18650 capacity loss test - garage vs basement vs fridge

Since I have Sanyo 18650GA 3500 cells left over from my battery build I decided I want to do a capacity loss test.

What always concerned me was if having my bike in my hot garage which has fluctuating temps and especially high temps in the New York summer (95F / 35C) has an effect on capacity. I also want to see if keeping my bike almost fully charged at 4.1 volts has an effect on the capacity as I would like to keep my bike fully charged at 4.1 volts if it doesn't make much of a difference.


I choose 4.1, 3.66, and 3.3 volts to see if storing them at different voltages makes any difference. I also placed them in different locations to see if temperature makes any difference.

All cells tested for capacity at around 78F and written on cell before placing them in the following locations: As of July 13th, 2016 all cells have been placed in the following locations.

3 cells in garage, 4.1 volts, 3.66 volts, 3.3 volts TEMP RANGE through year from about: 32F / 0C to 95F / 35C
3 cells in basement, 4.1 volts, 3.66 volts, 3.3 volts TEMP RANGE through year from about: 60F / 15C to 78F / 25C
2 cells in refrigerator at 32F / 0C, 3.66 volts and 3.3 volts
2 cells in freezer at 0F / -18C 3.66 volts and 3.3 volts
1 cell in attic, 3.66 volts, TEMP RANGE through year from about: 32F /0C to 120F / 49C Has the highest temp fluctuations compared to all locations.

Just wanted to add that I actually had an unused 18650 cell left in my hot summer garage for the last couple of months. What is odd is after I let it cool off in my basement for over a day and then cycled it, it tested higher capacity at 2.54 AH of capacity.

This came as a surprise because all the other unused cells in my more stable and cooler temp basement tested at 2.53. I expected this garage cell to actually test lower capacity than the cells in the basement because it has been exposed to temp fluctuations this summer from 75-95F daily for a couple of months.

Just an odd finding, but since it is only a single cell I can't draw any conclusions. All the other cells tested 2.53 in the basement. So my charger is very consistent.

Cool bro, I'll check these experiments out!
Interesting on the cell in garage vs basement.
Not impossible I suppose- and good-on-you for doing the 'speriments

I too am interested in your findings.

Tom

I don't want to discourage you, but similar experiments have been done by professionals in the past.

The format of the cells may slightly affect the outcomes, but you should be getting something close to this:

This is where I decided to place my cells and at what voltage.

I used 3.66 as that was the voltage of the cells when I received them new.
4.1 volts as that is the highest I charge my batteries and curious to see how much faster they deteriorate if I decide to keep my bike fully charged at 4.1 volts most of the time when not using it.
3.3 volts as that is the lowest I discharge my batteries

All cells tested for capacity and written on cell before placing them in the following locations:

3 cells in garage, 4.1 volts, 3.66 volts, 3.3 volts TEMP RANGE through year from about: 32F / 0C to 95F / 35C
3 cells in basement, 4.1 volts, 3.66 volts, 3.3 volts TEMP RANGE through year from about: 60F / 15C to 78F / 25C
2 cells in refrigerator at 32F / 0C, 3.66 volts and 3.3 volts
2 cells in freezer at 0F / -18C 3.66 volts and 3.3 volts
1 cell in attic, 3.66 volts, TEMP RANGE through year from about: 32F /0C to 120F / 49C Has the highest temp fluctuations compared to all locations.


So now it is just a matter of waiting and seeing the results. Not sure when I will update anything, could be in a year from now.

Sunder said:

I don't want to discourage you, but similar experiments have been done by professionals in the past.

The format of the cells may slightly affect the outcomes, but you should be getting something close to this:

Click to expand...

Not for latest panasonic chemistry... It's more temperature resistant.

*Subscribed*

This I will follow.

Sunder said:

I don't want to discourage you, but similar experiments have been done by professionals in the past.

The format of the cells may slightly affect the outcomes, but you should be getting something close to this:

Click to expand...

This chart is very outdated and I believe it was from over 10 years ago when lithium chemistry was very new. I have not seen anywhere near these specs on my batteries.

Here is a 4s Lipo 5000 pack that I never used and had it since 9/21/14

Here are the capacity out numbers from 4.1 to 3.7 volts.

