SAMSUNG INR18650-29E cycle test.

[electricbike]

I have just started the cycle test of Samsung INR18650-29E.
4 cells under test with different parameters.

Cell Nr 1: discharge 2,7A. charge 1,9A. 4,15 V - 3,10 V
Cell Nr 2: discharge 2,7A. charge 0,8A. 4,20 V - 2,70 V
Cell Nr 3: discharge 2,7A. charge 0,8A. 4,20 V - 3,70 V
Cell Nr 4: discharge 2,7A. charge 0,8A. 4,05 V - 3,55 V

 

[DVDRW]

29E is 1.5x cheaper and larger capacity than 25R and 25R is never in stock @nkon
Lets see if cycle life justify the price.
Problably my next 80v 50ah triangle pack built with 29E cells 18p22s
Discharge ~0.5C
Charge ~0.2C to 4.1V

 

[DrkAngel]

Using Samsung INR18650-29E discharge graph ...

Using .2A discharge line, I was able to roughly map cell capacity.

• Method - I:
• used "paint"
• added .5V graduations
• resized to equalize pixel count to graduated measure.
• went 300% size to precisely gauge

and transferred data to graph.

Being derived from a low discharge graph ...
I would roughly estimate optimal charged voltage as 4.15V and optimal static DOD as >3.20V (Higher resolution at lower (.1A) discharge rate needed to better determine capacity map and optimal static voltages!)

INR18650-29E spec sheet - from Samsung

 

[Pedrodemio]

DrkAngel,

Have you done a capacity mapping of the NCR18650B?

(http://lygte-info.dk/review/batteries2012/Panasonic%20NCR18650B%203400mAh%20(Green)%20UK.html)

 

[Teh Stork]

Please, start a separate thread for capacity mapping and comparison between cells. It is a topic that poorly fits this topic.

Looking forward to data from this test

 

[savaoaknyc]

From the german forum

at discharge rate of 5a from 4.2-2.8v after 400 cycles we have around 11% lost.
But there is another tests which has around 5.4% after 100 cycles at mixed discharge rate.Looking forward to see this test

 

[DrkAngel]

Just got note that initial cycles will be sampled more closely to chart this early degradation.

 

[electricbike]

Uppdated in the first post with graf. everything looks good exept for the cell-4.
i dont know why yet, will do some more cycles and then test the cell with some other charger and se if it's the cell or something else, the charger maybe.

 

[Wheazel]

At first glance the cell 3 graph seemed to be malplaced, but then I realized it might be correct that only 600mAh are delivered down to 3,7V.

 

[Teh Stork]

Great to see some data!

I'd be fine if you added a full 4,2 V to 2,8 V discharge every 200 cycles for cell nr 3 and 4. You mentioned that temperature is logged - any chance cell 4 is put in a draft or similar (looks to be a cyclical event on a daily basis)?

If you could post some more info on your method it would be of great value (setup, pictures of setup, equipment, possibly schedule).

 

[DrkAngel]

Choice of test parameter seems haphazard?

Cell Nr 1: discharge 2,7A. charge 1,9A. 4,15 V - 3,10 V
Cell Nr 2: discharge 2,7A. charge 0,8A. 4,20 V - 2,70 V
Cell Nr 3: discharge 2,7A. charge 0,8A. 4,20 V - 3,70 V
Cell Nr 4: discharge 2,7A. charge 0,8A. 4,05 V - 3,55 V

Cell 3 and 4 tests will demonstrate deterioration from charged voltage with ~equal capacity used = test with good valuable results.
Should yield excellent data quickly due to short cycle time.

But Cells 1 and 2 tests vary voltage range and charge rates.
There is no commonality to determine any relation of either charge rate or charged voltage or discharged voltage as a quantifiable factor of capacity deterioration. ... ?

• Recommend future tests to directly compare:
• different charge rate with same C\D voltage range
• same DOD with different Charged voltage
• different DOD with same Charged voltage

Additional test? - 1 C\D cycle at .1A Discharge from 4.25V to <3.00V (w/ all data) - to precisely map capacity (energy density)


Mapped Discharge area and percentage.
Charge point approximated - Discharge point actual.
Lesser sag 3.0A discharge line used as closest reference to 2.70A actual discharge rate.

Cell 3 & 4 closer capacity than graphed.
Wanted clear separation.

Approximated from graphs:
/ 2750

Cell #1 - 4.15 - 3.10 2370/ 2750 = 86%
Cell #2 - 4.20 - 2.70 2695/ 2750 = 98%
Cell #3 - 4.20 - 3.70 635/ 2750 = 23%
Cell #4 - 4.05 - 3.55 675/ 2750 = 24.5%

 

[DrkAngel]

"Lined" Test areas against near static discharge capacity map.
Note: 1C discharge "stretches" 3.00V blue line of graph to near 2.70V line of Cell #2 test Line

0 to 100mAh additionally scaled as oem rated capacity % (100% = 2750mAh)

 

[electricbike]

Great to see some data!

