A Lithium Ion ramble

[Simen]

This post might be better suited in Storage, but since i'm going to ramble a bit, it's better placed here in user diaries...

I've been playing around with LiIon batteries for a while now, and what i learned first, was that there isn't easy to obtain good and precise information regarding that chemical soup that all goes under the name Lithium Ion, or LiIon for short...

A good starting point was to study the BatteryUniversity site; it gives a nice introduction to the different materials used in LiIon cells.

Lately i've been grabbing with me some old laptop battery packs from the recycling center, to break apart and rip out the cells. in all the packs i've taken apart, there's only one cell(group) that's bad inside the packs; the rest are ok, though, with reduced capacity ofcourse. Laptop batteries are not always treated well by the internal BMS/charging circuit.

Now, what really bugs me, are the markings on those cells, and finding info about them. Since there are LiIon cells with different chemistry, with different capabilities, it's important to know what kind of cells these are, to safely recharge them and what loads they can handle.

For me, there are three main types of LiIon: Lithium Cobalt Oxide (LiCoO₂), Lithium Manganese Oxide (LiMn₂O₄), and Lithium Iron Phosphate (LiFePO₄) (For short: LiCo, LiMn and LiFe) There are a few others too, but they are not yet widely used, or still experimental.

Roughly said, their characteristics are:

LiCo; 3,6V, charge voltage: 4.20V:
Very high specific energy; typically 150-190Wh/kg, but limited power load, usually max 1C.
Needs a good protection circuit and cell balancing to prevent thermal runaway, which happens at 150deg. Celsius.
Found in laptops, cell phones etc. where small size, high capacity and moderate power drain are required.

LiMn 3,8V, charge voltage: 4.20V:
Good to high specific energy, 100-135Wh/kg and high power load capability, often around 10C.
Safer than LiCo, but still needs cell balancing and voltage protection. Thermal runaway happens at 250deg. Celsius.
Found in powertools, e-bikes etc. where higher power drain are required.

LiFe 3,2V, charge voltage: 3,65V:
Average specific energy, 90-120Wh/kg and high power load capability, often 30C and above.
Safest of them all, but still needs cell balancing and voltage protection. Thermal runaway happens at 270deg. Celsius.
Found in newer e-cars/bikes/motorcycles etc, where high power drain and charge cycles matters (1000-2000 cycles for LiFe, opposed to 500-1000 cycles for LiMn and LiCo.)
Edit: LiFe can be used as a replacement for FLA stationary batteries, since 4x cells = 12.8V, charge = 14.6V and 100%DoD = 10.0V

There are one more worth mentioning, and that's Lithium Nickel Manganese Cobalt (LiNiMnCoO₂) (Short: NMC). One of the newer types, wich can be tailored by the producer to give very high energy or high power, but not both at the same time.

Now, over to the Playing part...
As said, i've collected som cells from old laptops, and i also have an old 36V pack from an e-lawnmower, and a couple of months ago, my trusted (not so, anymore) 36V, 16Ah LiMn battery for my e-bike started to get less and less mileage after each charge... which led me to invest in 24 new LiFe cells, 3.3V, 10Ah each.

To cope with all these cells, i invested in a 10A universal LiIon charger that handles up to LiIon 10 cells or 30 NiCd/NiMn at a time. it has loads of different charge/discharge settings, usb port for logging to LogView, internal resistance measurement for each cell/bank, and most important: it can run directly from an 12V RE bank! Another neat feature: when running a discharge on a battery/cell, the charger can use the RE bank as dumpload... (It's an 'iCharger 1010B+' from Junsi)

I have now an assorted lot of 18650 cells, and to test them, i first charged them up, and then ran an controlled discharge down to 3.00V, and then a charge up to 4.20V. The charger then tells me the mAh on both cycles.

If cells are below 3.00V/2.5V, thay should be trickle charged with max 0.1C to get them above 3.00V first. If the cell are below 1.5V, they're trash..

Btw:
18650 are the physical size of the cells in millimeters; 18mm diameter, and 65mm in length. The LiMn cells in my bike are 26650 cells; 26mm diam, and 65mm length. Why the zero at the end? i don't know...

Most cells i got from laptops are Sony, and they're not exactly informative on their markings on the cells. Here's what i found out:

Sony US18650 and US18650GR:
G3 = 1600mAh ?
G4 = 2000mAh
G5 = 2200mAh
G6 = 2200mAh
G7 = 2400mAh
G8 = 2600mAh
GH1 = 2800mAh

The 'Gx' code are on a new line.
The letters after the size indicate the anode material; no letters: Carbon. GR: Graphite. It's my understanding that cells with Carbon anode have an average voltage of 3.6V, and 3.7V with Graphite anode.

