Proven causes of battery cycle life reduction

[cwah]

Hello all,

I've read and tested cells in the past, and I've found out the key cause to reduce life cycle are:
- Final charge voltage. The closer to 4.2V the faster the cell loses permanent capacity/increase resistance. Charge up to 4.05V can increase up to 300% (need ref)
- Charge rate. I've noticed on my trial that the higher the charge rate, the faster the cell loses permanent capacity/increase resistance. So ideally charge as low as possible and only use fast charge in case of urgency
- Discharge rate and heat. So for high discharge try not to do it continuously. Ref needed


Does anyone knows if final discharge voltage have an impact? Does discharging to 2.5V decreases life cycle? I know there is a risk of permanent damage but is it related to capacity loss?

 

[circuit]

Was temperature controlled during charge rate tests?

I did not see any proof of degradation due to low SOC, only high and temperature.

 

[spinningmagnets]

Justin posted some info at ebikes.ca about 4.15V, 4.10V, 4.05V top charge leading to dramatically increased lifecycles...

 

[cwah]

Was temperature controlled during charge rate tests?

I did not see any proof of degradation due to low SOC, only high and temperature.

I didn't do that at the time, but over 50 cycles of charge and discharge the trend was obvious. There was a bit more or less capacity at the end of charge during different time of the day, but on average after 50 cycles I could clearly see a faster and predictable capacity loss slope with higher end of charge.

Circuit have you also tested cells with various end of discharge cycle? That would really help as I don't have the equipment anymore since I moved.

 

[TheBeastie]

I never done any lab tests but I strongly believe in the theory that over discharge of lithium cells makes them go crappy.
When I had a 48v lipo pack on a 36v LVC controller a few years ago I once rode the bike all the way down to that LVC cut off which must of been like 2.5v each cell. After that the pack was never quite the same again always felt sluggish on the bike.

http://www.ebikes.ca/product-info/cycle-satiator.html#satiate-for-enhanced-cycle-life

I have my analogy description of how I like to imagine is happening with lithium battery cells with balloons instead...
https://endless-sphere.com/forums/viewtopic.php?f=14&t=54202&p=890904#p890904

 

[dogman dan]

No proof or testing to back this up, but I feel it makes common sense that the rate of discharge as you use the last 5% gets important. This is based solely on how hot a battery feels when you are finished. The basic concept is this, don't cook your battery.

Anything that makes the pack really hot is bad right? I don't mean body temp, I mean need a glove to handle it hot. Want to get a pack that hot? Give it a high discharge rate as you drive the cells over the cliff. I did this to four 5s Lipo packs once, the first ones I bought. They died in about 5 months. I never treated a pack that way again. From then on, when I know I'm draining that last bit, I lower the c rate to .5c or possibly even less. Quite a few packs treated better like this have all lasted at least 3 years, or at least till mechanical damage caused the demise. Those packs were always charged to 4.2v, and often stored that way for a long time too. But the ones I flogged hard to 3v and got hot died, the ones I milked down to 3v cooler did not.

If you think about it, as you discharge the pack, it's resistance increases. this is fact. So it makes sense that if you want that last 5%, you better take into account that the resistance will be high at that point on the discharge curve, and lower your discharge rate accordingly. So my take on this, is that how you discharge that last bit past the cliff on the graph matters a lot.

Ideally, keep those cells above that cliff voltage, resting. Then you have less need to be perfectly balanced. But once you are below that cliff voltage under load, slow down and milk em home without cooking your cells.

I think we can certainly agree, that getting your pack too hot to hold bare handed is always a bad thing. Bad on the charge, bad on the discharge, bad in storage.

 

[district9prawn]

As expected, 100% DOD has a big impact on cycle life.

But surprising to me how little there is to be gained by keeping SOC between 10-70% as long as DOD is kept at 60%.

 

[spinningmagnets]

The Prius and tesla both have battery packs that are far exceeding the warranty, and that is under near daily driving conditions. For both, cutting out some of the bottom and top of the charge envelope is frequently cited by battery engineers...

Get a large enough battery pack that you can just charge to 80% (Lunagizer or Satiator), so...charge to 4.05V and LVC at 3.2V?

 

[cwah]

As expected, 100% DOD has a big impact on cycle life.

But surprising to me how little there is to be gained by keeping SOC between 10-70% as long as DOD is kept at 60%.

hmm.. are these comparable? it seems it also change the initial charge? in other term, are they all starting from 4.2V?

 

[Teh Stork]

As expected, 100% DOD has a big impact on cycle life.

But surprising to me how little there is to be gained by keeping SOC between 10-70% as long as DOD is kept at 60%.
Effect of DOD on cycle life.jpg[/attachment]

Source?

 

[ecycler]

You have to be specific on the chemistry as it is drastically different between them and can not be generalized.

 

[cwah]

You have to be specific on the chemistry as it is drastically different between them and can not be generalized.

Lithium ion are all quite simjlar and I can't see how it can be drastically different between let say NCA vs NMC.

