Using low capacity cells for powerwall?

Hello guys,

I have crosschecked a lot of threads here, starting my first harvest. I have got around 140 notebook packs.

From first 30 pack, I have got mostly Samsung (the blue one, rated 2800mAh - I am not home and will edit the post with specs). I have got those packs in average for around 1.2E/pack - that I already consider too much, but in a future I hope to get better source.

Now my question is, has anyone of you ever tried to build a pack from 1500mAh batteries/or in general from batteries withless than 60% of original capacity? Most of my Samsungs has around 1500-1700mAh and IR=120-160mOhm. I try to cycle them and they looks solid, not much change in the capacity, itdoes not go down, but neither much up. They also do not heat.Everywhere wecan find a charts, how the li-ion batteries behaves when you cycle them to 80% capacity, then they are considered 'dead', but around those 80% the curve does usually look 'almost flat (linear)', isn't it that some chemistry will be still working just fine and handle another hundreds of cycles with much smaller capacity loss?

I know, that some of you has a problems with old Sanyo or Sony, that heats up, or dying fast with every cycle, but those Samsungs looks solid and stable at this point, does someone of you have tried to use them in your power walls with such a 'small' cap? What is your opinion?

You can use the lower capacity cells. There's nothing wrong with that as long as the packs in the series all have the same capacities. The IR is a bit high, but expected, and usually not an issue at the sub-C amp draw these will experience.

The big issue is the current capacity compared to original capacity. If you're aware that they have 60% of their original ratings, and you keep an eye on them, that's fine. I would recommend putting them in packs that would be easy(easier) to replace them. Unless, you build two strings. One has all your high value cells, the other with all the lows. Then as you collect high value cells, you replace the lower strings (as a whole, not individual packs).

Example: You have enough 80+% cells to make a 14s80p, or 1120 cells. And you have enough subPar cells to do 14s40p, or 560 cells. Then you could have two strings that are connected in parallel, either at the ends (requiring two bms strings) or each pack in parallel with the other string (using 1 bms string). Then you get another batch of 1120 80+% cells, you make another 14s80p and pull the 14s40p out and drop the 80p set in.

Korishan said:

You can use the lower capacity cells. There's nothing wrong with that as long as the packs in the series all have the same capacities. The IR is a bit high, but expected, and usually not an issue at the sub-C amp draw these will experience.

The big issue is the current capacity compared to original capacity. If you're aware that they have 60% of their original ratings, and you keep an eye on them, that's fine. I would recommend putting them in packs that would be easy(easier) to replace them. Unless, you build two strings. One has all your high value cells, the other with all the lows. Then as you collect high value cells, you replace the lower strings (as a whole, not individual packs).

Example: You have enough 80+% cells to make a 14s80p, or 1120 cells. And you have enough subPar cells to do 14s40p, or 560 cells. Then you could have two strings that are connected in parallel, either at the ends (requiring two bms strings) or each pack in parallel with the other string (using 1 bms string). Then you get another batch of 1120 80+% cells, you make another 14s80p and pull the 14s40p out and drop the 80p set in.

Click to expand...

Thanks for your answer, yeah it makes sense, something like that I wanted to make. I was thinking about having two invertors - one lowpower (1500W/48V)with 'bad'cells like mentioned above and use it to power lights in the house and second (3500W/48V) with a good cells for powering plugs. I have gas heating.

I need about 15kwh (with reserve, willing to use pack only between 3.3v-4.1v), so I was thinking to have 2 independent strings (one will be 5kwh pack with separate BMS, second will be 10kwh pack). Of course I still need to collect much more cells, I was just curious about those Samsungs, I have a lot of those, so it would be a pitty to just throw it away. Does it makes sense?

I have another question, also considering the Samsung cells, I have already seen leaked, or burned all other cells - Sony, Sanyo, Panasonic, Chinese nonames, but I have not seen leaked Samsungs and LG cells yet. I know I still have very bad statistics with just about 140 packs, so that cannot be objectively said, that is why I am asking what is your experience in that case? Have you seen leaked Samsung/LG cells?

