18650 used/mostly new sitting for 1-2yrs uncharged

[JonCombat]

I am working on building a 14s 102p powerwall using a combination of new/used batteries. Most of the batteries were unused but sitting uncharged for 6months - 2yrs. I am using an xtar vc8plus to charge/test capacity. I find the capacity varies +/- 100mah if you retest the same unit using the VC8plus but it gives a ball park number. I have batteries that range from 2500-2790 mAh with the majority of the batteries being > 2600 mah. I still have a couple of hundred more batteries to test before I can start putting them in packs. After charging I waited a few weeks to check for self discharging batteries. I found some bad one and discarded. Here are my questions:
What would be considerd a "Good" voltage after a few weeks? I am seeing on average between 4.12-4.15V.
Should I discard battereis < 2550mah to make a more robust pack? This is my first attempt at building a power wall.
Once I get close to testing all the batteries I will look into a charge controller/inverter. Any suggestions? I live in the Connecticut /USA and would like to get some solar in the future but it is so expensive unless you DIY your system. Need to put a new roof on my barn before I can even think about it.

 

[bbbbrass]

Hi Jon, great to have you as part of the addiction!

I use the XTar testers as well and have found them to not be the most accurate or consistent, but it does give enough of a picture for overall cel health and capacity to be useful. The IR feature on the XTar is why I chose it, but later I learned it isn't very helpful and a true IR tester is very worth the $50.

I would use any cell that is over 80% of original capacity, and stays above 4.1V after letting it rest for a few weeks. Since you are building 102P (why that number?) you will have plenty of room for variation in the pack between cells. Shoot for healthy IR as well, and I would measure that before you even try capacity testing or charging. When you say the cells were sitting uncharged, what voltage were most of them resting at?

Also, what kind of cells? It makes a difference. Try looking up some of the processing flow charts on the forum, and also look at Wolfs IR charts to know what cells are healthy.

This is a great group, ask away with any questions you have! Better to learn first before buying stuff (which was a mistake I made). To know what charge controller and inverter we need a better idea of what you want this power wall to do. How much power, for how long, in what environment.

 

[OffGridInTheCity]

In my experience (12,000+ cells), after a full charge and a couple of weeks, as @bbbbrass suggests, a resting voltage > 4.1v is fine. I often find 4.15 or 4.13 etc... these are minor differences. Sometimes I don't like this rule and will even close my eyes and accept a 3.95v cell because of all the work I did to get it free and test it - but the goal is to avoid self-discharging cells that can drag down a pack... causing a maint issue / headache down the road. Healthy packs will stay in balance.

My goal is as many thousands of cycles in my Powerwall as I can get, so I tend to discard cells <90% of original capacity - but 80% if perfectly reasonable for a long life of good use. If the battery is for occasional use - say <50 cycles / year - then down to 70% should work for several years as long as you evenly distribute them among the packs. But I've never personally gone below 85% so I don't know this based on hands-on.

I'd recommend grouping cells by 100mah into buckets and then evenly distribute them thru the packs you put in series. You want each pack in series to be very close to the same total ah. A +/- 2% total ah (or less) for each pack in series is my goal.

If you do laptop cells (e.g. cells with 3a or 4a max discharge) then you'll have better results (minimal voltage drop on load, longer life) if you limit the max load to ~0.5a/cell. Let's say you do 14s102p, then that would be 60a @ 48v (nominal) - around 3000w max draw. If you parallel 2 of them then 6000w max draw etc. Even if you use ebike cells with higher max discharge specs such as 10a then for longest life I'd still stick with 0.5a-0.7a/cell design goals to get nice results but you can do higher loads with less voltage drop.

 

[Korishan]

To add to the above comments, getting a decent IR reading of the cells will also help a lot. Get yourself a 4-wire IR meter, don't use the DMM in resistance mode, as it won't give accurate results "of a battery". They need to be tested with AC (alternating current) to get a good reading and also use 4 wires (2 wires for the power, 2 wires for measuring).
I think the one that most here use is the YR1035+. There may be a few other variants as well.
They aren't too bad price wise, and in the end, you'll have a much healthier battery pack.

 

[JonCombat]

Hi Jon, great to have you as part of the addiction!

I use the XTar testers as well and have found them to not be the most accurate or consistent, but it does give enough of a picture for overall cel health and capacity to be useful. The IR feature on the XTar is why I chose it, but later I learned it isn't very helpful and a true IR tester is very worth the $50.

I would use any cell that is over 80% of original capacity, and stays above 4.1V after letting it rest for a few weeks. Since you are building 102P (why that number?) you will have plenty of room for variation in the pack between cells. Shoot for healthy IR as well, and I would measure that before you even try capacity testing or charging. When you say the cells were sitting uncharged, what voltage were most of them resting at?

