Volgood 48V20Ah LiFePO4 Repair by Amberwolf

amberwolf

Administrator
Staff member
Joined
Aug 17, 2009
Messages
40,463
Location
Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion
This is probably going to be similar to my Vpower/CammyCC repair thread here:
http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=22750&start=45
except that this is a repair for RTLSHIP, whose pack drops by several volts not long after charging.

Based on his results, including a test he did of unplugging the balance wires from the BMS right after charging completed, and getting the same voltage drop then, it is likely to be a cell or cell-group issue just as mine (from Bluestreak) was.

If it were the BMS, it should not have dropped so far down after unplugging the balance wires from the BMS, as it could not then drag cells down (assuming a problem with the balance circuit(s)).

I suspect a bad cell dragging down a group, or a spotweld problem, or both.

Bonnie the chief inspector had to approve it upon arrival to the repair facility, of course. :)
View attachment 24

It was well-packed, and has not suffered any shipping damage that I can see.
View attachment 23

View attachment 22

View attachment 21

View attachment 20

View attachment 19

View attachment 18


Total pack voltage is 52.9V, across the BMS pads.
View attachment 17

I tested each cell (pics attached but not inline to make reading post easier) via the balance wires a few minutes after it arrived. I unplugged the balance connectors from the BMS and taped them to the top of the BMS for easy hold-down while I measured things. Note that the reading on screen is not the cell-voltage; I just noticed that it must've snapped the pic at a moment I didn't have good contact. :( It was the same as the next one up, 3.56V
View attachment 16
Also note that due to my arm being in the flash's path for most of the pics, it's a little hard to read the values on screen in these smaller versions of the pics. :(



There is only one cell group, number 6 from the most negative end, that is significantly different in voltage, and it reads very dead, at 0.75V.
View attachment 15

Marked on connector for easy locating next time:

and it happens to be the balance wire with the most blue markings, thus easy to trace, too.

Everything else reads about 3.48V up to 3.65V, with the majority at 3.56V or 3.57V.

More info later, after I begin disassembly to verify connections, welds, etc., and start determining the actual problem cell(s).

After that, I'll have to find a source for the cells to repair with; I have a couple of leads here on ES but have to see if they still have any.

View attachment 16

View attachment 14

View attachment 13

View attachment 12

View attachment 11

View attachment 15

View attachment 10

View attachment 9

View attachment 8

View attachment 7

View attachment 6

View attachment 5

View attachment 4

View attachment 3

View attachment 2

View attachment 1
 
In preparation for actually repairing this pack, I decided to first check out hte problems on my own, to refresh me on pack construction style, test and measurement, etc., so I can make all the mistakes on my own pack, and do it right on this one. :)

http://www.endless-sphere.com/forums/viewtopic.php?p=446576#p446576

So far it's going ok, so I shoudl be able to start this one tomorrow, after I finish a bunch of household stuff. Worst case, Monday evening once I get home from work.
 
Due partly to my little crash on Monday morning, I didn't get this started until a couple of hours ago. But it might be fixed already. :)

First I took off the plastic wrap holding the BMS to the pack, and carefully rolled up that plastic to put it back on for shipping back to RTLSHIP when it's done. Then I looked at the duct tape wrapping, peeling up the strips at the corner/edge where all the wiring goes in, to see if it was across the top like mine, or some other way.
View attachment 25
It is across the top, but this pack has some foam padding between the inner and outer layers of tape, whereas mine had thin cardboard. The foam will tear to pieces if I try to unwrap the tape like I did mine, so I have to cut the tape very carefully, all the way thru the foam but not into the second layer of tape (don't wanna cut wires by accident, or nick cells). This took some time, but worked, and then I could peel back the top of the pack's wrapping to see the wiring scheme.
View attachment 24
That was necessary so I would know which half of the pack the bad group is in, so I would only have to open up that half (and maybe not even that, depending on how the rest of it was packaged).
View attachment 23
It was on the "right" half with teh wiring end towards me; you can se the brown marker I used to trace the balance wires on either side of the bad group from the balance connector.

While there, I also noted that the very corner cell has it's shrinkwrap damaged; I can't tell when it might have happened or why, but there isn't anything else I can see wrong with the cell.
View attachment 22
I also noticed that the duct tape comes in at least three types, two of which seem to be much better tape than the innermost. I am suspecting at this point that the pack has been apart before, and the tape under the balance-wires' hotglue is the original, and the other two colors are "new".


