Bad Lipo Cell

[retro]

I've been running 5A5s Turnigy 20C lipos in parallel, then in series, 10A37V, for about 23 charges and today my iCharger discovered a weak cell.

My setup is: soldered the discharge leads in parallel and kept the balance plugs separate, so when I charge using a parallel board, two packs at a time, I have two balance taps plugged into the board and one discharge plug, plugged in. My first question is how would my charger note the individual cell voltages when in parallel? It only reads 5 cells, while 10 are actually being charged. I have always balance charged my packs and final voltage always showed 4.1 per cell, that's my limit. Today my charger wouldn't show anything except a "low voltage" warning, beep at me and nothing more I can do.

I separated the bad parallel pack and checked the voltage from the balance taps and there is one bad cell in each pack. Probably one bad cell which pulled its mate down with it. .83volt in one cell and 1.36volt in its mate. I know this is crazy bad and I have my CA shut off at 33 volts and never went below 30% capacity. I did actually get to that point one day, but never went beyond the 33 volts.

What are my options? Pull the packs apart and make one good battery out of the good cells? Is there any way to gently revive these cells or are they garbage?

And, is there anything I should have done to catch this before it happened or is this something that was out of my control?

 

[Ykick]

For starters, maybe the iCharger is out of calibration? I had one that drifted and there's a calibrate menu for adjusting each balance channel.

You must have other ways to check cell voltages and not rely only on iCharger! DVM will work but tedious - a CellLog and/or BM6 come in very handy for verifying volts at the cell level.

But you're talking about parallel brick power leads but not balance plugs, correct? That seems odd to me, I only parallel after series connection for my final output voltage. 12S/15S and then parallel however much capacity I require.

What you're doing sounds like you should just go ahead and parallel balance leads as well as power leads and treat the resulting "bricks" as one large AH 5S unit series together for 10S or 37-40V.

Anyway, cell may not be ruined - if not puffy worth a try to revive but don't take only the word of the iCharger verify using something else to be sure the iCharger isn't lying - wouldn't be the 1st time.

After you know iCharger is good perhaps rig a simple connection and try to charge only the funky cell 1S Lipo from the iCharger. I do this all the time for my weak cells rather than waiting for the slow bleed balance of the RC charger. You might have to start 2S NiMh/NiCd or use motor driver mode until cell voltage is high enough so that 1S Lipo will commence charging.

Obviously, maintain low charge current 0.5A or so and keep an eye on that brick. If any evidence of swelling/puffing stop any attempt to revive and move on to scavenging the brick by removing the bad cell(s).

Ypedal has great video how to rework bricks. I've done it before as have some others - not hard but must have right soldering iron and understand the risks and prepare accordingly.

 

[liveforphysics]

If this is modern LiPo, it's dead empty at 35v, and you're just slaughtering it taking it to 33v. I'm not surprised at all.


4.15 to 3.5v That's the range to run modern LiPo.

 

[retro]

If this is modern LiPo, it's dead empty at 35v, and you're just slaughtering it taking it to 33v. I'm not surprised at all.


4.15 to 3.5v That's the range to run modern LiPo.

Thanks for the information, a little late is better then never.

Sadly I thought I was doing well keeping the battery above 33volts - http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=9170&start=30

I read that thread very very carefully and I don't recall seeing 35volts - 3.5volts/cell as being a recommended minimum. In fact, recumpence was the only one that set a recommended minimum throughout that thread. If there had been a higher voltage recommended, I would have heeded that. I guess that the thread was not talking about "modern" Lipo.

I agree, but disagree.

Here is what I mean;

In 5 years of running over 150 lithium polymer packs and many thousands of charges, I have never, EVER had one pack go out of ballance! Not even one! Lipo packs just tend to hold ballance. I also disagree that problems occur when smaller capacity RC packs are used. RC packs are typically the highest quality cells out there (nrmally by a very large margin).

Every single Lipo problem I have ever seen has been related to charging, not the type of cells used and/or ballancing or BMS.

