 are actually 3.36 for every single one as I actually tested the wires together.

So, any clue what's going on? It seems that the "cell unit" on wire #10 is draining more quickly(or it just has a lower capacity than the others during discharge which might suggest that one of its 5ahc cells in parallel disconnected/shorted-out/"fried-maybe" somehow). Might that be a faulty BMS, cell, or ??? I guess what I'm asking is where should I look next?

 

[dogman dan]

The oscilating is the bms balancing the cells, by discharging the high ones, and trying again. If you weren't letting it do this, there's your problem. Leave it plugged in all night for a few cycles. Later on, you still won't get 10 full ah out of the battery, but should get a lot closer in the future. Running till cutout will be more likely to get the battery out of balance than shallow discharges in my opinion.

 

[swbluto]

I guess that might make sense. I'll leave it charged in for another 9 hours or so(HOPEFULLY, that'll leave it as balanced as possible) and test the capacity.

But, I'm having the suspicion the battery attached to wire #10 has one of the paralleled cells busted off somehow(I think there is two cells paralleled per wire) leaving that particular cell group at 5AH nominal. Since the neighbors are still operating at 10AH and there's enough time to balance between the lower-voltage-in-the-ride 5AH cell group and the surrounding 10AHs, that particular cell group should last a little longer than 5AH.... like something like 7AH(You'll notice the voltages around wire#10 in the discharged-table is slightly depressed compared to other voltages.). If I left less time in balancing(Such as I used the battery all in one shot), I suspect the usable capacity would be correspondingly less. But, I'm assuming that each cell group has its own little LVC which I don't know about.

 

[dogman dan]

That would also make a big difference. Or a cell could have just gone bad if it got far enough out of balance. When my pack was new, it needed to balance every charge, and then it stopped and stayed balanced enough since then. So it may take a few cycles to get fully in balance and stay that way.

 

[swbluto]

That would also make a big difference. Or a cell could have just gone bad if it got far enough out of balance. When my pack was new, it needed to balance every charge, and then it stopped and stayed balanced enough since then. So it may take a few cycles to get fully in balance and stay that way.

I'm starting to think a cell has gone bad as I've *gasp* delved into the pack by ripping off the duct-pack and there doesn't seem to be any obviously broken wires. But, it seems there's a soldered bridge between the two-halves of the pack which would seem to deter me from breaking the cell packs apart to fully investigate each cell, so I don't know at this point and until I figure out some other way to investigate each cell...

But, yes, I've pinpointed the two cells that I think are suspicious. But, thinking that either the white or grey wire's Solder-green-board part might be detached from one of two cells, I noticed that the left cell near the white wire looks unusually recessed. I'll check that out.

 

[swbluto]

After looking at the unusually recessed cell on the white wire(the left suspect cell in the image), here's what I found.

It looks like that cell is dangling onto the white wire. So... I'm guessing that's the problem. Assuming it's disconnected or nearly disconnected, how would I solve it? Do I have to de-solder the solder bridge and rip apart that part of pack and then just solder that seemingly loose connection back(however that could happen)? I don't really want to mess up the pack.

Edit: Nevermind. Apparently that "dangler" is really a foil tab that's soldered onto the green pcb bridge. I checked all the foil tabs with that cell group and it appears all are connected properly so.... garsh... it seems like it might be a bad cell however, testing *that* would require much more effort as it appears I can't just simply isolate the cell without ripping up the soldered connections.

 

[dogman dan]

I wouldn't know how to tell which pouch is bad in a cell group. I'm not sure how the terminology is supposed to be, but I call each group of pouches a cell, and each pouch a subcell. I don't know if you can test subcells without disconnecting them or how if you can. I suppose with a wattmeter you could load test the capacity of each cell, and find out which is the low capacity one. Getting a one cell charger and charging each cell and timing it could tell you which one is not taking as much charge without a wattmeter. Load testing and timing it would work too I guess. Too bad you didn't find something obvious like a disconnected tab.

 

[swbluto]

I was also hoping for an obvious fix. :lol:

I just realized that the cells are disconnect-able from the BMS through a ribbon cable(I thought it was permanently attached until I read another post) so I can individually test each group of 2 sub-cells. Also, since the foil tabs are technically accessible(Haha getting alligator clips on it with all the rubber and paper in the way and the little available clearance), it seems the individual sub-cells are accessible if need be(But it doesn't seem you can isolate them without breaking the foil tabs). I think I'm going to test the cell-group until "depletion" and time the charge to find out the capacity of that cell-group, and then test the voltage of each individual subcell to tell which one is the bad one(But since the two subcells are in parallel, they'd "balance out" any voltage differences, eventually, but I doubt they'd balance THAT quickly to severely distort the testing results). If it turns out nothing is bad, then it might be that I have the wrong cell(My reasoning dictates otherwise) or the BMS could be implicated with the seeming theoretical possibility it's drawing too much off that cell or it isn't "balancing" it enough. But... I think that might be a remote possibility.

But then comes the issue of charging that cell-group back up.... hmmmmm...

I'd need a charger with an SLA/LiFePO4 protocol that outputted something meant for a 3 volt cell. I was thinking I could simply do a voltage divider(Like how a potentiometer could be used) so I could use my 12v 1-amp SLA charger, but in choosing the two resistance values I seem to run into the problem of too much resistance(too little charging current) and, conversely, too little resistance(Too much "leakage current".) which is a problem since I only have 1-amp available to me from the charger.... I think.