I split the ducttape and foam all the way thru right at the seam between the pack halves, on both ends and the bottom, so that I could unfold the pack open, after I found the end with the connecting tabs (opposite the pack output end with the wires on it).
View attachment 21

View attachment 20


Laid open, it has exactly the same cardboard piece and rubber sheet that my Vpower/CammyCC pack does, to separate the two halves. Niether is glued down.
View attachment 19

Pulling those off,
View attachment 18

I can see there definitely was a repair done, to the lower left cell in the pic, which has been soldered back to the string on the end. (most negative, IIRC).
View attachment 15

There was also a popped weldset on the "topmost" cell gropu on this half of the pack, on teh far left end of that group. Easily fixed with the 80W Weller iron, with about 1.5 seconds of soldering. Couldve been popped already, or done so while opening the pack.

There was a bit of solder debris hiding in there on a strip, probably from the factory but possibly from the cell repair.
View attachment 17

View attachment 16


Close examination of the rest of the tabs on this inside of the pack (rest was left wrapped in it's tape and foam) found at least one blown-out weld, like this:
View attachment 14
View attachment 13
THe rest are ok, so I guess it's safer like that. :)

THis is the group that's bad, with the pen on top of it
View attachment 12


I hooked up the group to the meter via balance wires via my contraption made up for my own pack's testing.
View attachment 11
View attachment 7
So next I start cutting the tabs between cells in the group, to do a binary search for the bad one.
View attachment 10
To cut them, I just scored them deeply with a sharp blade, then gently lifted the cut with the edge of the blade from underneath. All cuts were made only to the inside ends of the cellls.
View attachment 9

I highly recommend AGAINST doing it this way, becuase you could drop the blade onto cells and short stuff out; i was lucky and didn't but I thought about it just as I finished the last cut....


I first cut the two 9-cell groups from each other, hten the side that drops in voltage is cut in half, then next and next.
View attachment 8


Instead of used solderwick braid, thie cut tabs on this one will be fixed with strips of metal (either tin or nickel) that came from a big old console tv, formerly used as ground straps for the CRT.
View attachment 6

View attachment 5

They're not jus tsteel, as they solder very easily.
View attachment 4

I put strips of card stock under the tabs to be repaired, so pressure on them with the iron doen'st melt thru the cell shrinkwrap an short anything.
View attachment 3

Each group of cells came back up to 2.1-2.2V by this time, without any input from anything, except for the isolated cell, which had dropped to 0.12V by this time.
View attachment 2

I soldered the tab repairs on the other cuts, and started charging that group with the Sorenson on 3.65V 250mA, where it will stay until it gets to 3.0V, and then I'll go up to 500mA. That was about 530am, and was at 2.87V by around 6am. (added note: while typing this, it reached 3.11V so I upped it to 500mA, at 704am)
View attachment 1


After it has recharged that group, sometime tomorrow or the next day, I'll test the pack as it is now on DayGlo Avenger for a short ride to see how it works. Since it has a different connector, I'll temporarily solder on one of my big Anderson Multipole connectors to the BMS outputs to do the testing with. If it works ok, then I'll see about getting a replacement cell for the bad one.

If it still shows any issues, I'll further troubleshoot it as needed.

Right now all cell groups except the very first one on that side (the one with the soldered-in cell at the pack corner) are 3.48V to 3.51V. That group is 3.25V. Might just not be balanced, or it might have a problem. I'll keep rechecking it after the rest of the pack is working ok, and as I do testing with it.

Now I'm off to take a nap, as I've dozed off several times typing this.
 
nice work. a single bad cell can lead to pack failure?
 
Thanks for posting your work here AW.
About to start surgery on my own 48V20Ah pack that wont charge past 36V at the moment.
Good to see some pics of what to look for.
Cheers,

Glenn
 
Glenn_M said:
About to start surgery on my own 48V20Ah pack that wont charge past 36V at the moment.
Ouch. Are you sure it's the pack and not the charger?

(and if you don't have a thread for your pack, start one and we can move this troubleshooting there and link it here)

Before you do anything else with it, just disconnect the two balance plugs from the BMS, and measure between each set of pins (writing down the readings in order or taking pics of them) like I did for the first post.

Any group that is below about 3.3-3.5V when it's "fully charged" has probably got some sort of issue--it could be broken spot welds leading to damaged cells from overloading those still connected, or bad cells internally to start with, or even a bad BMS channel dragging them down or keeping them from fully charging.

If all cells appear to be about the same, it's probably not the cells but rather the charger not outputting enough voltage to recharge the pack.


RTLSHIP said:
a single bad cell can lead to pack failure?
Yes, apparently:

When the cell is internally defective so that it is a low resistance "short", it drains not only itself but everything in parallel with it. So the entire cell group of 18 cells goes down, and can be killed by it if left that way.

