Building a small A123 Battery Pack

[daniel.weck]

I need your expert hands and tools to fix my pack!
(no sexual jokes please )

Seriously, one of my A123 cells died and I need to either (1) reactivate it by applying a high voltage for a short moment or (2) replace it. Either way, I am gutted because this type of repair is outside of my comfort zone (soldering plates on live high-discharge batteries). My charger/balancer also failed, I am not sure yet whether the Turnigy Accucel-6 is damaged, or if it is juste the main 12V power supply. And...my controller also seems to be kaput (I think it is the source of the original system breakdown...domino failure it seems). I need to get my hands dirty and find out what is going on!

http://www.pedelecs.co.uk/forum/electric-bicycles/6956-diy-stage-9-a123-nano-battery-pack-photos.html

 

[jerrysimon]

Sorry to hear that Dan.

I rely on my bike every day, so glad I got it fixed. I do have my daughters pack now as well to fall back on if necessary.

Regards

Jerry

 

[jerrysimon]

Damn looks like another cell has gone. Got to work ok, but lost some power towards the end of the ride and getting the same error on the CellLog.

Its the same 6s1p pack. Will be interesting to see which cell it is.

Regards

Jerry

 

[Scottyf]

Its a good job you have those extra cells knocking about. It sounds like its becoming a bit of an issue.
Still least your handy and rebuilding the cell packs.

Its something I looked at when I was opening up my old 36v battery but in the end there's just to many cells to contend with in the cheaper packs all at about 2ah each. Just amkes for loads of connections and getting at one cell is a nightmare.

Least your pack is managable. anything with 40+ cells is such a pain.

 

[jerrysimon]

Yep it should not be too difficult to fix. Its charging now. The CellLog is not reading any errors and all the cells are at 3.38 with one at 3.35 so not sure what is going on.

Its the same 6s1p pack as before so ordering another one of those with welding tabs already fitted and connected up is pretty cheap. I still have two more spare single cells though

I do actually have another full 2*6s1p pack, buy my daughter is using that on her leccy bike a lot now so daren't take it.

I am pretty happy with almost two years use and my 4-500 charge cycles. Specs say these cells should be good for about 1000 charge cycles though.

One other thing, one of the cells in the suspect pack was definately slightly warm when I got to work this morning

Regards

Jerry

 

[Pedalo]

It might be interesting to peel the wrapper off one of your dead cells to see if you get any clues as to whether it is a genuine A123 or not. I think I bought mine from the same supplier as you. They're all still fine though I've not used them in anger for several months. They did seem a bit too cheap to be genuine...

 

[Scottyf]

Are you using two packs together or just the one?
I wouldnt expect as many life cycles if your making the cells push 4c or 3c. Least with a larger pack your not stressing the cells out so much. Which would show up the weak ones pretty quickly.

Maybe one of the connections isn't quite there or the repair?

 

[jerrysimon]

Yes Pedalo I just read that the genuine ones will have A123 Systems stamped on the metal casing. I will check tonight.

I think they are genuine. You can get even cheaper ones that are marked as DeWalt rejects.

I have just recharged the pack and all seems ok. I will attach the CellLog and put it in logging mode for when I cycle home, as this will then show me what is happening throughout the ride.

Regards

Jerry

 

[jerrysimon]

I use two 6s1p packs connected together in series. I am pulling no more than 10 amps in any one burst and mostly only 6 amps. These cells are rated to something like continuous discharge of 30C, 69 amps (60C, 138 amp bursts!) so I am well within specs.

The connections are fine. I suspect a second cell is dying. I will check tonight.

As I said for about £40 for each 6s1p pack, a couple cells in one pack failing after 4-500 charge cycles, 22 months and nearly 2500 miles means I don't feel too cheated


Regards

Jerry

 

[jerrysimon]

Ok battery seemed to perform ok on the way home and cells were failry balanced (within 100mV) at the end with no errors

Attached is the result from the CellLog displayed using the program LogView. Note the CellLog meter is attached to only one of the two 6s1p packs i.e. the one with the suspect cell.

The ride time is showed on the scale along the bottom i.e. 26m40s. The overall drop in voltage of the whole pack (big blue wavy line, scaled on the left) is shown for the whole ride. The individual six cell voltages are the ones at the top, scaled on the right. Five of the cells are pretty much aligned apart from cell three (green line) which is about 0.1 volt down on the others. I am pretty sure this is the same cell that I swapped out, though can't confirm unless I take the pack apart again.

I am going to run it all again tomorrow to see what happens. The new cell might just take a few charges to sync in with the others ?

Ps It appears that these are genuine A123 cells, certainly from checking on line and taking off the green outer wrapper of the failed cell to check.