I'd be fine if you added a full 4,2 V to 2,8 V discharge every 200 cycles for cell nr 3 and 4. Yep, will add full discharge every 200 cycle. You mentioned that temperature is logged - any chance cell 4 is put in a draft or similar (looks to be a cyclical event on a daily basis)? Dont think so but will check the logs.

If you could post some more info on your method it would be of great value (setup, pictures of setup, equipment, possibly schedule).A picture of the setup is coming now./quote]

My test setup.

 

[TheBeastie]

Ahh full cycle testing on Samsung 29E, my main pack is build on 29E its 12S7P..
I am so happy, this is like my birthday!
Subscribed!

I actually charge my pack via a CB86 which gives 6amps max per channel, I am actually charging each half of the pack instead of in 1 go.. but I am going to split the pack and add a plug/switch to conveniently separate/join the pack into a 6S14P for charging.
That would mean I am charging each cell at 0.423amps which should in theory give me super performance if all this data is right.
I am not that fussy about specific making any charge requests etc as what you guys are doing now is pretty wide set of tests..
[youtube]MVmWno3zJJE[/youtube]

 

[Joe90]

It's surprising that the cell that are fully charged at 4.2 volts keep going way better than the 2 others that are no fully charged.
Slow charge to full 4.2V seem good. this is surprising.
Nice tests

 

[DrkAngel]

Cell nr 4 itself seems likely to be defective.

Exhibited cycle chart might indicate a cell suffering from self discharge.

Self discharge test: (1/1000th V meter recommended)
Charge cell to any specific voltage, near full works best.
Note precise voltage after charge completed. (charger removed)
Note cell voltage at set intervals, 12hours? daily?
Compare Voltage loss over time...
Any notable voltage loss indicates some internal short = defective cell.
Might be confirmed by cell being noticeably warmer than typical after full charge.

 

[electricbike]

All 4 cell uppdated in first post. Will pause test on 29E cell because of 1 week holiday from monday and will restart the tests after that.
Will also do a more thoroughly tets on cell 4 then.

 

[DrkAngel]

Make sure cells are charged,
Remove from test devices,
carefully measure and record voltages.
Compare to voltage after 1 week trip.

Should pinpoint any cell with self discharge.

 

[electricbike]

Yes i will do a self discharge test to.
I also have this idea to put up a fully automated cycletest and send it to a live website for everyone to take part of.
Then i don't have to put up all the data to a graf by my self, it takes a lot of time to do that.

 

[Teh Stork]

The cells sat in storage for 1 year before I sent them to you electricbike. If cell 4 was affected by self discharge, it would have been dead. I also checked the voltage of all cells (3,703 V) before sending them out.

Possible error sources are still external temperature disturbance and the reactors (cell testers). Judging from the photo of your setup, I doubt that the variations are caused by temperature - but try to swap cell 3 and cell 4 position. See if things change.

Edit: I can send another pair of cells from the same batch, and you can replace cell 4.

It is also obvious the NCA chemistry dislikes the 0,7 C charge regimen (cell 1). As the 29E is rated to C3 charge, charging them with C3 is important and not a spec that should be broken.

If you have reactors available I'll send you two cells.
One to replace Cell 4.
One to add another test, call it cell nr 2a: discharge 2,7A. charge 0,8A. 4,10 V - 2,70 V (basically the same as cell 2, with a lower top-charge voltage)

 

[DrkAngel]

Then i don't have to put up all the data to a graf by my self, it takes a lot of time to do that.

Rapid deterioration during initial cycles benefits from close sampling.
After 50 - 100? cycles, sampling every 10 - 20 - 50? cycles is more reasonable.

 

[Pajda]

Cell nr 4 itself seems likely to be defective.

I think that this should be an unexpected but common behavior for cell cycling in this SoC region (4,0-3,6V). See my results of LG ICR18650-D1 cell. Here you can see also a "capacity swing". But when I made full capacity test after ~240 cycle, there was no significant capacity degradation.

 

[Pajda]

I also have this idea to put up a fully automated cycletest and send it to a live website for everyone to take part of.
Then i don't have to put up all the data to a graf by my self, it takes a lot of time to do that.

Yes I have the same issue. My idea for the beginning was to use something like Google Sheets (like the Tesla battery degradation chart made by Maarten Steinbuch)

 

[Offroader]

I still believe the best way to test properly for capacity is after every 50th cycle, for example, test the full capacity of the battery at say 1 amp discharge from 4.2 to 2.5 volts. Making sure to get exactly 4.200 voltage before discharge.

The other important thing is to make sure the temperature is exactly the same. I learned that when testing capacity in my cold basement at 14C had serious capacity loss compared to tests done at 25C.

The benefit of doing it this way is the graphs are all equal and we can see how capacity is falling without making another graph to show percentage change. I also believe the results will be more consistent and accurate because we are measuring total capacity in the cell and not partial capacity, but this may not matter.

Also running the cell from 4.2 to 2.5 volts every 50th cycle should not have any significant deterioration in cell capacity.

I finally got my Sanyo 3500 MAH cells and will do a capacity test like mentioned.

 

[electricbike]

Teh Stork: Yes i have one extra Reactor free for another test now.
That would be nice if you want to send another two cells, then maybe the cells have arrived when i am back from my holiday.