Need help figuring out bad cels in GM 36 volt 16 AH battery.

[tex tiles]

Hi you guys. I have a GM 36 volt 16AH that started deteriorating after about a month from new. It is now over 2 years old but has only about 150 charges on it. I won't get into the lack of warranty or other stuff here, but the end result was I have a $550 battery that was defective and I'm on my own with it. I would like to try to salvage it if I can as I can't afford another.
 
It only goes a kilometer or so before cutting out etc. I have some A123 cels (what the GM cels are a copy of) that a friend gave me and I was wanting to replace the defective GM cels BUT.....
When I pulled the pack apart all the different wires to each pack were showing the correct specified currents. SO....

I tried again, this time trying to run the battery down as much as I could and then pulling all the connecting tabs inside each pack so I could measure each cel individually.  And they all spec out, that is all cells in each pack show the same charge......

Do the cels compensate between each other that fast? (20 minutes)

How do you figure out which cells are faulty?

PS I am not a technician, but can measure voltages and solder etc.

Thanks in advance for any help.

[Leslie]

Depending on the construction of the pack, if you can get some alligator clip mulitmeter leads and attatch them over one cell at a time on the battery, when it is at its malfunctioning depth of discharge.  EG when it cuts out,  take a reading on each cell when pushing the ebikes accelrator slightly while holding the ebike at a stand still.  You will notice the defective cell upon a small load volatge drop significantly compared to the good cells just before the BMS will trigger Low voltage cutout for this cell.

[tex tiles]

Thanks very much for your reply.
How can you do that when the cells are in packs? I have pulled the metal strips off that tie the cells together into each pack to test the cells and will have to solder them back together to run the battery.
I guess I could test the separate packs that way but not the individual cells.....
I will have to replace all the cells in the defective pack anyway because the A123 cells are slightly different specs.

[Leslie]

Then your free to deconstruct the pack at will and use the reamining cells for intersting projects,  the remaining good cells if Lifepo4 can balance without a BMS with some tinkering if theyre charged at 3.5v max.


You can test the cells indivdually and maybe make a jump pack for high voltages and a faster speed. 

You can use the cells in parallel also to run P7 3,6v led lights @ 2.75A to make awsome torches.  This makes one very long life torch that is easy to charge.

[Leslie]

Lifepo4 self-discharge between 5-10% / month at 25 deg. celcius

If you happen to leave you pack in a semi of high discharged state and or over 12 months without a charge topup your pack may fail..

I haed no idea Lifepo4 self discharges like this,

But Lifepo4 may all check good on the voltage until you put a load onto it. 

Here's an idea.

IF you were to leave the bike switched on for a few days to slowly discharge the entire pack the rouge cell might just rear it's ugly head.

Your cycle amount on your good old cells may allow the BMS to balance relaibly with your new cells if the internal resistance and max capacity between old and new is within 10%..

[Leslie]

Also a failing series connection may trigger LVC and over time get worse.  Give everything a good check over before you proceed.

[tex tiles]

Also a failing series connection may trigger LVC and over time get worse.  Give everything a good check over before you proceed.

like a cold solder joint or something? I doubt it is that as I alway get current.

[tex tiles]

But Lifepo4 may all check good on the voltage until you put a load onto it. 

I suppose I could solder the packs back together and try and run the bike under load and measure the separate packs. I didn't think about that. Perhaps that would show the offending pack and then I could just replace all the cells in that pack.
I'll try that. Thanks.

[Leslie]

You will only see a voltge drop on a bad cell or poor connection under operation of the motor under a load. 

EG using the bike motor with weight creates this load..

If you have a dry joint it raises the resistance though the connection. But full voltage can be allowed through a single hair thick wire when testing with a multimeter under no load..

Between the cell tab and the wire connection you may see a .5v or higher voltage drop but only on heavy currnet draws exceeding higher amps. 

If you draw 20 amps through a dry joint you might see a 2v drop between the tab and series connector.  This will equate to 40 watts adding heaps of heat to the terminal of the failing connection, the cell draws the heat into the cell and can cause damage to the cell and further damage to the connection. 

You cells would appear to behave more like a saggy lead acid battery on high loads to the BMS if the sense wire is connected after the dry joint.

I have experienced this on my SLA packs where I was having diffcullty securing screw connections to the terminals. One loose connection saw my battery get very hot on the end cell plate, to the point the plate boiled off all the electrolyte away and I lost the cell voltage, increased the resistance of the affected cell 10 fold. 

Typical comparison is thermal runnaway in LIPOS where the internal resistance gets so high self discharge become unmanagable spiral and the cells explode. 