Regards

Jerry

 

[Scottyf]

New cell will probably need a few cycles to ensure it gets to full capacity and discharge levels.
So your guess of it settling in sounds about right

 

[jerrysimon]

Well the same thing happened again this morning on the way to work. I definately put more load on the pack on way to work (it was very windy again) than on the way home.

This time I caught it on the CellLog as shown below viewed in LogView. Again the blue line is the overall pack voltage and the pink line is cell three starting off low and dropping in voltage till eventually it dies completely.

I took the pack apart and it is the new replaced cell that has failed

I wonder if I have just used a duff cell by chance. I have another couple of spares so will fit one of those. Not sure if to try a few more cycles first. Strangely it does recover when I recharge it. I must admit to being a little confused about the whole thing. I would not have thought a new cell would struggle in an older pack.

Regards

Jerry

 

[Scottyf]

You have to take them under load though unfortuantly. Its clearly not up to the job.
Keep on it it may get better with a bit more cycles under it belt. But looking at that graph its not looking good.

 

[jerrysimon]

Thanks Scott. I was chatting with one of the electrical engineers at work. He liked my graphs. He has suggested charging the single cell extra on its own to make sure it is fully charged before I put any load on it. I think one of the problems is that I may have put the cell into the pack in an unmatched charge state.

As you can see even when the pack is fully charged its still the lowest cell, voltage wise.

He also said when it dies the other cells effectively push charge back through it the other way, causing it to warm slightly and probably making things worse.

He said the battery pack has held up remarkably well and I have more than got my monies worth. Anything extra is just a bonus

Regards

Jerry

 

[Scottyf]

Thats eactly correct. The other cells will try compensate and put chage back into the cell to get it up to the same voltage. Hence if cells or packs are not withing a few 100mah the current flow causes the weaker cell alot of stress. Making things even worse.

I would have throught your changing process and balance charge would have brought the cell back in line?

Or do you not have a balance charger?


Might be a long process but once its done hopefully the cell should perform.

 

[jerrysimon]

Or do you not have a balance charger?

Yes I do but I tend to only use that sometimes usually at the beginning after I have built the batteries, otherwise I use a standard 36v 1.6amp charger as shown earlier in the thread.

I am getting all the cell voltages in line as I speak.

Regards

Jerry

 

[jerrysimon]

Ok almost finished. Charged the 6s1p pack on my dedicated Lipo/Life charger and afterwards the cells were all about the same except cell 3, which was about 0.15V less than the others. I then put that cell on charge on its own and put in about another 0.7 Amps into it before it registered as full

I remember now, before I put the new cell in the pack I had only charged it to the smae voltage as the others i.e. about 3.3v. Clearly although I charged it to the same voltage I had not charged it to the same capacity as the others. Its strange how even with the same voltage, near enough, cells can be of such different capacities. Makes sense though I guess, given voltage does not drop until they are almost empty capacity wise. I should have charged the pack to full capacity before taking out the old cell and then charged the new cell to full capacity before I replaced it.

I am hoping this was the problem in that because it was so under charged, capacity wise, everytime I was charging the whole pack it was registering full because the other five cells were at full capacity leaving cell three only two thirds charged.

Having charged the whole pack and got cell three to full charge, I am now using the balancing option on the charger to balance all six cells.

Hopefully all will be well tomorrow unless I have now damaged cell three, in which case I will have to replace it again this time making sure I ensure it is fully charged first.

Certainly and interesting learning exercise

Regards

Jerry

 

[jerrysimon]

Well good news today. Managed to get all the way to work without the cell failing.

As can be seen in the graph the cell is running a little lower than the others but when it had been at rest for a few minutes after the ride, it was within 0.036v of the others. Hopefully it will fully recover

I am currently charging the two 6s1p packs in series using the generic 36v charger so will see where they all end up after that.

I have not balanced the two 6s1p packs for a while and yesterday I did them both so they were all balanced to 3.6v. Looking at the graph the other 5 cells are even closer than they were yesterday for the whole ride.

Maybe it would be worth running these cells through my charger/balancer every month or so which would make them last longer. That said I have realised that I have actually put them through nearly 1000 charge cycles already if you count a charge cycle as one charge or one discharge ?

I had previously assumed a charge cycle was one charge+one discharge ?

Regards

Jerry

 

[Scottyf]

Well glad you have virtually solved the issue.
Might be worth doing a few more balance charges as really the new cell should be performing better than the older cells.

 

[jerrysimon]

Yes thanks Scott.

I have just finished the charge at work using the generic 36v, 1.6A charger. Clearly the balancing of the cells has helped as all the cells are much closer now on both 6s1p packs at around 3.4V.

As you say I will do a few more charge/balance cycles with my dedictated LiFeP04 charger at home over the next few days.

After that I am going to balance both packs every month or so as clearly the generic charger then seems to do a better job when I charge the complete 36v dual pack.

Looking online I think a charge cycle is actually one charge+one discharge session ?

Regards

Jerry