Samsung, and probably other manufacturer use another marking system, for example:

ICR-18650-20

Here, the 'I' stands for LiIon.
The 'C' are the chemistry:
C - LiCo
M - LiMn
N - NMC
F - LiFe
And 'R' stands for Round/Cylindrical.

The '20' number after the size are capacity; here, 2000mAh. (20DesiAh? )

I have a couple of cells that i haven't identified yet. They're labeled 'LGDB118650' and two more lines with lots of letters and numbers...

What's neat with these laptop packs, are that the cells are usually 2.0Ah or 2.2Ah, and they're connected 2p3s or 2p4s, this means that it's easy to separate them in pairs, giving me 4Ah and 4.4Ah 'cells' (2p) @3.6V; perfect replacement for 3s NiCd/NiMn cells. Most of the cells that were ok, had around 3.4Ah left in them; not bad for free.

And now - to find a use for them...
Actually, against all advice, i've replaced one (completely dead) cellpair in one of my functional laptop (Dell), and the surgery was a success! the 2 remaining cellpairs in the pack tested @ 3.3Ah, so i put in an 3.3Ah cell, replacing the dead one. Now i can run around 1h20min on one charge, opposed to zero seconds before.
I've also made a couple of flashlights with 2s2p cells and a 1W powerled spot. Good light, and over 24hr runtime.

[Madscientist267]

If cells are below 3.00V/2.5V, thay should be trickle charged with max 0.1C to get them above 3.00V first. If the cell are below 1.5V, they're trash..

The hair on the back of my neck stands up when I see 2.5V (or less) on a cobalt based Lithium battery, particularly the prismatic polymer variety; be careful. Some chemistries are friendlier than others.

The cobalt variety has a nasty temperament, and does not tolerate deep discharges. It won't bite you during the discharge, but can get really pi$$y in the subsequent charging process. The more the cell has been brought down below the "safe" mark, the higher the risk that it will let go.

The reason you had 'zero seconds' available in your laptop battery is because the charge controller sensed that one cell (or pair, or trio, as fits), had dipped below this safe threshold.

When this happens, the BMS chip says "no more, it's not safe to charge the pack", and shuts down the entire thing, to protect the end user from catastrophe.

It sounds like you've done your homework; don't take this the wrong way... I'm not trying to point out anything in the "obvious" category.

Just trying to remind you (and anyone else) that certain lithium chemistries have not just enormous potential for powering things, but as a ticking time bomb when they're abused (and that it doesn't take much to make them fall into that category). 

Steve

[taylorp035]

Wink (It's an 'iCharger 1010B+' from Junsi)

For my car project, I was looking at those chargers and decided to buy the 3010B charger.....  30A and 10s Lipo, it will charge just about anything.  Especially with these new LiPo cells that will charge at 12-15C, the higher current is nice.  If I'm really impatient, I just put my two 4s Lipo's in series to make it 8s and then I end up maxing out my 500w power supply    The dump load thing works well too.... you can even tell it to do more than 1 full cycle (I think up to 10) and it will tell you all of the "mah" values of each cycle.  It will also do Pb, NiCd, Nimh, and other lithium cells.


I can't wait until A123 systems comes out with their experimental cells.... If I remember correctly, they will discharge at 180C.... not sure what that would be useful for much past drag racing, but it sure would be fun!

[Simen]

Steve:
Sorry, i was a bit imprecise there; the 2.5V are the normal LVD for LiFe batteries. 3.0V are for LiCo and LiMn.

About the Zero seconds and the BMS, you're right. That's the BMS' job after all.

Thanks Steve; it's always good to get reminders about certain dangers when having fun.

All LiIon cells today have a vent to prevent explosion if overcharged etc., and a ptc which will disconnect the Cathode (i think) to prevent exessive current draw. if any of the two happens, the cell are rendered useless afterwards.

On that note, i'd like to say that Charging LiCo cells with anything above 4.20V +/-0.05V are a scary and unhealthy excercise; more so with the older cells. The cell's temp will rise quickly, and an internal short will happen, making the cell vent gases very violently, and get so hot that it might set fire to other substances.

The same should be said about LiMn and LiFe, but their thermal threshold are higher and therefore a bit safer.

Taylor:
Yes, it's a very nice charger. I've just started to discover the Logview program also; it makes nice graphs...

[Madscientist267]

Most definitely.

One thing not always obvious either to someone new to lithium is that pushing up above 4.2V/cell doesn't really gain you much.

The risks of doing so are orders of magnitude greater than the extra capacity the cell will store.

