Li-ion cells cycle ageing

[multifrag]

Great work Multifrag :thumb:

I can confirm that your test results are valid. Your setup, methodology and test equipment are correct. Do I understand well to the yellow course "100% Dis. Datasheet" that it is a simple linear extrapolation of datasheet value, where the cycle life should be >70% of initial nominal capacity after 500 defined cycles?

Only small and not essential remark to the presented test values. I think that EBC-A20 can set 100 mA as the lowest value of current. Datasheet says that you should charge to 50 mA cut-off for nominal cycle. It is a good practice to state this value as well.

Thank you Pajda,

Yes, the yellow line is just an extrapolation of the datasheet in linear fashion.

Good catch about the cut off, it's been 2 months since I read the datasheet... Forgot about this limitation completely. I've adjusted the post.

 

[Pajda]

I've adjusted the post.

Sorry for bothering you, but the cut-off current is in mA, not mAh. :wink:

But more importantly I would like to present your work as excellent example of how easily with more than satisfactorily accuracy you can verify your favourite cell parameters. I think that 120-150 USD for a set of EBC-A20 tester with BF-2A fixture in not that much. Everyone has an old pc today for running the software

..Or of course you can continue to listen the old wives' tales, like those on BU.

 

[multifrag]

I've adjusted the post.

Sorry for bothering you, but the cut-off current is in mA, not mAh. :wink:

But more importantly I would like to present your work as excellent example of how easily with more than satisfactorily accuracy you can verify your favourite cell parameters. I think that 120-150 USD for a set of EBC-A20 tester with BF-2A fixture in not that much. Everyone has an old pc today for running the software

..Or of course you can continue to listen the old wives' tales, like those on BU.

Adjusted the post again. Thanks for pointing it out . I've spent a 80usd on EBC-A20 and 45usd on the BF-2A fixture that was capable of 20A. All in all it cost me 125usd and it is worth every cent. The amount of time and manual labor this tool has saved me is astounding. The only thing that I would suggest is hooking it up to a laptop. This saved the data couple of times when the power went out in my case.

The next cell I want to test is K-Tech INR18650P. As it's extremely cheap and would suit my UPS battery rebuild very well.

 

[Pajda]

The next cell I want to test is K-Tech INR18650P. As it's extremely cheap and would suit my UPS battery rebuild very well.

Teaser warning: I believe that you will be pleased with its performance (25HCCG model), particularly in cycle life. :wink:

 

[Pedrodemio]

I've adjusted the post.

Sorry for bothering you, but the cut-off current is in mA, not mAh. :wink:

But more importantly I would like to present your work as excellent example of how easily with more than satisfactorily accuracy you can verify your favourite cell parameters. I think that 120-150 USD for a set of EBC-A20 tester with BF-2A fixture in not that much. Everyone has an old pc today for running the software

..Or of course you can continue to listen the old wives' tales, like those on BU.

Padja, do you recommend a tester that can do more than 5 A charging and doesn't break the bank? I want to play with fast charging profiles and thermal control in the future

 

[Pajda]

Padja, do you recommend a tester that can do more than 5 A charging and doesn't break the bank? I want to play with fast charging profiles and thermal control in the future

Good question, yes there is a solution from ZKE model EBC-A40L with internal 35A charge and 40A discharge current, but it is way more expensive, about 260 USD is best price. In this price range you can find some better modeler battery charger (Cellpro Powerlab 6) which can also do cycle test with logging function. But personaly i do not trust them and I am prefering the "standard" battery cycler data output format generated by the ZKE, so you do not need to mess with tons of data.

 

[Pedrodemio]

Good question, yes there is a solution from ZKE model EBC-A40L with internal 35A charge and 40A discharge current, but it is way more expensive, about 260 USD is best price. In this price range you can find some better modeler battery charger (Cellpro Powerlab 6) which can also do cycle test with logging function. But personaly i do not trust them and I am prefering the "standard" battery cycler data output format generated by the ZKE, so you do not need to mess with tons of data.

Thanks, it's not that bad

If this is to be trusted can be had by around $150

https://www.aliexpress.com/item/4000465060713.html?spm=a2g0o.productlist.0.0.39a1c515QIsXAf&algo_pvid=53a00069-433e-44bb-8cf2-a17666e54be9&algo_expid=53a00069-433e-44bb-8cf2-a17666e54be9-1&btsid=0bb47a1a16052917948615783ef6be&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

 

[Pajda]

If this is to be trusted can be had by around $150

https://www.aliexpress.com/item/4000465060713.html?spm=a2g0o.productlist.0.0.39a1c515QIsXAf&algo_pvid=53a00069-433e-44bb-8cf2-a17666e54be9&algo_expid=53a00069-433e-44bb-8cf2-a17666e54be9-1&btsid=0bb47a1a16052917948615783ef6be&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_

Did you check shipping cost? This item is heavy and so the shipping cost is high. The price 260 USD which I posted is a good price with shipping included.

