Battery charging problem

Got a hardware problem? ask for help

Battery charging problem

Postby Mjolinor » November 4th, 2016, 3:43 pm

I have a number of (>2000) liion batteries. Each battery is 8 volts ish, two cells and a BMS with no marking on the magic smoke plastic dodads in there. One of the two cells has become somewhat discharged.

The centre of the two cells is not brought out but I can connect to it fairly easily with a pogo pin without damaging anything.

I can recover these batteries by tapping with the pogo pin and charging each cell individually until they are balanced again at which point they work as they should and give > 6 hours on time for the printers that they power, new cells will give the same on time so the batteries are fine.

The problem is that there are so many and doing them as I am it takes > 6 hours to do one.

I thought I could just use single cell liion chargers, two per battery and just charge them but this does not work, it lets the smoke out of the chargers when I connect them.

More looking and if I take one that has not been messed with and tap the centre I get around 4.10 ish for the good cell but I get minus 3.5 for the bad cell, 15 seconds on the bench PSU and I get around 3.5 volts on it and then I can charge it so there is something seriously up in the air with the BMS design or it may be deliberate so that 0 volts is presented to the device from the battery in order to know the battery si not right and flash the charge LED.

What I need is an easy way to revive these batteries, it seems such a shame to break them open and replace the cells then pack them back together when there is nothing really wrong with the cells inside.

I considered paralleling them up and charging a bunch at a time but this will lead to initially large currents between cells, probably bigger than they can handle so I would need a bunch of resistors to current limit for the first hour or so.

Anyone have any bright ideas of an easy way. I am using Thurslby bench supplies which are quite good but a difference of 10 mv will give a difference in current of 20 ma or so. I have to constantly monitor the voltages as they drift that much and it will not balance unless I am very careful.

I considered discharging the charged one down to the low level and letting the proper charger bring them both up together but this does not seem to give good results, discharging to a specific voltage is as hard as just charging both to 4.17.

Altogether a diddly of a problem that should have a simple solution without me having to build huge banks of chargers or replace the cells.

Any suggestions?
Mjolinor
 
Posts: 231
Joined: July 30th, 2013, 10:05 am
Location: Burnley, UK

Re: Battery charging problem

Postby bandersnatch » November 5th, 2016, 4:30 am

Hi,

You obviously know what you are doing, so I won't bore you with the usual caveats on working with LiIon batteries
and the black art curve-based voltage controlled + current controlled charging.

As I understand it, you already seem to have a method for regenerating the individual cells & are looking
for a way to speed up/parallelise the process for >2000 batteries.

It seems to me that you have 2 basic problems:

1/ Mechanical access to the individual cells
http://www.evdl.org/pages/hartregs.html discusses rejuventation of series-connected cells but the BMS is gonna get in the way
& the results of trying to charge both cells at once will probably be unreliable and unpredictable.
I reckon you are gonna have to bite the bullet & charge/rejuvenate the cells individually..

A photo would help here. If all batteries have the same format, you might be able to access both ends of both batteries by drilling
tiny holes in exactly the right location, to exactly the right depth.
However, a few minutes to open each pack is nothing compared to the 6 hours needed for charging each pack

2/ How to rejuvenate all batteries in less than >12,000 hours (!)
As you have already discovered, you cant simply charge multiple cells in parallel. You gotta have some sort of voltage/current regulation
for each cell. The official wisdom requires a complex combination of current+voltage regulation but your practical experience shows
that you can get good results by simply connecting a power supply & regularly monitoring/twiddling the voltage/current.

The suggestion below DOES NOT CONFORM TO ESTABLISHED LION CHARGING PRACTICE and is provided without guarantee or liability..