Once the group is below LVC, or close enough that drawing power from the pack to run anything drops it below LVC, the BMS will cut out all power from the pack to try to prevent damage to the cells that are low. If you have a dead group, the BMS simply won't allow power draw at all. Effectively, this "kills" the pack, in taht you can't use it until you repalce or at least remove the cell that's dragging down the others.


LiPo or LiCo I would not even *try* to recharge after this, (well, *I* would but I wouldn't recommend it to anyone else!) but I've had zero problems (that I'm aware of) with the cells that this happened to on my LiFePO4 pack, essentially just like yours, and the same is true of most of the other people I've read of or heard of that have done the same thing with various types of LiFePO4 cells that went below 2V, and then recharged them anyway.

In theory you'd get better results replacing the entire group of 18 cells, because they may not have the same capacity they used to, or deliver the same current. I don't know for sure. Best bet on getting groups of them is from Volgood, CammyCC, or Vpower; IIRC someone else reported getting a 9-cell pack for $20 or so (don't know if that included shipping from China), which if it did include shipping would add about $40 to the repair, for 18 cells to replace the whole group. If shipping isn't included, it could be a lot more.

The pack I have was already down below 15Ah before it failed for BLUESTREAK, and it now has about 12-13Ah before LVC hits, maybe a bit more. I suspect (know, actually) there are other cells in at least two other groups in mine that aren't up to snuff, but I don't have replacements for any of them so I'd rather keep what I have of their capacity than take them out. :)


But AFAICS it isn't the cell group that had a bad cell (in mine) that goes low first (even though it has less capacity in that group now), so my lowered capacity doesnt' appear to be because fo the bad/removed cell, but rather one of the other groups that always reads a bit lower than any others.


Your pack's status right now is still charging; as of an hour ago it's up to 3.41V with the Sorenson turned on, and drops to 3.32V with it switched off but still connected. Should be about 6Ah back in it now, I think. At 0.5Ah per hour going back in, it should be "full" to the original "20Ah" capacity in a bit over a day (~28 hours). :)



No heating of cells is noticeable compared to any of the others in the pack that aren't being charged right now, at least not in the 94F house. :roll:
 
Amberwolf, does trimmed mean that you remove a few of the really bad cells?
 
Finally finished charging, AFAICT. It reached 3.64V with sorenson attached, and stays the same with it powered off and connected, or disconnected.

Left it with the DMM on it for several hours while I did yard work and stuff, checking each time I came in for water, and it only dropped down to 3.58V. A couple hours later and it was still the same, so Iv'e hooked the BMS back up and will see where the groups go from there.

My work schedule for next week isn't out yet, so I wont know till tomorrow night (saturday) when I will have the chance to put this on DGA to do some test runs, and then on CrazyBike2 (which can put more of a load on it for longer, without risk to the motor).

The bad cell is still physically present, just not hooked up at it's positive end. Still need to get hold of a replacement cell.
 
Amberwolf, you saved the other cells in that string but not the bad one? Is there any way of measuring how many cycles this pack has gone?
good work :)
 
THIS WAS MY OLD PACK IT GAVE ME 15.5aH I DID SOME WORK ON IT... IT GAVE ME 16.5ah

http://www.endless-sphere.com/forums/viewtopic.php?f=3&t=8807&hilit=ICECUBE57+GOLDEN+MOTOR&start=45#p141794

Read here for a history... the photos are down but skim through a few pages it will get u caught up on my saga.
 
true, icecube57, I used it maybe 1000 miles so far. Not giving up yet on it.
 
Glenn M could let us know some more infro on your what brand how old about how many cycles. The fact that it's lower than 36v bring up some red flags. Could be other issues the dead cells ( cell ). Can be easier then the knife.
 
That's part of the problem - I don't know. Was bought of another person.
I *suspect* its a Vpower battery / BMS (and GM 1000W hub). 18650 cells (lots of them) All I know at the moment is it appears to not be the charger or the BMS
Don't know the age, # of cycles.

Its a 15s pack, so using AW method in this thread here are the voltages at the balance wires:

1s- 3.6V
2s- 0
3s- 0
4s- 0
5s- 3.36
6s- 3.35
7s- 3.35
8s- 0
9s- 0
10s- 3.35
11s- 3.21
12s- 3.35
13s- 3.35
14s- 3.35
15s- 3.35

I am hoping for an easily repaired broken connection between the strings of batts.
Surgery to come :twisted:

Cheers,
Glenn

Continued over in this thread http://endless-sphere.com/forums/viewtopic.php?f=14&t=31310
 
too bad the pictures are not available neither here or the Vpower thread.
 