 

[Pedrodemio]

Did you check shipping cost? This item is heavy and so the shipping cost is high. The price 260 USD which I posted is a good price with shipping included.

Didn't, but it's $170, no deal, but shipping of anything is ridiculously expensive to get it here, I'm not on the US

 

[docware]

I checked precise 4 wire voltmeter (see page 13) on PDVS2mini precision digital voltage source. To my surprise accuracy and linearity of this cheap chinese voltmeter is unbelievable. In the range 0,5 - 4,3 V absolutely accurate without any deviation. Bellow 0,5 V stable offset -0,1 mV. In the range 4,4 - 10 V, after switching to lower resolution, stable offset -1 mV.
Ambient temperature 23°C, 30 - 40 min warming-up, internal supply 6 pcs NiMh AA, U = 8 - 8,4 V.

PDVS2mini calibration record: maximum voltage output error is ± 4 μV, calibration instrument DMM KEYSIGHT 3458A.

 

[john61ct]

excellent!

any purchase sources more reasonable than direct?

don't mind waiting. . .

 

[docware]

any purchase sources more reasonable than direct?

Second round of testing : in the range 0 – 4,3 V is accuracy ± 0,1 mV, range 4,3 – 10 V again stable offset -1 mV.

In my opinion 6,- USD for voltmeter which has accuracy 0,004 % at 2,5 V (0,0024 % at 4,2 V) is very reasonable.

 

[Pajda]

Great news, docware is back! By the way stylish blanket :wink:

 

[Stealth_Chopper]

:thumb:

 

[john61ct]

In my opinion 6,- USD for voltmeter which has accuracy 0,004 % at 2,5 V (0,0024 % at 4,2 V) is very reasonable.

Of course! I only found sources selling for $200+

 

[docware]

Manufacturer´s Ali-shop (RD tech, or also Hangzhou Ruideng Technology Co., Ltd) :

https://rdtech.aliexpress.com/store/923042

Go to Products/DC Digital/DC Voltmeter 5 Digit

 

[Pajda]

docware can we expect more of your excellent test data soon? :wink:

I am now at 9th month of calendar life test (direct comparison of storage at 100% and 50% SoC at RT) also have initial results of "torture test" of High Power cells HG2, 30Q ant VTC6 at 5C(15A) continuous discharge at 100% DoD.

 

[docware]

Inspired by your testing I also started calendar aging during summer this year. Twelve pairs of the cells are relaxing at room temperature, always one at 100% and one at 50 % SOC :

LG M36, M29, HG2, MJ1, M50T
Samsung 35E, 29E6, 30Q, 50E, 29E7
Sanyo GA
SONY VTC6

Each 100 days are cells checked for capacity and DCIR.
Currently 10 pairs of the cells passed the first 100 days, two remaining pairs will achieve 100 days in the first half of the January 2021.

 

[john61ct]

Manufacturer´s Ali-shop (RD tech, or also Hangzhou Ruideng Technology Co., Ltd) :

https://rdtech.aliexpress.com/store/923042

Go to Products/DC Digital/DC Voltmeter 5 Digit

No luck, just the junky ones. Might be my phone browser, di not install the app

https://m.aliexpress.com/storesearch/list/5%20digit%20.html?sceneId=050000&keywords=5%20digit%20&searchType=storeSearch&sellerAdminSeq=202659690&storeId=923042

 

[Pajda]

I also started calendar aging during summer this year.

Excellent news, can we expect your "first impressions" soon?

 

[serious_sam]

"torture test" of High Power cells HG2, 30Q ant VTC6 at 5C(15A) continuous discharge at 100% DoD.

Now this will be interesting !

 

[docware]

Excellent news, can we expect your "first impressions" soon?

You know, 100 days is too short time for relevant results IMO. Trend of degradation will be different at 300 – 400 or even more days. Moreover, my calendar aging study is again little bit problematic because of unsufficient amount of samples (only one for each case). However, I can show the first results. Some numbers are really strange, especially Samsung 29E6 increased capacity at 50% SOC and decreased DCIR at 100% SOC. Interestingly, Samsung 50E DCIR is very low compare to the older samples.