On my "insane" laptop battery, I don't even really like to let it push that extra 5% of capacity into the cells for most charge cycles. Depth of discharge isn't the only way to reduce lifespan of lithium - going over the top (even when there aren't catastrophic consequences) degrades the cell's energy capacity by a small amount every time it happens. It's another reason that balancing is so important for these cells. It is very easy (much more so than lead-acid) to perpetuate an imbalance that ultimately cripples the whole pack.

My general rule is, charge to 95%, drain to no less than 30% charged. About once every 10 cycles or so, allow the pack to go all the way up to 4.2V/cell @ a few mA, and allow it to balance out completely. It takes longer to complete the last 5% or so than it does to get the first "95%", sometimes meaning several hours to complete. This isn't immediately visible in the graph.

The problem with balancing on every charge is that they spend too much time in the 4.2V range. It's a balance between usable capacity, cell balancing, and lifespan. Not really possible to get all 3 maxed out at the same time.

Steve

[zap]

Nice ramble Simen

[Simen]

Nice ramble Simen

Thanks.
I wrote this mainly for myself in a textfile on my pc, but it's good info, so why not a public ramble?

Most definitely.

One thing not always obvious either to someone new to lithium is that pushing up above 4.2V/cell doesn't really gain you much.

The risks of doing so are orders of magnitude greater than the extra capacity the cell will store.

Agreed; pushing the charge above 4.20V gains only a few mA, and guarantees a shorter life. But cutting it too low, and you loose much capacity, especially older batteries, like those in my graph above. That's my old ebike battery; 36V, 16Ah (a 4p10s battery with 4Ah 26650 cells; LiMn), which you now can see, are a 9Ah battery...

Looking at that graph, if i had cut the charge right after they closed in on 4.20V (around 1:06), i would have missed over 2A. In newer cells, i have observed that the capacity are much closer to 100% when reaching 4.20V than in older cells, so one can say that they're like humans; the older one gets, the slower one gets and can't hold on for so long time anymore...

One other thing i've learned; LiIon doesn't like trickle charging much. I think the Ions accumulate somewhere they're not supposed to, and by that reduces capacity and increases the chance of failure...
That iCharger i have, does have a setting for trickle charging LiIon after ended charge. I've tested it, and it seems that it only squeezes in some more mA's after reaching set end-of-charge voltage (Edit: ...but only for a limited time). Without it, the end of charge stops at a certain current draw from each cell. in the above graph, trickle charge was off, and it seems that the charge stopped at 70mA/cell (one 'cell' here are 3.7V, 16Ah)

[Simen]

I've learned something new since last ramble...

I mentioned that old 36V pack from an e-lawnmower, and it contains 20 Sony US18650VT cells. Now, the VT series from Sony are NMC cells, and these cells have a max charge voltage of 4.10V, opposed to 4.20V for similar LiCo and LiMn cells.
The pack i got these cells from was an Bosch pack, and the spec says 36V, 2.6Ah, so the pack are 2p10s and each cell are 1,3Ah, and i think the Discharge rate are a healthy 30C or more. Also, i see Sony says that LVD are at 2.5V...

The pack have no cell balancing circuit, so it seems that NMC cells manages fine without it. This pack are 4 years old and well used, but still the cells are within 0.01V of each other. By lack of knowledge, i charged them to 4.20V, and the came out with a capacity of 1.25Ah/cell, so charging them to the 4.10V spec, i'm guessing the cells will hold 1.15Ah or thereabout. (I'll test later...)

[tecker]

I was checking out the possibilities and saw Magnesium batteries were bing given a good out look

[Madscientist267]

The pack have no cell balancing circuit, so it seems that NMC cells manages fine without it.

Most of the time, the "balancing circuitry" in small lithium packs serves more as a safety device, even though it does also provide balancing. Some lithium chemistries will not get fussy when they are overcharged or over discharged, etc. While it may cause cell degradation, it doesn't pose a safety issue.

My first real use of a chemistry that didn't need them, but should have had them (in hindsight) was the iron phosphate version. I found that after a few cycles, the voltages had gotten badly out of whack in only a 4 cell pack (series). Some were approaching 5V, others were below 3. They never even so much as got warm, but the ones that were hitting 5V aren't holding a charge very well anymore (not that I know for sure that they did to start with).  

Point is, I never felt like I was in any danger, but I would have used a balancer had I known they would get out of whack that quickly.

Steve

[Simen]

Yes' i'd agree. LiCo would NOT be wise to run without a balancing circuit, but several of the other, newer ones manages without it seems. But for long life, a balancing circuit would pay off in the long run i think.

My new bike battery are LiFePo; 2p12s, and i haven't any balancing circuit on them, but my charger have the ability, so i balance them every 4-5 charge. (Actually, i haven't charged them more than 3 times yet, so... )

Tecker:
Magnesium? Would that be those Lithium Manganese Oxide (LiMn2O4) cells?