Also, what kind of cells? It makes a difference. Try looking up some of the processing flow charts on the forum, and also look at Wolfs IR charts to know what cells are healthy.

This is a great group, ask away with any questions you have! Better to learn first before buying stuff (which was a mistake I made). To know what charge controller and inverter we need a better idea of what you want this power wall to do. How much power, for how long, in what environment.

My parallel comination is not set in stone. I picked 102 because I am using rows of 6 cells and trying to maximize the capacity. The battery spec is a little confusing. Standard Discharge 0.2C (500mA) with a max of 10A. Trying to figure out what the actual capacity of the cell will be. I would like a 10KWH system without over charging/discharging the batteries. I plan on spot welding 5A glass fuses on the + side, not sure what to do with the negative terminals. I do not like the idea of fuse wire burning in the open air so that is why I am going with glass fuses. I am also looking for a configuration that will fit inside some sort of encolsure so it will be easy to access and possibly be mobile. In the future I would like to purchase LIFPO4 cells but they are very expensive. so far I spent <$150 usd on my holders/chargers. Once I get enough batteries tested I will figure out the BMS. As of now I have enough to build 10 packs of 102 cells. I need to scrounge around for more cells.

 

[JonCombat]

Hi Jon, great to have you as part of the addiction!

I use the XTar testers as well and have found them to not be the most accurate or consistent, but it does give enough of a picture for overall cel health and capacity to be useful. The IR feature on the XTar is why I chose it, but later I learned it isn't very helpful and a true IR tester is very worth the $50.

I would use any cell that is over 80% of original capacity, and stays above 4.1V after letting it rest for a few weeks. Since you are building 102P (why that number?) you will have plenty of room for variation in the pack between cells. Shoot for healthy IR as well, and I would measure that before you even try capacity testing or charging. When you say the cells were sitting uncharged, what voltage were most of them resting at?

Also, what kind of cells? It makes a difference. Try looking up some of the processing flow charts on the forum, and also look at Wolfs IR charts to know what cells are healthy.

This is a great group, ask away with any questions you have! Better to learn first before buying stuff (which was a mistake I made). To know what charge controller and inverter we need a better idea of what you want this power wall to do. How much power, for how long, in what environment.

I have some specs on the cells. <=60 mili ohms measured at 1kHz The company is Yiyang corun battery co. They supply battery packs for our products. My company had an early spring cleaning and sent old units (mostly new in box) to the trash, these 2 cell/3 cell packs were inside the boxes. The batteries look like they went through a few revisions since the materials used are different between the packs. (kapton tape, shrink tube, insulation materials, battery insulator color....the date codes were from 2019-2022.
Question: To increase or maintain the existing capacity and increase the lifespan of the cells (charge/discharge cycles) could you add 1-2 packs in series, reduce the max charging voltage to 4.05V and increase the low voltage cutoff to >3V. Output voltage range would be 48V min / 64.8V max. Are there any local groups in Connecticut that are DIY battery/solar enthusiests ?

 

[OffGridInTheCity]

My parallel comination is not set in stone. I picked 102 because I am using rows of 6 cells and trying to maximize the capacity. The battery spec is a little confusing. Standard Discharge 0.2C (500mA) with a max of 10A.

Standard Discharge of 500ma = 0.5a - as I was mentioning above will lead to lonest life and low voltage sag.
But don't confuse discharge specs with capacity. A 3000mah cell has 3000mah capacity - regardless of charge/discharge specs. The discharge specs give you a guide as to how fast you can discharge a cell and have less voltage drop and do less damage in terms of long life.

Trying to figure out what the actual capacity of the cell will be. I would like a 10KWH system without over charging/discharging the batteries.

Let's say you have an average of 2600mah/cell and 102cells per pack and 14 in series - that's 14s102p. Each p = 102cells * 2600mah = 265,200mah = 265.2ah. Capacity increases as you increase voltage. So if you have 1s102p, you have nominal 3.7v @ 264.2ah = (3.7v * 264.2v) = 977wh of capacity. If you do 14s that's 48v (nominal) so you have 48v * 264.2ah = 12,681.6wh or 12.7kwh.

Note: Technically that's 12.7kwh at 48v (empty) and 15.5kwh at 58.8v (full charge). Personally, I use 52v for my design calculations on 14s Lithium-ion because it's more representative of the middle of my operating range. 52v * 264.2ah = 13.7kwh When communicating with with strangers and without context, 48v is a common standard.

Next is depth of discharge. Lithium-ion ranges from 3.0v -> 4.2v / cell. But in practical terms the max is more like 4.15v -> 3.4v (discharge curve knee) or ~90%. So take the 13.7kwh number above and multiply .9 (90%) to get 12.3kwh of useable power out of 14s102p at 2600mah/cell.