All samples are stored in one cardboard box at ambient temperature 25 ±2°C.
Capacity testing : charge 1A to 4,2 V, 0,1A cut-off, 1 hour rest, discharge 1A to 2,5 V, all at ambient 25 ±1°C. All 100% SOC samples first discharged to 50 %, rest 1 hour, then charged to 4,2 V.
DCIR measured according IEC 61960-3:2017, paragraph 7.7.3. Measuring parameters are unified to 0,6/3 A at 50% SOC for all cells. Temperature of the cells 25,6 ± 0,2°C.
Equipment : ZKETECH EBC-X0510 for charging, true 4-wire fixtures BF-2A, electronic load Maynuo M9712, DMM GW Instek GDM-8351, thermometer Omega HH 520, K thermocouples.



Capacity decay at 100% SOC : M29 line hiden below VTC6 line (identical values), 30Q line nearly blends with GA.



DCIR at 50% SOC : 30Q line blends with 50E.

 

[serious_sam]

Not a criticism, just a comment:

The capacity graphs look really bad because the lines are sloping down so steeply (because you're not graphing from 0-100%).

But it's only @98% which (I think) is pretty good.

You rescale from 0-100% and those lines are nearly horizontal. Visually tells a different story.

Just saying...

 

[Pajda]

Thank you very much docware for this amazing Christmas gift

I noticed that you prepared space for percentage values of capacity fade for all four graphs but the values are visible only in 100% SoC capacity?

I think that most significant issue with this particular test is a ZKETECH EBC-X0510 accuracy (stability). I think that this can be an explanation for your 29E6 results, by the way I expected much more zig-zag values. This instrument has current accuracy(and stability) of 0.2 % + 0.005 that can be translate to ca ±7 mA and can produce up to 21 mAh error after 1A discharge. Calibration helps, but still I was not able to get under 5 mA in average. For example on my $$$ tester I can achieve under the same conditions a 0.12 mA accuracy and so 0.36 mAh total error. My average monthly capacity fade is in the range of 2-4 mAh including full nominal capacity test and both 10s and IEC DCIR test. After finishing one year I will run capacity test to substract the influence of 12 nominal capacity tests.

Second issue is temperature stability. As you instructed me :wink: , I am using temperature chamber with 0.1 °C stability.

But still I think your results will be very benefical. My intention of doing this test is to verify that 20% capacity fade per year when stored at 100% SoC RT is politely speaking totally outdated value.

Small teaser from me:

LG M36 capacity fade after 8 months of 100% and 50% SoC storage at RT, 
! including 8 nominal capacity test cycles with 10s and IEC DCIR tests !
 
100% SoC:  95.4 %  of initial capacity
 50% SoC:  95.7 %  of initial capacity

 

[Pajda]

Migrated to a new thread https://endless-sphere.com/forums/viewtopic.php?f=14&t=114473

Test eqiupment:
- 2pcs ZKETECH EBC-A40L
- 2pcs BF-2A four-wire cell fixture (30A version with "huge" golt plated contacts)
- tested in room with relatively stable temperature 23 ±2 °C

Remarks and comments:
- someone considers these testers a child's toy and so I am able present results on my own decision
- the most problematic is tester accuracy. It is 0.2 % ±0.01 A. The stability is relatively good.
- do not much look at the absolute values, particularly at low currents in nominal capacity tests, trends are here important results.
- will try to add more data as it will come (ca one week for 50 cycles) and finally add energy and 10s DCIR graphs as well
- each 50th cycle is inserted datasheet nominal capacity test with 10s DCIR test (see dots in figure)
- 100 % DoD should be with asterix because:
a) HG2 and VTC6 datasheets recommends to go down to 2V when cycled at 5C.
b) minimal charging CV phase cut-off current is set to 100 mA due to the EBC-A40L limitation (HG2 should go to 50 mA cut-off)

* There was ca 4 weeks pause in testing after 200 cycles 30Q and after 150 cycles VTC6
** Changed VTC6 tester from A40L to A20 model after 720 cycles
*** Missed data for 50S, 40T3 and P42A caused by the harddrive failure
____
added "Table" with values measured on precision battery tester (not ZKETECH). Those test were performed in temperature chamber under 25°C, TC is using forced air flow to maintan consistent temp. This 25°C air flow acts as cooling force on the cell surface, so the measured temperatures under high load will be affected by this.

added Samsung 30Q6 to the chart (A20 tester)
added Molicel P42 to the chart (A20 tester)
added Samsung 50S to the chart (A40L tester)
added Samsung 40T3 to the chart (A20 tester)
added Sony 21700VTC6A to the chart (A20 tester)
added Samsung 25R8 to the chart (A20 tester)

Migrated to a new thread https://endless-sphere.com/forums/viewtopic.php?f=14&t=114473