Li-ion cells cycle ageing

[vanturion]

IMO the constant for current discharge should be a selected C rate, not in amps.

:thumb:

 

[vanturion]

added 5C discharge voltage drop graph (30.05.2021)

This chart is so cool Pajda, it definitely deserves a bump, especially for those who aren't yet clued in to check your capacity test updates on page 20. HG2 is the lifespan king among the ~3000mAh cells, it's not even close! It's interesting to compare docware's 1C tests on page 17 comparing these 3 cells too, you can infer some lifespan behavior with his data on DCIR creep over cycles, but you don't get nearly the whole story with the massive difference in voltage sag under high load and total area under the curve at higher cycle counts like you do with your graph above. While neither test case is representative of real world loads/conditions of a ebike pack in use, seeing both testing extremes is ideal so again big thank you for sharing.

Speaking of telling the story, it's crazy how the VTC6 dips immediately to 3V before recovering somewhat, like taking a gut punch and getting the wind knocked out for a while before meagerly finishing out the round. Talk about voltage sag. Then with 30Q it's like a 90 lb'er that immediately takes a shot to the face and rolls around on the ground for the rest of the round, this at "only" 750 cycles too. Compare these to the seasoned fighter HG2 still packing punches at round 12, yeah HG2 dominates no question.

I've also been watching for your P42A updates too, and it doesn't look too far off the manufactures specs, although not amazing results so far. Kind of an aside, but yesterday I saw someone online posting a battery pack build using the Lishen 21700-LR2170SF. I'd initially ignored the brand not recognizing it, but the specs on paper are better suited (capacity) for my bike than the Molicels at the 12A/cell limit I'd run. Any experience or thoughts on these? I'm thinking I might just have to order one for cycling and put off my build a little longer.

 

[vanturion]

Adding my results from the Samsung 25R Capacity test using EBC-A20 Tester and ZKE Battery Rack. The cell I tested was unused having always sat at room temp, and the storage voltage read 3.59V at the beginning of the test sequence. Manufacturing code indicates September 2014 origin.





Some notes:
-1st cycle capacity measured 2445 mAh. Considering the original spec rates these cells at a minimum of 2450 mAh at 10A discharge, this seems like a great result for a 7 year old cell.
-Samsung spec quantifies a "typical" 25R cell's initial capacity at 2539 mAh so while I don't know what capacity it would have had in 2014, it's very likely the capacity loss was less than 100 mAh in 7 years.
-For a sanity check I also tested 2 other unused cells from the same batch and got 2434 and 2452 mAh initial cycle capacity.
-I attempted DCIR 10s and 30s measurements, but the calculated results were not inspiring in terms of accuracy so I've left those results out (could have also been influenced by room temperature swings as shown in picture above).
-Edit: 77 day pause (red arrow) at 225 cycles before resuming testing. Interesting that there is no capacity recovery like with the LR2170SF, but the opposite. Capacity holds steady for about 25 cycles before resuming original trend line. Odd.



In comparison to the 4A Charge/20A Discharge spec, the test cell exhibited more of a linear decay curve with less capacity loss. Not an unexpected result considering the 1.5A Charge/10A Discharge is pretty tame by comparison. I think what was most interesting from this test was how well the cell tested considering its age.

 

[Pajda]

Small update. The Samsung 30Q6 sample begins to die after 450 cycles at 5C. Nevertheless, I would like to call him the new king of this category.

I would also like to announce some news. First is that Nkon finally stocked the Samsung INR21700-40T3 on this friday and so its sample will enter the competition right after the final death of 30Q6. So we can look forward to the duel of P42A and 40T3, where a similar course can be expected as between 30Q6 and HG2 in the lower weight category.

Finally, I managed to order from Vapcell few samples of one of the most badass cylindrical cell today, the Samsung INR21700-50S. If it doesn't get lost in transit, it could also appear in the 5C chart during this summer.

 

[Talon]

Do you know anything about "Graphine Tech? It is rumored to replace Lithium Ion Technology?

 

[thundercamel]

-I stopped the test early at 225 cycles to begin cycling a Lishen LR2170SF cell when it arrived, but will resume to 70% capacity after I have enough data from the Lishen test to make a purchasing decision.

I'm interested in your Lishen cycle life results, though the LR2170SD suites me more than the LR2170SF for maximum capacity. At least they're available on LiIonWholesale as opposed to Samsung 50E or LG MJ1 / M36.

 

[Dak77]

-I stopped the test early at 225 cycles to begin cycling a Lishen LR2170SF cell when it arrived, but will resume to 70% capacity after I have enough data from the Lishen test to make a purchasing decision.

I'm interested in your Lishen cycle life results, though the LR2170SD suites me more than the LR2170SF for maximum capacity. At least they're available on LiIonWholesale as opposed to Samsung 50E or LG MJ1 / M36.

Yeh definitely, since prices and availability of Samsung & LG cells are becoming more painful every day it seems. A good low cost competitor is sorely needed. I lucked out and grabbed 100 50e while they were still under $5 , but they're now over $7 pretty much anywhere in the states.