Or I could just plug the battery directly into the charger. With balancing, it'd only take... what... a day?

This has taught me: Always buy a battery whose cells are easily accessible(Like screw terminals) and has easily findable/purchasable replacement cells. On the positive side, at least this purchase didn't cost me an arm and a leg(Only an arm. XD).

 

[ZapPat]

You could get just one of the DC-DC converters that Docbass uses for his per cell balancing charger (see thread). You'll need a power source for the DC converter between 36 and 72V dc, and that will handle the load your converter can handle.

You will also have to set the voltage to limit current, but that's easy to do, just adjust your converter's output voltage a little higher than the cell's voltage before connecting them together, and then trim the voltage up slightly to get your desired charge current that will taper off slowly. (again see Doc's thread for info about moding the artesyn converter as to lower it's current limit).

Hope this might help you recover your cells a bit. If it doesn't work just bypass the damaged cells and you will end up with a slightly lower voltage battery (and have to get a lower voltage charger too).

 

[swbluto]

I emailed ping, even though I wasn't the original buyer(I felt so weird in asking him), but he thoughtfully helped. It turns out you'd need an adjustable power supply(Which I don't have but I consider it an item I'll eventually need if I'm going to pursue any serious endeavor in electronics) but here it is just in case anyone else comes up with the problem(Which might be the case if your pack starts to seriously dip in usable capacity prematurely) and, ZapPat, it almost sounds like Docbass's method in concept:

"3.07v is not so bad and can revive. it could be caused by imbalancing. Do you have an adjustable DC power supply? If so, first charge your battery pack with your charger for more than 7 hours. Disconnect the BMS and use the DC power supply at 500ma 3.3v to charge the 3.07v single cell to 3.2-3.3v. Then use 1.5-2a 3.7v to charge this cell till it gets to 3.7v. Just leave it for hours or one day and see if the voltage drops to less than 3.2v. If not, congratulations, it has revived. Finally, connect the BMS and use your normal charger to charge the whole pack for hours."

I don't really know how to isolate the cell to charge the "single cell" without regrettably slicing and dicing, but I assume this prescription would equally work with two cells in parallel(*that* I can easily isolate). Hopefully it's not like a defibrillator in the sense it could help someone dying but it'd kill someone healthy.

Anyways, I just have a question. How does a cell "Get out of balance" when it seems balanced after charging(The cell's voltage is the same as everyone else's) unless the cell itself is "bad"? I'm thinking that if the cell is balanced after charging(same voltage as everyone else), then it should decrease in voltage at nearly the same rate as everyone else as it contributes the same amount of current. But my voltage charts shows this is not the case so there's some unbalance that I'm not understanding. Can anyone clarify what causes this voltage disbalance as the battery drains that was originally voltage-balanced when fully charged?

Later added:

I just did a discharge test of those 2 parallel cells using an M16 35watt light bulb and two multimeters(one acted as voltmeter and the other as an ammeter) and, at 1.41 average amps with a range of 1.34-1.5 amps with working voltages ranging from 3.7 to 2.75v and resting voltages between 3.7-3.05(I have the spread-sheet if anyone wants it), my calculations show 7.4 usable AHs which is pretty close to the amount my capacity available in the battery. It's clear these cells are the culprit and I can rule out a "Disconnected cell" or something like that since the discharged amount was greater then 5-6 AHs and it appears there are little resistive losses where the foil tabs connect from the lack of heat at the connections, so it's definitely the cells. Is there a way to restore their capacity? I'm not entirely sure ping's "revival" of the cells would restore their original capacity as they were already at 3.7V: What effect would pumping more amps through them have?

 

[swbluto]

I purchased a good old power supply(HY1803D mastech - 0-18V, 0-3 Amps. I'm sure it'll serve me well for nearly everything.) and followed ping's instructions. I'm in the middle of the cell test(Actually, I have about 5-6 hours to go) but, so far, it appears that there are 10.83 Ahs available from my voltage extrapolation. So... it actually seems like it was "unbalanced"!

Anyways, I'll have confirmation and detailed information later.

Added after testing:

Here's the test data

It appears that it still has a capacity of only 7.4-7.5 AH as I originally measured which indicates this cell-cluster is "weak"/"sucks"/"is bad"/"needs replacing". It appears the voltage curve over the discharge cycle was different this time as it maintained a working voltage between 3.2 and 3.3 volts for around 2.333 hours whereas the original discharge curve dipped below 3.2 volts within the first 6 minutes of discharge, however, it started plummeting much more quickly, this time, once it passed 3.2 volts so it seems the capacity of the cell-cluster is still the same.

 

[swbluto]

I just had a thought. I've heard that "bad cells will make their neighbors go bad" but(as usual) I wondered why. As I figured out that cells with ever so slightly different voltages in parallel will discharge from the higher voltage one into the lower one(I.E., the "good into the bad cell"), this would increase wear-and-tear on the good cells in parallel but NOT the ones in series. So, if I'm correct, the damage to my battery will be limited to the nearest 1-cell neighbor. Is this an unreasonable assumption? I'm just wondering if I keep using this battery, eventually that bad cell will spread its misfortune to the rest of pack. All I really want to do is keeping using it until I have the means to repair it, but I don't to inadvertently damage the whole battery in the process but it's OK if those two parallel cells die a horrible death.