 

[brumbrum]

What about storing cells for any period of time when fully charged, does this have a large impact on cycle life?

 

[parabellum]

Calendar life seems reduced in full charge storage conditions, at least for NMC (which have aging problems anyway). Is it still important factor for NCA?

 

[999zip999]

Cell should not be stored fully charged. Need more time to know about NCA, years

 

[cwah]

Full charge is an important for all lithium ion. I think it's one of the fundamental that isn't going to change

 

[DrkAngel]

Charged Voltage
"Proven" factor
Battery University

See -Optimal Charge Voltage

Rate of Discharge
Notable factor!
Observed deterioration and ongoing testing indicate discharge rate as a definite factor.

Charge Rate
Comparison above also directly compares charge rate and indicate charge rate notably affects cycle life!
INR-18650-25R 2.5A Charge\5.0A Discharge vs 4.0A Charge\5.0A Discharge

Discharged Voltage
I advocate not discharging beyond the top of the "cliff" common in discharge graphs.
This differs with different lithium formulations and varies continually with discharge rate.
Discharging below this cliff is needlessly damaging for minimal actual capacity at a sagging-pitiful voltage.
Worse ... any weak cell or bank will drop lower and can suffer unrecoverable damage!

I personally gauge battery use by its static voltage (not under load) and estimate actual voltage during use.
I keep charged and discharged voltages in the regions of good energy density.

Capacity Map derived from discharge graph

This cell demonstrates optimal use between ~4.10V >> 3.40V+

 

[DrkAngel]

Rate of Discharge
Notable factor!
Observed deterioration and ongoing testing indicate discharge rate as a definite factor.

Mapped and labeled #1

Updated #2

 

[ecycler]

You have to be specific on the chemistry as it is drastically different between them and can not be generalized.

Lithium ion are all quite simjlar and I can't see how it can be drastically different between let say NCA vs NMC.

Just the different formulations of the nickel, cobalt and aluminum between NCA batteries will alter their properties. The engineers at Tesla seem to think the differences are drastic enough between chemistries to make this a major consideration in the powerwall design:

An NCA battery typically has a shorter cycle life and a higher energy density (and less stability). An NMC battery generally has a longer cycle life, more stability, and less energy density.

http://fortune.com/2015/05/18/tesla-grid-batteries-chemistry/

 

[ridethelightning]

An NCA battery typically has a shorter cycle life and a higher energy density (and less stability). An NMC battery generally has a longer cycle life, more stability, and less energy density.

http://fortune.com/2015/05/18/tesla-grid-batteries-chemistry/[/quote]

this is odd, as I was under the impression that NCA was better all-round chem because the manganese eventually leaks through the cells(from cathode to electrolyte) and causes faster cell degradation..

 

[Cephalotus]

Ways to kill most Li-Ion cells quickly:

1. Try to charge them with high current at low temperature
2. Charge them full and store them hot.

 

[parabellum]

Ways to kill most Li-Ion cells quickly:

1. Try to charge them with high current at low temperature
2. Charge them full and store them hot.

Hammer a nail in it!

 

[circuit]

Looked through all the graphs. To me it looks inconclusive, as there is not enough data. For example, when discharge rate tests were done, was temperature controlled? I mean, is the degradation due to high current itself, but not due to higher temperature?

However I did see a dramatic loss in capacity by too quick charging of Panasonic B cell:

Source: http://e-motion.lt/2013/07/30/licio-celiu-pasirinkimas-4/

The result was ridiculous, so the test was aborted. Those cells should not be charged quicker than C/4. Did not see such behavior in any other cell. But again, temperature was not controlled.

 

[TheBeastie]

Looked through all the graphs. To me it looks inconclusive, as there is not enough data. For example, when discharge rate tests were done, was temperature controlled? I mean, is the degradation due to high current itself, but not due to higher temperature?

However I did see a dramatic loss in capacity by too quick charging of Panasonic B cell:

The result was ridiculous, so the test was aborted. Those cells should not be charged quicker than C/4. Did not see such behavior in any other cell. But again, temperature was not controlled.

Wow so looking at that blog post via google translate it seems those panasonic cells are pretty crap? Do Tesla vehicles have temp control systems in their cars for the battery packs or their powerwalls? Seems like part of the 18650s specs should be able to deal with at least a bit of standard room temperature.. I would assume they are just using different type of cells all together?
I noticed this conclusion was almost 2.5 years ago maybe they were a bad batch of cells and if the test was done now it would all be magically better?
http://www.wolframalpha.com/input/?i=days+from+07.30.2013+to+today

 

[circuit]

I did only one test run on a single cell, so it is possible that it was defective or whatever. However the cell (sample) was bought from my old source, together with thousands of other cells (other model) and we never had issues with quality. So it remains an open question on that cell compatibility with high(er) current charge.

I repeated the test at lower current (on a new cell) and results were much better.
Even though I did not control the temperature during tests, it was not too hot (not above 45C I guess).

I don't really know what Tesla uses, I heard it is Panasonic cell (without any letter). But it looks like this cell ir really not suitable for supercharging.