They're simply missing until all the attachments finish transferring to the new server; see the announcement thread at the top of each forum. ;)

Did you have a pack like this you need to repair? If so, you could start a thread about it with specifics about it's failure mode, details about cell voltages, pics of it, etc., and people could help you with it.
 
Thanks I read the announcements after .

The upcoming wiki feature sounds great it should cut down on search times.

Thanks to Justin and Mr Vass
 
I was going to be testing this pack on the bike in the last week, but with Bonnie being really sick plus working six days a week, I haven't been able to do much that doesn't involve just sitting at the computer (next to Bonnie on the bed). Bonnie died yesterday, so now it's just the six days a week (and dealing with the aftermath) that are putting a damper on time. Hopefully be able to restart work on this and other things by Wed, my next day off.
 
I always have a lot going on...just, sometimes, it's too much, cuz life tends to happen to me a lot all at once. :? I still gotta do what I gotta do to keep me and the dogs in food and rent, though.

Good news is I finally get a week's vacation starting this coming Sunday. I'd like to sleep thru the whole thing, but I doubt I'll get that chance. :lol:

Some news on the pack, though. I couldn't sleep at all last night, so while I was way too tired to try setting up a load test or put it on the bike to do a ride test, I managed to eventually measure all the cell groupss while hooked to the BMS, without any other load on it.

From most negative to most positive, line breaks indicate ends of balance connectors:
1.928 (this is bad news)
-----
3.400
3.439
3.429
3.421
3.433
3.441
3.431
-----
3.436
-----
3.425
3.413
3.418
3.432
3.344
3.359
-----
3.404
Pack voltage is 52.1.

The pack voltage and most negative cell measurement are from using the output ground for reference, but if I measure using the *charging* ground, I get 53.3V pack voltage, and 1.973V most neg cell.

BUT: if i measure between those two grounds, I get 0.001V. Say what? :?

None of those three numbers can be made to work together to give me the right sum or differnece to explain each other. :roll:

Anyway, I set the pack up next to CrazyBike2 that has the CA on it, and used the CA to measure the charger output and total power put back into the pack, then charged it for the 3 hours or so I had left before I had to go to work this morning.

When I checked just before taking it off the CA and leaving for work, the charger LED had gone green, and was at 59.4V, with 0.687Ah put into the pack. The low cell was at 2.931V, so I'll need to manually charge that group with the Sorenson.

Since it's the most negative group, I'm guessing its' what the BMS is powered off of, and it has a low capacity compared to what it should, so it drained a lot from it while sitting the 3-1/2 weeks since I was last able to do anything with it.

If it's not that, then there is at least one bad cell in this group, too. I'll find out when I do more testing, which I hopw is tonight after I get more sleep than the nap I almost took while riding home from work today. (which left me with enough adrenaline in me to keep me from even thinking of dozing for the last couple of hours so far :shock:).


Now I've gotta go and set up the Sorenson to charge taht single cell-group.
 
I dozed off again before I could get the Sorenson setup, but eventually I got it going:
DSC04821.JPG

The Turnigy Watt Meter is being powered by an old router's AC adapter (on the powerstrip) via the aux port, since the 3.65V running thru it to charge the cell group is too low for it to operate from. I just took the little two-pin connector off an old NiMH phone battery that was long-dead and unchargeable, and stripped the wire ends enough to secure to the unaltered DC connector on that router's PSU, and plugged it into the TWM's aux port.

Current limit on the SOrenson is at it's max, of 750mA, at 3.65V output.

BMS is disconnected from balance plugs, but still connected to the pack itself main + and -.

Cell group started at 2.95V, and a few hours later is at 3.45V, according to the same DMM used to do the previous measurements. The TWM shows it as about 0.05V higher; I trust the DMM more. TWM has counted 4.65Ah into the cell group so far, with a theoretical capacity of 20Ah, and it's probably going to take close to that to fully recharge it given it's starting voltage.


At 0.75A, it should take about 27 hours to fully recharge; probably it will take less as the capacity of the cells is likely less than 20Ah, and that 20Ah is probably at a higher voltage than 3.65V (I don't know).


Once that cell group is done, I think i would like to perform this procedure on all the other cell groups, too, and then let it sit on the BMS for a few days, noting the group voltages periodically during that time to see if any fall significantly.

THen redo the charge procedure on each group, noting the Ah it takes to recharge each one, let it sit without the BMS connected, and repeat the group voltage measurements.

If all the voltaes are about teh same on either, or less on the problematic group when connected to BMS than without, then probably it was just one very drained cell group being further drained by a BMS issue.

No more time to work on it tonight, so maybe Wed.
 
Back
Top