I have only harvested about 700 Cells, but there was no Samsung or LG which leaked. Only Sony and Sanyo's.

Since when cells actually leak?

The only way they can leak is by physical damage.

I've harvested also around 700 cells so far, and found a few Samsung's also corroded and leaking , but I think they're initiated from a scratch, nick, and then just when I took off the nickel strips, one cell casing tore a bit, and the rest had residue, so I just dumped them. There were quite a few Samsung's as well that had leakage from the top of the cell, between the positive and casing.

I use a lot of packs with low capacity cells, even with 10% capacity remaining. LED systems can do with these too.
I just make sure not to have damaged/rusted cells and I keep an eye on their health over time.

BlueSwordM said:

Since when cells actually leak?

The only way they can leak is by physical damage.

Click to expand...

No, they actually leak sometimes between the plus and the minus pole, as beserker786 described.

Here, fer example are some leaked NCR20700A from Sanyo:




Sometimes it's just a little, then you have to remove the shrink and insulator ring to see it.

If a cell leaks for any reason, toss it. Don't even use it for a flashlight. If you open a pack and get that destination smell, find out which one(s) are compromised and get rid of them.
Electrolyte is very corrosive to steel/iron. Nasty stuff

Korishan said:

If a cell leaks for any reason, toss it. Don't even use it for a flashlight. If you open a pack and get that destination smell, find out which one(s) are compromised and get rid of them.
Electrolyte is very corrosive to steel/iron. Nasty stuff

Click to expand...

Yeah I usually get rid of those and dont even test when I see just a bit of corrosion/leak...

To BlueSword: hehe it seems that you did not open much packs, I had a lot of pack without single mechanical damage with leaked cells. From my recent delivery it was about 10% of pack with leaked cells...

I have had 100s of leaking cells without damage. The cells arent 100% sealed in terms of tight seal. They can leak due to pressure issues or that the seal in the top is faulty.

daromer said:

I have had 100s of leaking cells without damage. The cells arent 100% sealed in terms of tight seal. They can leak due to pressure issues or that the seal in the top is faulty.

Click to expand...

I guess you are not using such cells right? Today I got two Samsungs leaked... It was barely visible at the first moment and than I removed a cover from top to see it clearly >

Satiriasis: not sure what you mean. The cell you're holding (or somebody is) probably leaked between the seal that is the CID. It probably failed or has an imperfect seam and allowed electrolyte leak out. This is what Daromer was referring to. There is no "visible" damage, but there are seams that can be compromised.

Korishan said:

Satiriasis: not sure what you mean. The cell you're holding (or somebody is) probably leaked between the seal that is the CID. It probably failed or has an imperfect seam and allowed electrolyte leak out. This is what Daromer was referring to. There is no "visible" damage, but there are seams that can be compromised.

Click to expand...

From this perspective it is better to see. Would you try to charge and check such cells? I normally throw them out, because of some reason the cell leaked, so I dont trust them.

No i wouldn't. I'd toss it. Daromer was stating that cells that don't 'appear' to have damage can still leak.

we are on the same page, but somehow we ended up on opposite ends of it

Oh god. That's rust.

It's been in contact with water. If it leaked, the positive pole would've been damaged too.

I would FULLY discharge it and then throw it out.

BlueSwordM said:

Oh god. That's rust.

It's been in contact with water. If it leaked, the positive pole would've been damaged too.

Click to expand...

Not always, the electrolyte with the oxide (LiCoO₂) in it will form hydroxide with the water from the air which also accelerate corroding. About the plus pole, I had many cells where the plus was without rust and under the insulation ring was the rust.

BlueSwordM said:

I would FULLY discharge it and then throw it out.

Click to expand...

That's for sure the best but i discharge all my bad cells before I bring them to the recycler.