I also disagree that the number of cycles is less than other cell types. For very long Lipo life, I recommend charging only to 98% capacity, keeping 40% capacity in the pack before recharge, keeping the cells over 3.3 volts under load, and keep them in a constant temperature (somewhere between 50 and 100 degrees. Very high and very low temp can shorten Lipo life as well). If they are run very hard repeatdly, yes their life will shorten. However, if you take care of them, you should see 1000 cycles or more out of your Lipo packs.


Matt

For starters, maybe the iCharger is out of calibration? I had one that drifted and there's a calibrate menu for adjusting each balance channel.

You must have other ways to check cell voltages and not rely only on iCharger! DVM will work but tedious - a CellLog and/or BM6 come in very handy for verifying volts at the cell level.

But you're talking about parallel brick power leads but not balance plugs, correct? That seems odd to me, I only parallel after series connection for my final output voltage. 12S/15S and then parallel however much capacity I require.

What you're doing sounds like you should just go ahead and parallel balance leads as well as power leads and treat the resulting "bricks" as one large AH 5S unit series together for 10S or 37-40V.

Anyway, cell may not be ruined - if not puffy worth a try to revive but don't take only the word of the iCharger verify using something else to be sure the iCharger isn't lying - wouldn't be the 1st time.

After you know iCharger is good perhaps rig a simple connection and try to charge only the funky cell 1S Lipo from the iCharger. I do this all the time for my weak cells rather than waiting for the slow bleed balance of the RC charger. You might have to start 2S NiMh/NiCd or use motor driver mode until cell voltage is high enough so that 1S Lipo will commence charging.

Obviously, maintain low charge current 0.5A or so and keep an eye on that brick. If any evidence of swelling/puffing stop any attempt to revive and move on to scavenging the brick by removing the bad cell(s).

Ypedal has great video how to rework bricks. I've done it before as have some others - not hard but must have right soldering iron and understand the risks and prepare accordingly.

Yes, I only have the power leads parallel. After separating the bricks, my one cell recovered to over 2 volts, enough for my charger to kick in and start a balanced charge. Now, I'll look into this other cell.
Would it be a bad thing to hook up the balance plugs of a couple of bricks on the parallel board along with the one with the bad cell and see if the low cell will balance up without the charger? Then try to give it a charge after?

 

[neptronix]

33v for a 10s pack has about 1%-2% of charge left.

You had a cell that was a little weak in the pack and it overdischarged it's parallel group.
Always, always test packs before you put them in big parallel groups to prevent things like this from happening, otherwise you get something like this occurring where you can make an entire parallel group have a dead cell..

Too bad nobody corrected that poster about the 3.3v/cell number back then; oh well.

Sorry if i am beating a dead horse, just wanted to make it clear for other people who are reading this thread.

 

[Red_Liner740]

key word in Recumbences post was UNDER LOAD!!!

a battery will sag under load but will recover when allowed to rest.

you were taking the packs to 3.3V at rest! so under load they were probably sagging like mad!

the issue here is that that one weak cell hit the cliff and plummeted in voltage when it hit empty. u then bulk charged and recharged the entire pack back up to 4.2V per cell. Which means two things. you never actually fully charged that one cell and you over charged the rest...at that point you drained the pack again and the partially charged cell got slaughtered again.

even if you could revive that one cell, i wouldnt suggest it. its been badly abused and all i would use that pack is for the good cells as spares. cut it open, solder the cells apart and store at 3.8V for when needed. 'beit a short which blows a cell tab, another weak dud etc etc.

 

[dogman dan]

Personally, I wouldn't trust either pack to be charged anywhwere but outside, or in some kind of bunker even if the cell comes back. Once you strip out the bad cells, you'll still have some good 4s packs. I'd consider any cell overdischarged or overcharged at least moderately hazardous.

No need to have a hot pack event on the bike when you could buy a 2s pack and replace those two cells. Or just run 19s for now. The best possible outcome is the revived cells kick the bucket pretty quick. The worst is you burn the house down. Chances are, they merely puff like hell very soon. But why take the chance for the cost of two lousy cells? Lets see, burn the house down or spend $20. I'll just risk the house.