Series connctions must be solid in all batteries with morderate to extreme loadings or else battery failure is inevitable 100% of the time..


Here is what I tell people who ask me about SLA's.   When you accelerate, SLA's may see a voltage slump over the terminals of up 1v on a 12v battery , thats four volts over 4 series 12v battery on a 48v pack. If you draw 20 amps, and see 4 v drop on the outer terminals.  4v*20amps = 80 watts of power loss.  Thats a sh!t load of heat, most inside the battery..  The batteries are working like a ressitor which is designed to limit both power and current by making heat.

With dry joints, the connection can be intermitant.  Lots of power when slow, bumping along the road ditzt dizt is when the problems occur, and hitting mounds in the roads can peak current draw in excess of 20 amps with direct drive gearless motors like the MP..

Just check the joints I cant hurt, and dry joints happen enough to make this worth a looksy..

[tex tiles]

what's a dry joint? I'm guessing that's when the metal touches but doesn't actually properly connect.....

[Leslie]

Yes, and also you can have cut out in contollers that has dry joint on the main input connections and the main power tracks.

Remember I said that heat can be created over a dry joint under heavy loads? 

The heat can expand the dry joint after a few minutes ride and either close the connection tighter or wider.   

Id rather think it's your battery cells problem like you suspect.  But wouldnt you be upset if you replaced the battery and the issue continues? 

So just a check of the whole power line circuit to be 100% sure before wasting a lot of money on a new pack.

[Leslie]

A few other things to check.

All connection between battery and controller.

Battery is charging correctly. Connection from charger to battery.  A dry joint on a charger plug or lose connection can give a false positive to the charger.

Check the BMS is balancing, all the BMS sensor leads are connected and test cell volts at the end of charge connected to the charger to see if one cell is over charging tipping the charger to off before all the other cells get a charge..

Check if output isnt slumping at the BMS fets.  Failed BMS trasnsistors can create high resistances before the controller and create a voltage slump that trigger cut off in the controller. 

If you can program your controller down to 24v this would be easy way to detect errors in both connections and BMS output fets.

I can think of more.  It's been a while since I did Ebike online help.

[Leslie]

People also have revived cells by manually charging each cell with a 4v mobile cell phone charger.

I use a 5v (5.25v) 3amp wall charger and a two diodes to drop the volts to a safe level to charge cells indivdual, Or a 2.2ohm 5 watt resistor to balance the high cells down to balance the pack, once it in forget for 2 mths AND DONT FORGET TO CHARGE...

Id rather no BMS at all.  In fact I don't use one on my Lifepo4 pack and havent for what almost seems like a year now.

I suspect a BMS has more chance of failure and screwing you over than a good lifepo4 pack has going out of balance, with charger set to charge at 3.5v (56v for 16 cell lifepo4) it should be fine.. 

I was always worried a sensor wire would come lose and the BMS would balance all bar the one cell. This situation will wreck a good pack over night, where I can ride a round and charge with no BMS for months..  BMS can lead to a false sense of security.  But make sure you charge before or after every ride. 

Again I rely upon my own connections to the charger, if I make a mistake its lights out, which would be my doing and not the manufacturer.  Every so often I make sure the charger does a long charges to make sure im getting juice into the pack a short charge would mean some of the cells are getting high, so I balance again.. 

Resistor over high high cells, tap tap the volts down on the high cells and its done.  Ride for another two months,

Manual balancing takes less than 10 mins out of my day every blue moon.

BMS woes no more, and I run a much more efficient power train without the BMS output fets in the way.  I always charge my pack as soon as I get home and never exceed my usual riding usual range to not risk over discharge.

[tex tiles]

I was just reading another thread about faulty battery and the instructions were to cut the individual cells apart by the tangs and not to solder directly onto the cells. Unfortunately I read this after I had soldered them all back together directly on the cells, so I suspect I may have wrecked the cells......

[tex tiles]

I have soldered all the cells back into the original pack sequence (I don't notice any difference after soldering directly onto the cells so I guess I got off lucky).
The cell packs all decrease in voltage when applying juice and go into LVC. I have measured all the packs individually and I suspect that these are just crappy cells that they used. The only thing would really be to replace all the cells, which probably isn't going to happen as I can buy a new battery cheaper.
I have purchased a 36 volt 20 amp battery for $329 from Codd Power via ebay and I hope for better luck this time around. I'll let you guys know how it turns out regarding reliability and lifespan.
The GM Bike kit I bought has been great and I am probably going to buy another soon for my wife, but the battery was a disaster as I got better performance from my old lead acid pack that only cost $75. $550 down the toilet.