9/21/14 = capout 3494
4/10/15 = capout 3450
7/12/16 = capout 3463

You can see very little difference in capacity. I believe the difference between 4/10/15 and 7/12/16 is because my basement is warmer at 7/12/16 so I will have to rerun these tests. But they shouldn't be that much difference and certainly not 4% a year of loss.

When using the lipo packs on my bike and with use and cycles they saw about 5% loss a year, but that was with use.

We will see how these 18650 cells work out, but I would not rely on that chart as it is old and outdated.

Offroader said:

Sunder said:

I don't want to discourage you, but similar experiments have been done by professionals in the past.

The format of the cells may slightly affect the outcomes, but you should be getting something close to this:

Click to expand...

This chart is very outdated and I believe it was from over 10 years ago when lithium chemistry was very new. I have not seen anywhere near these specs on my batteries.

Here is a 4s Lipo 5000 pack that I never used and had it since 9/21/14

Here are the capacity out numbers from 4.1 to 3.7 volts.

9/21/14 = capout 3494
4/10/15 = capout 3450
7/12/16 = capout 3463

You can see very little difference in capacity. I believe the difference between 4/10/15 and 7/12/16 is because my basement is warmer at 7/12/16 so I will have to rerun these tests. But they shouldn't be that much difference and certainly not 4% a year of loss.

When using the lipo packs on my bike and with use and cycles they saw about 5% loss a year, but that was with use.

We will see how these 18650 cells work out, but I would not rely on that chart as it is old and outdated.

Click to expand...

Aligns with my understanding aswell. Never saw such degradation from storage.
A more recent indication of capacityloss when storing fully charged 18650s can be found in a jehugarcia video.
https://www.youtube.com/watch?v=jAcgld6gtyo

Interesting video, basically he lost about 2% a year capacity keeping them charged at 4.2 ? volts. What is unknown is if they were charged to 4.2 volts, all he says is he fully charged them and written on the battery is 3.88.

But yes I agree that keeping batteries today charged at high voltages and left to sit only lose a couple of % of capacity a year.

This is exactly why lots of those old charts are outdated now, although must people believe them.

This is exactly why I wanted to do a test of my own to see exactly what the losses are. This way I won't care if I charge the night before to 4.1 volts and have the bike ready in the morning. Or if the heat of my garage is causing rapid deterioration.

We need to move away from the outdated research done on batteries. Even in the video he talks about people telling him that keeping them charged at high voltages will make them lose like 50% capacity in a few weeks.

[youtube]jAcgld6gtyo[/youtube]

One important thing I wanted to point out is that it is very important to check capacity at the same room temperature.

My basement was 80F or 26.7C when I took initial capacity measurements.

This is why I will have to check my capacity yearly to make sure I check them again at the same temperature.

From doing other cell tests, I have found that checking capacity during winter when my basement is cold at 60F or 15.5C, will reduce the capacity of the cells by about 3% to 4%.

subscribed

5 months to go!

Still a couple of months to go but I used 280 of these same cells in my bike.

After 11 months of use, but I hardly used the bike in the winter. I am showing only about 1% to 1.5% capacity loss on my max-e. However, my pack is so large at around 3.25 KW, I have done many partial cycles from maybe 3.6 volts to 4.0 volts. I always stay within 3.3 to 4.1 volts, and I also keep my bike well discharged at around 3.4 volts per cell when not in use.

I am also pushing the cells at a max discharge of 6.5 amps per cell on these 10 amp rated cells.

With my bike showing such a small amount of capacity loss, I think these test cells should have less than 1% capacity loss for the year.

It is showing that all those old battery university info where capacity loss was like 3-10% a year is all wrong and this has not been the case for new lithium batteries. Even my old lipo packs had a very small amount of capacity loss every year when not in use.

Here is the post for more info on the 11 month capacity loss for my E-bike if anyone is interested.

https://endless-sphere.com/forums/viewtopic.php?p=1289529#p1289529

offroader, I know you are only one user, but...I want to thank you for putting in the extra time and effort to track this, and then post the results.

+1 great stuff!