A good voltage to stop at is about 3.75 v, or higher. Leaves some space for a low cell, and if all is balanced, stops before you are all that discharged. So call it 38v for your LVC.

 

[theRealFury]

Sorry to hear about your battery issue. As stated 3.3v per cell is just a little too low for a LCV as it leaves no room for any imbalances. Along with an LVC i use 2 of these. http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=320720945085&ssPageName=STRK:MEWNX:IT on my packs and they work wonders, are really cheap and are acurate to 20mv (tested them with 2 different multimeters and 20mv discrepancy was to largest i found when testing).

these allow you to set a minimum voltage that you want for each individual cell, all you do is plug it in before you go for your ride and if you hear the alarm (you cant miss it as it s loud as HELL) its time to pedal back home. the really good thing is that you can set what ever LV setting you want (to within 0.1v) so i have mine set to 3.6v.

Lipo really has to be monitored at the cell level especially while riding, all it takes is one loose cable or broken tab and a cell can go out of wack real quick.

 

[El_Steak]

Along with an LVC i use 2 of these. http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=320720945085&ssPageName=STRK:MEWNX:IT on my packs and they work wonders, are really cheap and are acurate to 20mv (tested them with 2 different multimeters and 20mv discrepancy was to largest i found when testing).

these allow you to set a minimum voltage that you want for each individual cell, all you do is plug it in before you go for your ride and if you hear the alarm (you cant miss it as it s loud as HELL) its time to pedal back home. the really good thing is that you can set what ever LV setting you want (to within 0.1v) so i have mine set to 3.6v.

Those buzzers are interesting, especially if they are loud enough to be heard with a full-face helmet at 40mph.

Any idea what is the current draw on them when they are plugged in ? Wouldn't want to kill my pack because I forgot to remove them overnight.

 

[retro]

key word in Recumbences post was UNDER LOAD!!!

a battery will sag under load but will recover when allowed to rest.

you were taking the packs to 3.3V at rest! so under load they were probably sagging like mad!

Actually, 33volts was UNDER LOAD!!! At no time did I say it was at rest, the CA shut it off at the magic number.

Thanks for all the information. I'll chalk this up to a learning experience and move on.
Remove the bad cells and work with what I have.

 

[neptronix]

Load varies with how many amps ( C rate ) you are pulling.
Like for example, i can get 2 volts of sag on my 10s pack if i only use 5ah, or 0.5 volts of sag if i am using 20ah.

When we talk about a battery voltage under load, it is useful to know how far it is sagging from it's base voltage.
Either ways, below 34v ( resting ) is pushing it, some lipos have worse cells in their packs than others.

Sounds like you can make a 14S-13S pack now, just need some new 5s bricks.. hehe

 

[mvly]

Yeah some recommendation is to balance charge everytime. It does take longer but In The end it might save your battery. As for the 33v, yeah rat is low. I would recommend to cell monitor during the ride. It helps to know individual cell voltage.

 

[neptronix]

You absolutely don't need to balance charge every time, and it will not save you from this unless you already have a cell that is awfully misbalanced.

When making big parallel packs, always do a full discharge test until a cell hits 3.0v to determine if you are sticking a pack with a dud cell in your parallel group!

You can balance charge all day to the top, but at the bottom of the charge, nothing's going to save you from overdischarge unless you have a BMS.

Out of 13 lipo packs i recieved, 2 packs had cells in them that were extremely low in capacity. That is a higher than 10% rate of duds that would have caused the problem down the road that the OP had.

 

[bobale]

always do a full discharge test until a cell hits 3.0v

We're talking here about under-load voltage of 3.0V, right?

 

[neptronix]

Under load works, you don't precisely have to hit 3.0v.
The 3.2-3.1v it will spring back up to is close enough. ( see my graph ).

You just need to know, most of all, that you don't have any packs that have cells that are hitting 3.0v or below really early on.

