Li-ion cells cycle ageing

[electric_nz]

Very interesting to see the internal resistance of many of these remain consistently low over the full 700 cycles. It appears this new generation of cells (vtc6, hg2) has both high power, relatively high capacity and outstanding cycle life. Assuming these formulations are the same used in factory built electric vehicles; the future for electric propulsion looks very bright. Amazing work, thanks!

 

[madin88]

But we have to consider, accoring the first post, the discharge current is 2.5A only. Yeah, its good over the datasheets discharge rate and closer to reality. But I think many e bike riders here load their batteries with higher constant current per cell. Or do you want to commute 30mph (45kmh) with a BBS02 750W assuming 25A consumption and carry a 14s10p (10p!!!) pack with you?

Yes it would be good to have another cylce test, with lets say:

5A for LG M36 and MJ1
5-10A for 29E7 and M29
10-15A for HG2 and VTC6 with 2A charge current or so

That would be way closer to how the cells usually being used in real world.

 

[Centurio]

That would be way closer to how the cells usually being used in real world.

I was thinking the same. But: Why not to monitor a battery under realy conditions. Monitor the discharge, charge, cycle, ambient temperatur, celltemperatur (optional), ++. It sould be something easy, anyone can hook up his pack, analize and if desired, shere the data on ES.

Maybe little off topic. But anyway, maybe one knows a link.

 

[Cephalotus]

Yes it would be good to have another cylce test, with lets say:

5A for LG M36 and MJ1
5-10A for 29E7 and M29
10-15A for HG2 and VTC6 with 2A charge current or so

That would be way closer to how the cells usually being used in real world.

Typically the battery in an electric bike should/will last for 1-3 hours of continous riding. On average this translates to C/3 to 1C discharge rates. So 2,5A per cell is perfectly fine.

Who would build an ebike that has an empty battery after just 12 minutes of driving? (VTC6 or HG2 at 15A continous discharge)

 

[john61ct]

The point is not the average

some use cases need peak currents, that may not last that long

to be many times the usual amps capacity

 

[docware]

New data : Samsung 35E 673 cycles, Samsung 50E 650 cycles, LG HG2 658 cycles.

 

[madin88]

Typically the battery in an electric bike should/will last for 1-3 hours of continous riding. On average this translates to C/3 to 1C discharge rates. So 2,5A per cell is perfectly fine.

Who would build an ebike that has an empty battery after just 12 minutes of driving? (VTC6 or HG2 at 15A continous discharge)

You are right there. This is true for e-bikes most people ride with or the average one from the large companies.

If someone start building his own bike with tuneable controller, or when i think about E-mopeds like Surron, the peak C-rate usually is much higher there.
Surron as example has 6kW peak power and around 1900Wh battery (Panasonic PF) so already above 3C or 9A per cell. Still not true continuous in most cases but the controller allows it if you want. Thats the point.

The 2.5A tests are more a best case scenario and such 5, 10, 15A tests then would show the worst case and real lifetime then can be expected to be somewhere in between.

 

[john61ct]

Note that most lifecycle testing is done at steady CC well below 0.4C

which most people here consider unrealistically low.

So yes using closer to 1C is better from that POV, if your usual usage is well below that then you're likely to get even longer lifespans.

 

[docware]

If someone start building his own bike with tuneable controller, or when i think about E-mopeds like Surron, the peak C-rate usually is much higher there.
Surron as example has 6kW peak power and around 1900Wh battery (Panasonic PF) so already above 3C or 9A per cell. Still not true continuous in most cases but the controller allows it if you want. Thats the point.

The 2.5A tests are more a best case scenario and such 5, 10, 15A tests then would show the worst case and real lifetime then can be expected to be somewhere in between.

Those are unrealistic expectations. Cells warming in the closed battery pack will be probably too high even for 5 A discharge continuous with significant temperature inhomogeneity of the cells. Without suitable thermal management of the pack we can expect faster degradation of the cells.

https://endless-sphere.com/forums/viewtopic.php?f=14&t=96360&p=1447905#p1447905

https://endless-sphere.com/forums/viewtopic.php?f=14&t=99047&p=1452025#p1452025

 

[docware]

The final batch of data - Samsung 35E, 50E and LG HG2 finished 700 cycles.