The charging method you are already using violates many LION charging rules anyway, so my suggestion is based on a way of
speeding up the process you are already using...
One possible solution is to build multiple (e.g. 10-20) examples of the simplest possible battery charger regulation circuit so that you can
automatically charge 10-20 cells at once.
Depending on the charging current, you might only need 1 resistor & 1 Zener diode for each cell...
This means only buying 20 resistors & 20 zeners.
If the charging current is too high for cheap zeners then you can use the zener to drive junkbox transistors
i.e. like: http://www.electroschematics.com/6253/r ... d-charger/

I suggest you first experiment with a single pack to find the simplest circuit that successfully charges a single cell automatically while you drink tea and watch Dr. Who.
When you are sure that this works ok, then just build 20 of the suckers & drive them all from a single psu..
This will reduce the "effective" charging time to less than 20 minutes per cell
Even allowing for the time to build 20 charging circuits & manually opening each pack, the total time will definitely take less than 12,000 hours...

STFB...
bandersnatch
 
Posts: 150
Joined: September 17th, 2014, 12:06 pm

Re: Battery charging problem

Postby Mjolinor » November 5th, 2016, 4:53 am

I can easily get at all three necessary terminals with three pogo pins but it does not seem to be necessary, the red and black wires on the plug are connected directly to the battery. I think the BMS is supposed to let the mid point drift one way or the other to keep balance. I would have to build a new holder though as the 18650 holder is not tall enough.

As you say it is the problem of building so many holders. The ideal solution would be to blast them somehow with enough power to bring the voltages almost equal in a few seconds but the risks therein are more than I want to handle. The risks at the moment are pretty low even though the rules are being broken.

http://roughtop.co.uk/john/battery.jpg
Mjolinor
 
Posts: 231
Joined: July 30th, 2013, 10:05 am
Location: Burnley, UK

Re: Battery charging problem

Postby bandersnatch » November 5th, 2016, 5:27 am

Yeah,
An attractive solution would be to apply a bit of good old "shock treatment" to the faulty cells & then charge normally thru the BMS.
But I would advise against this unless you are working in a bunker via remote control of a surplus Chernobyl robot.

There are many links in the net claiming success with shock treatment of LION batteries but I am dubious of these claims.
I have had great success using shock treatment for rejuvenating NiCd & NiMH batteries but LION batteries use a completely
different chemistry and are much more dangerous to mess around with.
(https://www.youtube.com/watch?v=MDohk-byqFM ) :^))))

For a volume of 2000 units, I reckon it is worth the effort to spend a couple of days to see if a resistor/zener combo will work..
Once you have a simple circuit, the time invested in the mechanical construction of even a 10x parallel charger will still pay off massively in the long run...

Good luck & be careful!!!! :^0 !!!!

STFB
bandersnatch
 
Posts: 150
Joined: September 17th, 2014, 12:06 pm

Re: Battery charging problem

Postby Mjolinor » May 8th, 2017, 12:58 am

Well I am now about 1000 batteries in using this in banks of 8:

http://roughtop.co.uk/john/battery2.jpg

The short red wire is connected to the +ve 18650 terminal on the holder and there is a sharp pin that pierces the heat shrink in the right place.

I am still somewhat confused about what goes on while these things are charging. The two small boards on the left are MP1584 based and they are terrible for holding a constant voltage. LM2596 based boards work a lot better but still not good and I am back to just using the Thurlby bench supplies as the source.

This is still not ideal because I want to get each cell to 4.20 volts and the PSUs will drift. If they drift up to 4.23 then the BMS in the battery will disconnect the input and it will stay that way until I draw current from that cell so in order to check the charge I have to let the bank of 8 current draw drop to < 10 ma or so then disconnect and take power from the cell then measure the voltage.

Initially I had a short piece of nichrome wire where the short red wires are in the picture but it seems not to be necessary, maybe the BMS is limiting (I am still charging through that) or maybe the current that does flow between unlike charged cells is not that big.

It seems that there are another 4000 of these things but at £7 per battery it is turning into a nice little earner. :)
Mjolinor
 
Posts: 231
Joined: July 30th, 2013, 10:05 am
Location: Burnley, UK


Return to Help me! Hardware

Who is online

Users browsing this forum: Bing [Bot] and 3 guests