Can we build a 18650 battery recycle centre?

[Skalabala]

Hi members

I was wondering if one can build a 18650 battery recycle centre.

Like having 20 cells on it and each gets monitored.
Discharge and each gets a capacity rating?

http://www.ebay.com/itm/Digital-Lit...510?pt=LH_DefaultDomain_0&hash=item2a4f4c0cce

 

[jaunty]

Good idea...im about ready to hire someone and i havent even got through 50 cells yet...

180 more to go.

 

[Skalabala]

Yeah I have about 500 for this batch :?

Something like: Space for 20 cells. Each cell has a resistor something like 5-10W and 4-8ohm resistor and then the control circuit with a relay on each cell
When the cell gets discharged to pre programmed voltage then the relay disconnects the load and the mah is displayed on a simple LCD.

 

[Syonyk]

You could, but to do it properly would likely be rather expensive.

You'll want individual cell loads, with voltage/current monitoring for each cell - so two digital voltmeters per cell (one for the cell voltage, one for the shunt so you can calculate current).

You can build a moderately accurate (10 bit) 4-port voltmeter from an Arduino's analog ports, so 2 batteries per Arduino - this would get expensive.

It's probably better to build a totally custom board, but that requires a decent bit of engineering work.

 

[amberwolf]

There are at least two threads about designing battery testers like this, dunno how far they got. Both were recent, within the last few months. (this year)

 

[Skalabala]

I just can't see how you can test a bunch of old used cells in a string accurately

 

[dogman dan]

I know squat about this, but for a recycled tool or computer pack, charge some and check voltage would be the starting point for me, before doing any cutting them apart. True duds will be at 0v or whatever. Obviously less than the others.

Then once cut up, sort by internal resistance. Tedious for sure!

 

[Syonyk]

I just can't see how you can test a bunch of old used cells in a string accurately

You can't. You need to test each one individually. That's the only way to get per-cell data - in series or in parallel, you're not testing individual cells, you're testing a pack. And, if they're in series, you're probably driving the lower capacity cell below it's minimum voltage.

I know squat about this, but for a recycled tool or computer pack, charge some and check voltage would be the starting point for me, before doing any cutting them apart. True duds will be at 0v or whatever. Obviously less than the others.

Then once cut up, sort by internal resistance. Tedious for sure!

If you want to test for utterly basic functionality, yes, that would work.

A much more useful test would be to have chargers and load banks for each cell, and be able to automatically test each cell in parallel for several cycles to ensure the data is sane. Discharge it to the specified cutoff voltage, charge it to the proper voltage (while monitoring current & temp), discharge it, measure power/amp-hours/etc, repeat. Do this 2-3 times for each cell, and you'd have high confidence data that could be used to group cells.

Unfortunately, this is not a trivial system to build, and anything commercial capable of doing this will be very pricey (plus they don't seem to exist).

All this, to measure cells that are worth maybe $0.50-$1, at best. It seems like a financially infeasible option.

 

[Lurkin]

Depends whether you are trying to obtain enough cells for yourself, or set up a business reselling them. They are two very different objectives.

You would have to find the voltmeter/ammeter at a crazy low price. I think the last one I bought (which has the shunt built into it) was AUD4.40, which would get very expensive if you wanted to test lots of 100 and above. Also has no ability to log the charge. Pretty hard to make money this way. I imagine Doctorbass would be the man to ask.

Alternatively, it's just a slow process to obtain enough good cells for yourself.

 

[Syonyk]

You'd want something with a serial interface or otherwise on it that you could hang off a small computer (the GPIO pins on a Raspberry Pi would work great for this) to do logging.

Another possibility, if throughput isn't a big concern, is to build one charge/discharge unit, and have a rack of batteries it can cycle through - just a set of relays or something to connect or disconnect various batteries. It would take a long time, but running 24/7, it would be able to get through a few test cycles per day. You could also bulk charge/discharge by connecting multiple batteries to the bus.