Thanks guys, you discharging them using some resistor I guess, or is it possible to discharge it on my IMX6 charger with Nimh settings to be able to go below 3V? What is your simplest and fastest method - I do not want to spent hours just doing that...

BTW: Today I got finally package with my Liitokala Lii 500 chargers and I have got interesting results. My Oppus charger shows with my Samsung batteries much higher IR. For example, on one cell that has about 2000mAH it was showing 260mOhm (when I played around with contact, the smallest number I get was about 180), but Liitokala shows just 54mOhm, now the question: "which charger to believe"? ... I am testing the batteries again and for all my Samsungs, the Oppus is showing around 110-200, where Liitokala just 40-110... I have seen here in forum someone who mentioned that Liitokala is much more precise in testing IR, but that guy had just a small differences, not that big as I do have.

Satiriasis said:

Thanks guys, you discharging them using some resistor I guess, or is it possible to discharge it on my IMX6 charger with Nimh settings to be able to go below 3V? What is your simplest and fastest method - I do not want to spent hours just doing that...

BTW: Today I got finally package with my Liitokala Lii 500 chargers and I have got interesting results. My Oppus charger shows with my Samsung batteries much higher IR. For example, on one cell that has about 2000mAH it was showing 260mOhm (when I played around with contact, the smallest number I get was about 180), but Liitokala shows just 54mOhm, now the question: "which charger to believe"? ... I am testing the batteries again and for all my Samsungs, the Oppus is showing around 110-200, where Liitokala just 40-110... I have seen here in forum someone who mentioned that Liitokala is much more precise in testing IR, but that guy had just a small differences, not that big as I do have.
Not so sure about that maybe in mAhs but IR I don't think so

Click to expand...

Discharging.
Build yourself a cheap 20p board and get a good 1? 30Watt Resistor.Wire thick enough to take 5A (AWG 16) to be safe and discharge 20 of them in a safe environment. Maybe a small fan to cool the resistor. No need to attend let it sit a day or so and done.

As far as IR measurements is concerned.
Brand new batteries test the best as they have no foreign nickelartifacts attached which cause a definite increase in measured resistance.
The LiitoKala is much better at IR than the Opus but it also has its limitations.
The Opus is for sure way off all the time just by the instruction manual alone it admits a 10% to 20% error plus you need to deduct30m? from the results. Check hereYou can test the battery 10 times and get different results every time. Who has the time to do that?
The LiitoKala well again it is closerwith brand new batteries but it also fails when testing harvested cells. It is better than the Opus but it has an upper limit of 125m?. I guess they figure anything that high is no good anyway.I insert the cells look at the IR and then unplug and replugthe unit and reboot so to say and there is usually again a different IR reading. So what to do? My solution is to trust a YR11030 tester as a dedicated IR tester and that's it. I can test a battery 20 times with this unit and 20 times I am going to get a reading within .1 to .2 m? of each other.
BTW from this particular vendor it comes with all the bonus attachments. (No affiliation).

Wolf

Satiriasis said:

Thanks guys, you discharging them using some resistor I guess, or is it possible to discharge it on my IMX6 charger with Nimh settings to be able to go below 3V? What is your simplest and fastest method - I do not want to spent hours just doing that...

Click to expand...

I read now many times about the discharging of cells befor thrashing them.

Now, just as i again read about the resistor many people recommend, an idea popped up in my mind to use an oldfashioned light bulb as load. Advantage of them is, that they have a weak constant-current behaviour, that means current does not drop according to Voltage in Ohms law. In effect, the same can be used on any voltage of the cell, will drain quite high currents (depending on the lamps data) plus show when the cell is still alive.

Combined with one of the cheapo chinese sockets its easy to build a discharging station which is quite favourable in winter, it gives heat AND light simultaneously .
I guess a 6V 5W lamp would be quite nice, giving reasinable load at all Voltages. 6V 3W would still have nearly half an Amp.