How much capacity loss is ok?

Wattsup

Wattsup

Member
Joined
Feb 19, 2017
Messages
512
Hi all,

Given most of us are using recycled cells, I'm curious to know what amount of capacy loss / degradation do you consider acceptable for power wall use.
I know the chargers most of use are not the most accurate instruments out there but just to get an idea..

My powerwall will be using Samsung 30Q only at this stage and with these my cut off is 2700 mah.

Where do you draw the line?
 
I think most people, including myself, aren't looking at capacity loss. We're looking at remaining capacity, what ever it is, as long as it's above a certain mark. I think most are aiming for 2000mAh and above. Some have set their bar at 1800mAh, others at 22-2400mAh, and some even as low as 1500mAh.
So, I think that's the primary determining factor.

However, I do understand the sense of Original Capacity vs Current Capacity. If a cell was rated at 3000mAh, and it's now 2000mAh, then it's seen a hard/long life as 1/3 of it's gone already. But, that doesn't mean it'd not last a long time in the powerwall application as it'd be running very gently and those last 2000mAh worth of life could last years.
 
Korishan said:
I think most people, including myself, aren't looking at capacity loss. We're looking at remaining capacity, what ever it is, as long as it's above a certain mark. I think most are aiming for 2000mAh and above. Some have set their bar at 1800mAh, others at 22-2400mAh, and some even as low as 1500mAh.
So, I think that's the primary determining factor.

However, I do understand the sense of Original Capacity vs Current Capacity. If a cell was rated at 3000mAh, and it's now 2000mAh, then it's seen a hard/long life as 1/3 of it's gone already. But, that doesn't mean it'd not last a long time in the powerwall application as it'd be running very gently and those last 2000mAh worth of life could last years.
Click to expand...

Interesting.

I am definitely looking at capacity loss, I personally wouldn't use a cell that has lost 1/3 capacity or even close to that.

I guess it can depend on the amount of cells you have to work with as tohow choosy you get.

I do process around 2-300 cells per week, so can afford to cherry pick :D
 
Yep, that's one way to look at it ;)
I'm sure others will chime in, such as Pete and Daromer as they have really large packs/strings.
 
I dont look at capacity loss either. I have yet not seen any chart or proper test shown that after x loss the battery will drop faster. Most of the ones I have seen have had a linear capacity loss down to and below 50% left.

Due to this I dont care much about it. I have set my limit to 2Ah. Though I only test 2-3Ah cells. I have not gotten anything above that.
I process 80 per day and I couldn't bother checking everyone's original capacity and calculate it :) Im instead running such a big bank that if one cells drops faster thats not a problem.

The only problem if so is if 1 cell starts to self-discharge rapidly. But as of now I dont have any proof that capacity left is related to self-discharge. If anyone find any scientific datasheet or test please share! I have not been able and I have read through many of them so far without luck.


If I were to process 10x what i do today then i would most likely cherry pick alot more and only pick the ones that are close to its original.
 
I'd say the lowest I'd go is 50% capacity. At that point, it just goes into DIY powerbanks or flashlights.
Because while modern cells do lose capacity, their internal resistance doesn't rise much, if at all.
That means that even very low capacity 18650 cells, they still have higher energy density than NiCD/Lead Acid batteries, which means they can still be very useful in powerwall applications such as ours.
 
daromer said:
I dont look at capacity loss either. I have yet not seen any chart or proper test shown that after x loss the battery will drop faster. Most of the ones I have seen have had a linear capacity loss down to and below 50% left.
Click to expand...

I just read somewhere that capacity loss indeed is linear and that their isn't a falling-off point. I think it was a page that was comparing lithium to lead acid, and that was one of the pros of lithium.

Of course, my cells are unused, but just recently I commented on a thread about battery sizing, and derating for aging in sizing calculations was a point that came to mind. So I would think that one's power budget vs. capacity would be a main point in the decision for a cutoff, as well as how much hardware one wants to assemble to achieve a capacity that fits the budget. In which case, it becomes simply a math problem.

I'm not sure that I expressed that clearly. Energyrequirements would have probably been a more accurate term than power budget.

No sleep last night. zzzzzzz
 
skyfridge yeah I aggree. I stick to 2Ah because lower than that and its to much work per capacity in my case.