Each 1/10 of a volt you go down from full charge significantly increases life span. If you limit max charge to 4.1v or 4.0v or even 3.9v you get increasing lifespan but lower DOD and less useable power. Here's a old but interesting page on this - https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries that shows possible life space increase by charging to a lower max.

The point is, you can design your battery for longer life by making it larger than you need and avoiding full charge. For a multi-year, cycle every day powerwall (365 cycles/year) life-span is more important than for an ebike ridden once a week (52 cycles/year).

In your case, a 14s102p at 2600mah/cell would achieve the goal of ~10kwh useable power between 4.0v (max charge) and 3.5v/3.4v (low) and likely give you 1,000(s) of cycles.

FYI - My powerwall design is 14s100p at 260ah/pack. My oldest 14s battery has 1,563 cycles at an average DOD of 36% and no detectable degradation. This subject is near/dear to my heart as I've invested A LOT of effort for low DOD but it will take another 4-5 years before I know what happens at 3,000 cycles and beyond.

 

[JonCombat]

Standard Discharge of 500ma = 0.5a - as I was mentioning above will lead to lonest life and low voltage sag.
But don't confuse discharge specs with capacity. A 3000mah cell has 3000mah capacity - regardless of charge/discharge specs. The discharge specs give you a guide as to how fast you can discharge a cell and have less voltage drop and do less damage in terms of long life.


Let's say you have an average of 2600mah/cell and 102cells per pack and 14 in series - that's 14s102p. Each p = 102cells * 2600mah = 265,200mah = 265.2ah. Capacity increases as you increase voltage. So if you have 1s102p, you have nominal 3.7v @ 264.2ah = (3.7v * 264.2v) = 977wh of capacity. If you do 14s that's 48v (nominal) so you have 48v * 264.2ah = 12,681.6wh or 12.7kwh.

Note: Technically that's 12.7kwh at 48v (empty) and 15.5kwh at 58.8v (full charge). Personally, I use 52v for my design calculations on 14s Lithium-ion because it's more representative of the middle of my operating range. 52v * 264.2ah = 13.7kwh When communicating with with strangers and without context, 48v is a common standard.

Next is depth of discharge. Lithium-ion ranges from 3.0v -> 4.2v / cell. But in practical terms the max is more like 4.15v -> 3.4v (discharge curve knee) or ~90%. So take the 13.7kwh number above and multiply .9 (90%) to get 12.3kwh of useable power out of 14s102p at 2600mah/cell.

Each 1/10 of a volt you go down from full charge significantly increases life span. If you limit max charge to 4.1v or 4.0v or even 3.9v you get increasing lifespan but lower DOD and less useable power. Here's a old but interesting page on this - https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries that shows possible life space increase by charging to a lower max.
View attachment 29001

The point is, you can design your battery for longer life by making it larger than you need and avoiding full charge. For a multi-year, cycle every day powerwall (365 cycles/year) life-span is more important than for an ebike ridden once a week (52 cycles/year).

In your case, a 14s102p at 2600mah/cell would achieve the goal of ~10kwh useable power between 4.0v (max charge) and 3.5v/3.4v (low) and likely give you 1,000(s) of cycles.

FYI - My powerwall design is 14s100p at 260ah/pack. My oldest 14s battery has 1,563 cycles at an average DOD of 36% and no detectable degradation. This subject is near/dear to my heart as I've invested A LOT of effort for low DOD but it will take another 4-5 years before I know what happens at 3,000 cycles and beyond.

Thank you for the fast response. It takes time to find and read all the info out there. Talking to experienced people that have "been there,done that" weeds out a lot potential dangerous and expensive mistakes. Curently am looking into a holder that is 6x17 for 102 cells. It will require an extra bus bar but makes a shorter pack. Finding a safe enclosure to insulate the bus bars/fuses and possibly circuilate air to help keep the batteries cool would be nice. Has anyone found off the shelf encolsures that are inexpensive to cover these battery packs?

 

[OffGridInTheCity]

Thank you for the fast response. It takes time to find and read all the info out there. Talking to experienced people that have "been there,done that" weeds out a lot potential dangerous and expensive mistakes. Curently am looking into a holder that is 6x17 for 102 cells.

A very common (and less expensive) 18650 cell holder is 4 x 5 but you can buy a range of sizes like this - https://batteryhookup.com/products/18650-cell-holders?_pos=2&_sid=d9807a15f&_ss=r You could do four 4x6(s) + two 1x3(s) to get 6 x 17.
You can also cut 4 x 5(s) (or other sizes) leaving an original edge to snap onto the base until you get 6 x 17.

t will require an extra bus bar but makes a shorter pack. Finding a safe enclosure to insulate the bus bars/fuses

Not sure what you mean exactly to insulate the bus bars/fuses - I don't find insulating bus bars necessary. Maybe a pic or drawing of what you have in mind would help. They do make large size shrink-wrap (would fit around 6 x 17 pack) that some have used around their packs.

and possibly circuilate air to help keep the batteries cool would be nice. H

I also 'worried' about cooling cells - but at a powerwall level (e.g. lower stress / <= 0.5a/cell) they don't even get warm - not necessary to worry about this in my opinion unless you're load (or charge) is >1a/cell continuously.

as anyone found off the shelf encolsures that are inexpensive to cover these battery packs?