Such as this 10 cell pack ( 2x 5s in series ); one cell sags a lot.. the other one sags a lot and has low capacity too.
The above pack would be one that should not be used. One cell is missing 500mAH or so, and starts it's descent towards 3.0v far before the other cells.

And another note on balancing.. good lipo packs don't fall out of balance very often. You would be surprised at how well they stay in balance over dozens of charges.

 

[EBJ]

so, when a dud cell is paralleled with a bunch of good cells, would that help reduce the effect of the dud cell?
example: in a 4p setup w/ one "slacker" cell, wouldn't the other three cells "boost" the performance of the dud? or at least take up the slack and work a little harder?
I guess another question i have is: In a paralleled group of cells, can a weaker cell drop to a voltage lower than the other cells? or does the paralleled group always drop voltage together?

 

[neptronix]

Yes, it helps reduce the effect of the dud cell.
They literally work as if they were one cell, each balancing each other out constantly.
The more good cells you put in parallel with it, the better.

But that cell group will always be a bit more stressed over a long period of time; eventually it will be the one that dies before all the others.

For example if you have a 20ah pack, and one cell group has 17.5ah, you are giving that cell group a fuller depth of discharge every cycle. You are effectively making one cell group work 12.5% harder.

 

[markcycle]

so, when a dud cell is paralleled with a bunch of good cells, would that help reduce the effect of the dud cell?
example: in a 4p setup w/ one "slacker" cell, wouldn't the other three cells "boost" the performance of the dud? or at least take up the slack and work a little harder?
I guess another question i have is: In a paralleled group of cells, can a weaker cell drop to a voltage lower than the other cells? or does the paralleled group always drop voltage together?

I had a bad cell in parallel with a good cell and the bad cell pulled the voltage down of the good cell. There is a distinction to be made, a weak cells effect can be minimized by paralleling sure but most weak cells are just bad cells that self discharge on there own and those will pull down the paralleled good cell.

 

[bobale]

Neptronix, how much load did you put on those cells in the graph?

 

[neptronix]

Bob, that's an 8 amp load.
Also note that the internal resistance increases starting at about 3.3v.

While i'm at it; if you have an iCharger or celllog with a USB port, you can produce these same graphs with the included logview software.

 

[bobale]

Unfortunately, I only have Accucell-6, which I managed to connect to a computer, but it can discharge 6S battery at an unbelievably miserable rate, so I was thinking could I set my charger to Discharge, connect it to LogView, and then add some other load like light bulbs to a battery?

I know bulbs are resistive load, so current will vary with voltage, but at least I could load them properly and log voltage drop. What do you think?

 

[neptronix]

That's how i did it, halogen bulbs. Not an accurate way to measure amp hours, but a great way to determine if you're going to have a cell poop out early.

 

[bobale]

That was my think as well. After all, it's much simpler to measure capacity when charging (give or take charging inefficiency).

 

[retro]

I decided to scrap the parallel cell which was dragged down from the bad cell when I cycled my battery pack on the charger. I looked closely at the voltage of each cell and noted that on the lower end of discharge, 3.6volts, the paralleled cell was getting down to 3.4volts while the other 4 cells were still above 3.6v. Taking heed of the valuable information people have offered, the cell is gone... I now have a reassembled pack from the two affected packs, with 3 spare cells for the future.

A big When my 2 5s packs were in parallel via the discharge connection only (not the balance taps too), why would the exact same cell (2nd cell in 5s pack) be affected... was this just coincidence? I can't see how 1st cell would match to 1st cell, 2nd to 2nd and so on through the discharge wires.

 

[neptronix]

retro; my theory is that your balancing charger is doing a very bad job of balancing.
If you are not paralleling your balance leads ( you should be doing it ), that could be the cause. The batteries could be a little over or under full after a balance.

My iCharger drifted a little bit recently. Setting my LVC to 3.45v/cell saved me from catastrophe ( a bank of cells went to 3.3v while the others were around 3.4-3.5 ). I found out that it was off by 0.03v at the top on one cell, which actually made a big difference at the bottom.