If you are doing this, be sure to have a nice fireproof enclosure for the batteries under test that can handle the whole test lot going off at once. A flare stack might be useful.

 

[Skalabala]

I already did ask Mr Doctor
I actually saw this idea on the internet but I can not find it again
Even a device that can do 10 cells will be nice.

I do not really want to end up buying 4 or more rc chargers
And an accurate meter for Ir will be nice

If we can get in contact with this guy
https://www.youtube.com/watch?v=x8nbHYOc8ns

 

[Skalabala]

This looks nice
http://www.ebay.com/itm/DC-12V-Batt...fe-Internal-Resistance-Analyzer-/171761180936

How many cells do you guys think you can discharge with this rc charger?
http://www.aliexpress.com/item/Radi...ne-time-Professional-For/32287810909.html?s=p

 

[Skalabala]

So far this looks like a good option.
I am thinking of getting two of these so I can do 8 cell at a time. Should be able to do realistically 24 cells a day
http://www.hobbyking.com/hobbyking/...x6S_Lithium_Polymer_Battery_Pack_Charger.html

 

[Skalabala]

Still wondering, because my two quad chargers got lost in transit as if I have too much money

How about 10 of these? Built into a nice format on wood board etc?
http://www.ebay.com/itm/18650-li-io...071939?hash=item2807a6ecc3:g:j54AAOSw14xWMdXq

 

[DrkAngel]

I bulk charge large numbers of cells to 4.20V
Let set for some days to eliminate self-dischargers!!!
Then perform a quick IR test, usually on pairs (~4400-5200mAh rated). - removes bad, obviously weak or poor cells.
String 60 similar IR cells as 30s2p and discharge with 2 x 60w 120v light bulbs = 1Amp per hour. (pairs linked with neodymium magnets)
This rates all cells in batch with a reasonably accurate comparative capacity.

Quick cheap and easy!
Cells are labeled with EG 3000mAh + 3.84V.
Test should be ended when 1st cells near empty, planning and stopping test "on the hour" simplifies further testing.
Re-series (30s) better residual voltage cells with similar from other batches for continued discharge and re-label EG 4000mAh 3.65V etc.

 

[Skalabala]

I bulk charge large numbers of cells to 4.20V
Let set for some days to eliminate self-dischargers!!!
Then perform a quick IR test, usually on pairs (~4400-5200mAh rated). - removes bad, obviously weak or poor cells.
String 60 similar IR cells as 30s2p and discharge with 2 x 60w 120v light bulbs = 1Amp per hour. (pairs linked with neodymium magnets)
This rates all cells in batch with a reasonably accurate comparative capacity.

Quick cheap and easy!
Cells are labeled with EG 3000mAh + 3.84V.
Test should be ended when 1st cells near empty, planning and stopping test "on the hour" simplifies further testing.
Re-series (30s) better residual voltage cells with similar from other batches for continued discharge and re-label EG 4000mAh 3.65V etc.

Hey bud, yes I know of your method. But the cells I have are not of the same year and brand and thus I prefer the full per cell tests

 

[DrkAngel]

Testing in pairs is just as effective and much simpler than separating and individually testing cells.
If a pair has poor IR or very low capacity ... then separate and test - one cell might be poor-bad.

Capacity is capacity, cells in series all discharge equally at 1 amp ... regardless of brand or capacity.
Comparative capacity not good enough?
Bulk discharge batch 3000mAh, then continue discharges in pairs, add 3000mAh and metered discharge for total capacity - saves many hours of testing.

Most importantly IMO ...
bulk charge all cells to identical voltage then
separate and let set for some days then
measure carefully for voltage loss - self-discharge!!!

Smaller batches of same type cells?
Metered series discharge with cheap 90V capable V\A\Ah capable meter.

Cells connected with small neodymium magnets between them. Thin steel wire on ends bridging strings.

Add external power to meter and you can disconnect and label cells with capacity as the reach your desired discharged voltage.
Just important to closely monitor cell voltages as they approach desired voltage - most cells begin discharging rapidly when they drop off the discharge curve "cliff".