 

[vanturion]

Small update. The Samsung 30Q6 sample begins to die after 450 cycles at 5C. Nevertheless, I would like to call him the new king of this category.
...

Pretty impressive improvements, I especially like how the initial capacity loss isn't as severe as everything else in the first hundred cycles. Only scenario I'd probably go for the HG2s over 30Q6 is for a budget commuter seeing regular use considering the 1000 cycle lifespan, if you could even find them for a decent price.

Nice, looking forward to more 21700 comparisons.

 

[vanturion]

I'm interested in your Lishen cycle life results, though the LR2170SD suites me more than the LR2170SF for maximum capacity. At least they're available on LiIonWholesale as opposed to Samsung 50E or LG MJ1 / M36.

I'll post up a couple charts when I hit 200 cycles, about 50 more to go now.

The price-to-performance level is really attractive, especially at 3.85/cell with 18650batterystore in bulk as opposed to the premium prices the well-known manufacturers are charging as Dak77 said. This is where I got my test cell btw. Here's some thoughts I can share in the meantime:

-There's a great amount of variation in the specs including the min vs rated capacity, the individual cell weight, and most importantly to me, in the cell dimensions including the diameter. This was not encouraging, but it was the cell diameter + tolerance above 21.7mm that was the deal-breaker for me. The cell would not fit 2 types of ABS plastic holders that I intended to use without removing the orange cell wrapping and I did not like that. If it weren't for this loose manufacturing tolerance, I would be building my new pack with these.

-My cell tested well with about 4400mAh initial capacity, and acceptable end temp < 64C after full 10A discharge to 2.5V.

-Capacity decay is very similar to the results Pajda shared for the Molicel P42 in the first 100 cycles.

 

[Pajda]

It is a good point, it seems that at least for next one year we have to look on other, mostly chinese production.I have good experience with K-tech cells, but this company is now also fully occupied with production for large customers. So other potentially interesting manufacturers are Lishen and BAK. I hope that Nkon will soon receive a shipment from BAK and I will definitely start testing N21700CG as 5000mAh HE competitor and then some samples of size 18650, capacities of 2900 and 3350 mAh. Unfortunately graphs from these tests of mine will not be public, at least in the near future. I'll just leave comments.

 

[Dak77]

It is a good point, it seems that at least for next one year we have to look on other, mostly chinese production.I have good experience with K-tech cells, but this company is now also fully occupied with production for large customers. So other potentially interesting manufacturers are Lishen and BAK. I hope that Nkon will soon receive a shipment from BAK and I will definitely start testing N21700CG as 5000mAh HE competitor and then some samples of size 18650, capacities of 2900 and 3350 mAh. Unfortunately graphs from these tests of mine will not be public, at least in the near future. I'll just leave comments.

Nice , thank you sir

 

[Dak77]

I'm interested in your Lishen cycle life results, though the LR2170SD suites me more than the LR2170SF for maximum capacity. At least they're available on LiIonWholesale as opposed to Samsung 50E or LG MJ1 / M36.

I'll post up a couple charts when I hit 200 cycles, about 50 more to go now.

The price-to-performance level is really attractive, especially at 3.85/cell with 18650batterystore in bulk as opposed to the premium prices the well-known manufacturers are charging as Dak77 said. This is where I got my test cell btw. Here's some thoughts I can share in the meantime:

-There's a great amount of variation in the specs including the min vs rated capacity, the individual cell weight, and most importantly to me, in the cell dimensions including the diameter. This was not encouraging, but it was the cell diameter + tolerance above 21.7mm that was the deal-breaker for me. The cell would not fit 2 types of ABS plastic holders that I intended to use without removing the orange cell wrapping and I did not like that. If it weren't for this loose manufacturing tolerance, I would be building my new pack with these.

-My cell tested well with about 4400mAh initial capacity, and acceptable end temp < 64C after full 10A discharge to 2.5V.

-Capacity decay is very similar to the results Pajda shared for the Molicel P42 in the first 100 cycles.

Yeh the dimension variances are not cool. It sucks to design an enclosure only to find that some cells of a particular model will fit while others won't.

 

[Pajda]

Last week, both new Samsung 50S (from vapcell) and 40T3 (from Nkon) 21700 competitors arrived, together with 48X and 50GB which are also very interesting cells but from different category. I performed an initial test in the temperature chamber on a precision tester, you can see the results in the updated table. If all goes well, both competitors will start their 5C cycle life tests in early August.

 

[vanturion]

Lishen LR2170SF cycle results:


Unwrapped to check fitment in cell holder after initial cycle.




I thought it'd be cool to show the performance decay every 100 cycles.

So ideally I'd be running this test at 3C discharge or 13.5A per max rated discharge per manufacturing spec, but my battery tester is only rated to 10A so not ideal testing conditions. Capacity decay trend was still clearly established though. Also, there's a small recovery/blip at around 105 cycles - I paused the test for a day or so here so that probably influenced results slightly.

Yeh the dimension variances are not cool. It sucks to design an enclosure only to find that some cells of a particular model will fit while others won't.

Definitely. Normally I'd 3D print holders to account for the larger cell diameter, but I can only print PLA somewhat unreliably currently and for how close I need to pack the cells (22.5mm spacing) in the rectangular arrangement I prefer, PLA would be a bad choice IMO.