Just as a matter of interest, here is capacity decay comparison without Samsung 50E offset :


Capacity decay comparison without zoom :

 

[Pajda]

Should I ask for DCIR rise comparison in % without zoom?

 

[docware]

Sunday afternoon, no brain capacity for such work
But have something better, source Excel file :

View attachment DCIR rise Excel file.xlsx

Edit - DCIR without zoom :

 

[Cephalotus]

Hey,

very cool. Thank you.

Do you have cycle 1 vs cycle 700 charts for capacity and Ri for the HG2 cells?

 

[docware]

Currently preparing charts for all three cells.

 

[docware]

Samsung 35E has 87,4 % of initial capacity in the cycling interval 4,1 - 3,4 V / 2,5 A discharge. In standard range 4,2 - 2,5 V at 1 A discharge has 93,6 % of initial capacity and 93,6 % of initial energy.



Samsung 50E has 93,8 % of initial capacity (offset variant) in the cycling interval 4,1 - 3,4 V / 3,75 A discharge. In standard range 4,2 - 2,5 V at 1,5 A discharge has 95,7 % of initial capacity and 95,8 % of initial energy.



LG HG2 has 92,1 % of initial capacity in the cycling interval 4,1 - 3,3 V / 2,5 A discharge. In standard range 4,2 - 2,5 V at 1 A discharge has 94,7 % of initial capacity and 95,1 of initial energy.



Updated SOC comparative chart after 700 cycles :


Want to again remind that this was not comprehensive testing but only laboratory comparative testing of cycle aging at particular conditions. Conditions quite favourable to the cells, but on the other hand typical for lot of ebikes. However, in the real life there are other factors affecting the cells degradation, not only cycle aging.
Any opinions, comments, remarks ?

 

[john61ct]

Thank you.

I have to say even the lower-performing cells did pretty darn well. . .

 

[Cephalotus]

Very good and very helpful.

Thank you very much!

 

[Honk]

Now that you have confirmed cell performance at modest cycling then perhaps it's time for a more brutal approach.
E'g full 4.2V charging at high currents and high discharge rate down to 3V to simulate how folks tend to build their light weight packs and kind of abuse them hard.
It could be fun to see which cells that can keep up without to much damage!

 

[mighty82]

I would also like to see at least some of the cells tested at 4.2v. With this much great information it is a golden opportunity to see what difference 4.1v vs 4.2v actually makes with the other parameters being the same.

 

[serious_sam]

How much are you guys paying docware for performing all this extra testing at your request? Just curious.

 

[Honk]

Zilch - so I don't expect or demand anything from him.
It was just some basic curiosity from my side.
A few promising cells could be selected to narrow the effort.
And it might interest docware himself, based on curiosity!

 

[john61ct]

I agree such requests appear entitled.

It really is not that hard to start doing your own tests if you want to.

 

[mighty82]

Sharing ideas for future tests is not being entitled. I'm sure he's going to continue testing cells in the future anyway.

The effort put in here as well as the execution is amazing, and I am very grateful for it!

Having this many different cells tested at the same time is gold. Comparing cells tested at different times, by different people with different equipment is impossible.

 

[mighty82]

Any idea why the cells have lost more capacity percentage wise inside the cycling interval than total capacity degradation? Have you tried retesting the cells with the previous parameters after it has been fully cycled to see if some of the capacity within the cycling interval has been regained? Might be a posibility since they also regained capacity during a resting period as you mentioned

 

[docware]

I'm sure he's going to continue testing cells in the future anyway.

Very probably not, cells testing became boring. Vegetable growing is much more interesting and enjoyable.

Comparing cells tested at different times, by different people with different equipment is impossible.

Wrong, it´s not easy, however it´s possible.

Any idea why the cells have lost more capacity percentage wise inside the cycling interval than total capacity degradation?

No.

Have you tried retesting the cells with the previous parameters after it has been fully cycled to see if some of the capacity within the cycling interval has been regained?

No.