The "inexpensive" is difficult I find that 14s (cells horizontal) packs side by side come up just a little less than 48inches. So you'll need shelves or boxes or cabinets at least 48 inches wide and appropriately deep for you're packs.

 

[100kwh-hunter]

What would be considerd a "Good" voltage after a few weeks? I am seeing on average between 4.12-4.15V.

If you charge them at 4.2, put them on a shelf for four weeks, all below 3.9 i consider as a self discharger.
But i found out that some cells stop droping v at 3.9 even after waiting a extra month.
So where is this going to lead me?

My bms is set to max v of 4 volt, and only on a rear occasion we have one month no sun, but they will charge anyway back up.
As others stated before, if the IR is good i see no reason to discard of them.

Little side note.
Those cells have there own pack! added in the 100p pack.
That pack is checked every year.
But all my cells have 1a fuses, so if they self discharge to much the fuse will blow.

With this in my mind i was ready to put a extra panel set on the roof to cover up the losses.
And a extra set for the inverter.

From what i understood the cell self life is approx 15 years. concidering you have them second handed, you have 10 years left.
with 1000(80%dod) you will be fine

 

[100kwh-hunter]

I also 'worried' about cooling cells - but at a powerwall level (e.g. lower stress / <= 0.5a/cell) they don't even get warm - not necessary to worry about this in my opinion unless you're load (or charge) is >1a/cell continuously.

Plus one.

The more you draw the warmer they can get, but when the pack is empty, and how qiuck?
Most of us when they put there PW in to use, most have 100p packs of ~250 to 300 Ah
Most wil not hit 1A discharge per cell, or your name must be igor and have a shed full of machinery.....
For normal house holds you mostly will hit indeed lower than 0.5 if the freezer is jumping on.
Or electric coocking.
But how long wil you draw 1?A per cel times 1400 cells?
1400*50v=7kwh, that is one big cooking stove.

Make sure that ALL youre connections are true, check the mechanical ones, once a month.
Those will create hotspots.
Make your pack wire also to the max size, in that way you keep them cool.

If a cell gets hot (above 50C) it is at the end of his life.
If a terminal or o connector gets hot and the others not, check.

If you feel a slight rise in temp, check
If the temp feels comfortable 40C check
If the temp feels warm 50c take action soon
If the temp feels hot 60C take action.
If the temp feels uncomfortable to the touche 70C, take immediately action.

Yes you can train your hands to feel temps, but only when there is room enough to feel

With best regards Igor

 

[paddy72]

I'd like to second Korishan for testing the IR with a decent 4-wire 1kHz-tester. Try to put similar IR-cells into the parallel packs and check that each pack of 102p cell (e.g.) has a simalar IR alltogether. 2. Pay attention to get very close capacity ratings over all packs in series. There are free calculators/tools available to select the cells that each pack has very close capacity.

If a cell drops from 4.2 to 4.1 in 3-4 weeks that well o.k. imho. I prefer to test from a lower level, after the final voltage has settled for at least 24 h.
Some cells drop 50mV from 4.2 in 10 hours, some drop even to 4.1 in 15...20 h but may rest there for the next weeks. So take the first reading after settling for 24h and than a second reading 30 days later - thats my way and its more consistant imho. After 30 days they should not drop more than 50 mV, maybe 60 or 70 if you like. Really good cells drop just a few mV/year at cool temperature (cellar).

 

[JonCombat]

Thank you, very informative. What IR tester/charger/capacity tester do you recommend. Hopefully it will not break the budget! I will probably charge up my cells and run the 24hr/30 day test before starting to put them together. I am still in the testing phase. I am using a VC8plus for capacity which I cycled some units through and the results do vary. The issue is the capacity test takes about 8+ hrs to complete for four (4) batteries. I have two of them so If I put fully charged batteries on the test side once the discharge completes I record the capacity and move them to the charge side and start over with new fully charged cells. Very time consuming but I have made it my mission to complete the testing. It has not become a ritual!

 

[Wolf]

What IR tester/charger/capacity tester do you recommend.

RC3563 or YR1035+

IR Meter Recommendations

Does anyone have a recommendation for a good meter to check IR (Internal Resistance) of various battery types? I haven't checked this in the past as the testers I use (Opus) aren't very accurate. There have been a lot of requests to test this in some of my videos, so I'd like to start doing so...

secondlifestorage.com