 

[vanturion]

Last week, both new Samsung 50S (from vapcell) and 40T3 (from Nkon) 21700 competitors arrived, together with 48X and 50GB which are also very interesting cells but from different category. I performed an initial test in the temperature chamber on a precision tester, you can see the results in the updated table. If all goes well, both competitors will start their 5C cycle life tests in early August.

Nice, it'll be interesting to see how these stack up to the P42!

 

[Pajda]

Lishen LR2170SF cycle results:

Nicely done :thumb: So do not waste time and order another EBC-A20 tester with cell fixture immediately! :wink:

 

[thundercamel]

Thank you very much Vanturion!

 

[vanturion]

No problem thundercamel, happy to share the results here.

Nicely done :thumb: So do not waste time and order another EBC-A20 tester with cell fixture immediately! :wink:

Thanks - haha well... now that I was finally able to make a decision on which cells and satisfy my curiosity with my old unused 25Rs, I basically accomplished everything I wanted to with the tester until I need a new pack hopefully many years down the line. Something did come up though with the Lishen cycles that I want to test with the extra Molicel P42 I ordered though.

I noticed a significant stall in the capacity decay rate after resuming cycling after a period of 24+ hours before resuming linear or near linear capacity decay:

Cycle 103 - 3.864
Cycle 104 - 3.861
Cycle 105 - 3.858
Break 24 hours
Cycle 106 - 3.864
Cycle 107 - 3.865
Cycle 108 - 3.864
Cycle 109 - 3.865
Cycle 110 - 3.863
Cycle 111 - 3.861
Cycle 112 - 3.856

I know this capacity "recovery" behavior is probably not new information to everyone here, but I'm interested enough to commit the extra P42 cell to a 1 cycle/24 hour test to see how this capacity decay rate would stack up to the results you found. This would also better mirror my own usage as the most use my bike ever sees is 1 cycle/day. Planning to start this test once I'm finished with the Lishen cell.

 

[Pajda]

small update: Samsung 30Q6 was declared dead after 700 cycles, Samsung 50S at 5C (25A) melted plastic spacer in my "strongest" BF-2A cell fixture with huge gold plated terminals which has been declared to withstand 30A now it is fixed and the test continues. Samsung 40T3 started cycle life test yesterday. Nkon seems to have found some amount of Sony/Murata US21700VTC6A cells. I'm thinking of adding samples to the test after the Moli P42A will finish. The other option is Moli P28 or Samsung 25Rx.

 

[savaoaknyc]

Hi Guys,

any experiences with LR2170SD?I plan to buy 100 of them for some packs.I might be able than to run some cycles.I have ebc-a20 gen1 Can we attach them to raspberry pie or it is windows only?

GReets Sava

 

[From-A-To-B]

Hi Guys,

any experiences with LR2170SD?I plan to buy 100 of them for some packs.I might be able than to run some cycles.I have ebc-a20 gen1 Can we attach them to raspberry pie or it is windows only?

GReets Sava

I just bought 52 2170SD cells for a new battery. I was planning on building the pack with Samsung 50E cells, but the price of those cells has climbed since the beginning of the year. Based on posts from this thread and some comments on Reddit, I figured I'd give Lishen a try. I have no test equipment so I don't think my experience with the cells will be all that helpful, but I thought I'd comment to let you know about my order.

The 2170SDs just came back into stock at 18650BatteryStore about two weeks ago. Given that they were out of stock, I'd say there's interest in these cells.

 

[Pajda]

Hi Guys,

any experiences with LR2170SD?I plan to buy 100 of them for some packs.I might be able than to run some cycles.I have ebc-a20 gen1 Can we attach them to raspberry pie or it is windows only?

GReets Sava

ZKETech software is windows only. Lishen LR2170SD look very good according to the datasheet and especially the prices. I have no experience because, together with the BAK 21700CG cells, they are not yet reasonably available in the EU. However, let's hope that Nkon will soon have them in stock.

This week, unfortunately, I was hit by a technical glitch when the hard drive in the tester control PC died :? . So i lost the last 50 cycles in the running tests and it will take a while for me to replace the pc to continue with testing.

 

[Pajda]

The testers are running again. I don't know exactly how many cycles out of 50 the individual cells had before the crash of the PC. For simplicity, I have continued since the next fifty. The fact that the latests results are a bit erratic is mostly caused by the actual weather. The PC crashed in full summer and now it is cold and rainy, but not enough to start with room heating which will again stabilize the results. But at 5C discharge test it is not significant.

It seems that Nkon will finally stock 21700 cells from Lishen and Sony.

 

[Pajda]

Finally! Lishen LR2170SD and BAK 21700CG samples are now running nominal capacity test in precision battery tester. Samples of Sony US21700VTC6A are on the way and I expecting them till Friday. Unfortunately, the tester's channels are never enough, so Sony will have to wait for his 5C torture until the P42A finished 1000 cycles.

 

[savaoaknyc]

Hi Pajda,
i am about to buy very soon Lishen .How is the internal resistance between the cells, is it the same or?