I have heard rumours that there is a higher likelyness off issues with capacity left less than xx% but have yet been able to confirm those numbers or rumour :)
 
daromer said:
skyfridge yeah I aggree. I stick to 2Ah because lower than that and its to much work per capacity in my case.

I have heard rumours that there is a higher likelyness off issues with capacity left less than xx% but have yet been able to confirm those numbers or rumour :)
Click to expand...

2Ah for 30kWh! Holy cow! I have 5Ah, and I'm overwhelmed when I'm only aiming for 4kWh.

It has been slow getting started. I don't know how you all do it. I'm still trying to get going on soldering the contacts.
 
I would just bunch the cells that have like 1/3 capacity loss together in a pack or two. That way if they fail (higher chance then with healthier cells) you will have those all together and your healthier cells can just run. Less maintenance work.

daromer said:
skyfridge yeah I aggree. I stick to 2Ah because lower than that and its to much work per capacity in my case.

I have heard rumours that there is a higher likelyness off issues with capacity left less than xx% but have yet been able to confirm those numbers or rumour :)
Click to expand...

Isnt it more like no one really knows how much life the cell has left in it after the manufacturer released life cycles of like 500 or whatever (are there even some with 1000?)
Then of course there is the internal resistance. Older cells with more cycles and less capacity will have a nigher internal resistance and coudl cause problems (well not with our current draws but in theory right)
 
I am surprised that capacity loss isn't given much consideration by some of you guys.

However this whole powerwall thing is not really an exact science... and that's part of the appeal really, you're project and the way you build is your own choice!

If I'm not mistaken there are probably not many builds out there more than a couple of years old to know what is best practice.

For me if I'm putting the time and effort into putting it all together, I want to use the best possible cells I can!
 
Yes its a good question but I have given that alot of consideration and its in fact simple. I did spend hours and hours reading big scientific documents and not 1 of them could for sure state that the degradation was anything but linear. So 2 cells with same amount of capacity left even though 1 was with 2500 from start and the other was with 3000 and same linear number of cycles they would continue to age at same rate.

Untill that is proven otherwise I will continue to run my systems. And no none of us have had it running long enough nor have proper fact for it. Neither less if we take in account the extra time it takes to do those extra consideration when building the packs that makes up for keeping all cells instead in my case.

With that said I can build 2 strings at same time as i would spend doing more intense tests and at same time i would most likely have 2x the amount of cells not testing it more. So in long run yes i might end up with uneven packs due to above issue but in short term i end up with 2x the capacity and half the wear = should potentially last me longer.

My oldest packs have been running for almost 2 years now without sign of issues :)

Hope that gives you the answer in terms of the consideration I made :)

The-J-Man: You can potentially read out the cycles from the laptop packs but is it worth it? Dunno. My end goal is that if I can stress them less with keeping more cells in the pack that will out grow the other factors :)

skyfridge: There are no 18650 with 5Ah. max is somewhere around 3.5+
 
The-J-Man said:
I would just bunch the cells that have like 1/3 capacity loss together in a pack or two. That way if they fail (higher chance then with healthier cells) you will have those all together and your healthier cells can just run. Less maintenance work.

daromer said:
skyfridge yeah I aggree. I stick to 2Ah because lower than that and its to much work per capacity in my case.

I have heard rumours that there is a higher likelyness off issues with capacity left less than xx% but have yet been able to confirm those numbers or rumour :)
Click to expand...

Isnt it more like no one really knows how much life the cell has left in it after the manufacturer released life cycles of like 500 or whatever (are there even some with 1000?)
Then of course there is the internal resistance. Older cells with more cycles and less capacity will have a nigher internal resistance and coudl cause problems (well not with our current draws but in theory right)
Click to expand...


Manufacturer spes forlife cycles indicate the minimum number of life cycles that one should expect from a new cell that wilbe used underthe prescribedof the spec before the usable capacity falls below a certain percentgage (usually 80%) of the rated capacity. When a box of cells leaves the factory, the cells in the box won't all hAve identical capacities, due to tolerances, possible flaws, etc. Damnedlies and statistics.

Does anyone know the life expectancy of any cell. Is anything a certainty? No, but the available data should give a good indicator.
 
But they arent 18650 ;)

32650 is alot bigger
 
You must log in or register to reply here.
Back
Top