@offroader,
What is going to be your method of testing these once you bring them out of storage?
Do you have access to a charger that can record discharge and charge cycles to an sd card or pc? (eg, iCharger 308/4010 duo)

I suggest you measure the voltage at which they are once you take them out of storage, discharge them and record the capacity, then charge and discharge 3 times while recording the capacity that goes in and out of them. If you have access to a charger that can measure internal resistance, that would probably be really good data on the health of the cells.

PXSS said:

@offroader,
What is going to be your method of testing these once you bring them out of storage?
Do you have access to a charger that can record discharge and charge cycles to an sd card or pc? (eg, iCharger 308/4010 duo)

I suggest you measure the voltage at which they are once you take them out of storage, discharge them and record the capacity, then charge and discharge 3 times while recording the capacity that goes in and out of them. If you have access to a charger that can measure internal resistance, that would probably be really good data on the health of the cells.

Click to expand...

I have placed cells in various locations and temperatures, I also have them at all different stored voltages. This was designed to see how best to keep my bike battery stored when not in use. I guess it makes sense to still discharge them first right after storage, before charging, the more info the better.

I will let them sit at the same mid July basement room temperature that I first tested them at, which is around 78 degrees, for a couple of days to equalize all the battery temps. I have found that room ambient temperature makes a huge difference in capacity testing so it needs to be very consistent here.

I use one of the most accurate charges available from revolectrix which seems to be the most accurate and high-quality charges. I find it to be very consistent when checking my cells.
http://www.revolectrix.com/GT500_description_tab.htm


I am not sure if I can check internal resistance for single cells with my charger, but it does give a reading when charging packs. I have to see how that internal resistance test works and if I can make it work for single cells. Maybe I need a 1s lead connector. For my discharging, I stop charge exactly at 3.3 volts so if the internal resistance is bad the discharge stops early and it shows up in capacity.

Most likely there will not be any difference in any of the cells within 10 MAH capacity, and my charger only gives me a 10 MAH capacity resolution. New lithium cells are probably much better than the old stuff and just doesn't deteriorate all that fast. But we will see, that is just my hypothesis. So far these cells on my bike have shown very small capacity loss after 1 riding season.

Offroader said:

I have placed cells in various locations and temperatures, I also have them at all different stored voltages. This was designed to see how best to keep my bike battery stored when not in use. I guess it makes sense to still discharge them first right after storage, before charging, the more info the better.

Click to expand...

Yes, the initial discharge would capture the self discharge characteristics of the cells with respect to location and storage voltage.
For example, the cell in the fridge at 3.33V might be at 3.23V after the year and the cell in the basement at 3.33V might come out at 3.1V. The very first discharge would capture which cell self discharged more which might or might not correlate to the highest loss in capacity after a full charge (it should).

Offroader said:

I will let them sit at the same mid July basement room temperature that I first tested them at, which is around 78 degrees, for a couple of days to equalize all the battery temps. I have found that room ambient temperature makes a huge difference in capacity testing so it needs to be very consistent here.

Click to expand...

Yes! I agree. Batteries can be very temperamental in this respect.

Offroader said:

I use one of the most accurate charges available from revolectrix which seems to be the most accurate and high-quality charges. I find it to be very consistent when checking my cells.
http://www.revolectrix.com/GT500_description_tab.htm

Click to expand...

Shamefully, the GT500 does not have any data logging capabilities. That would truly bring this experiment to the next level as you would be able to graph everything. I don't think data like this is available in any research paper within the last decade.

Offroader said:

I am not sure if I can check internal resistance for single cells with my charger, but it does give a reading when charging packs. I have to see how that internal resistance test works and if I can make it work for single cells. Maybe I need a 1s lead connector.

Click to expand...

Your charger doesn't have an internal resistance check feature but it does display it while charging.

Offroader said:

For my discharging, I stop charge exactly at 3.3 volts so if the internal resistance is bad the discharge stops early and it shows up in capacity.

Click to expand...

So you don't discharge them fully? These cells go down to 2.5V. At 3.3V no load you still have roughly 17% capacity left in them. I don't know what effect this would have on the results of the test. I imagine you will still see a capacity loss on the cells left at 4.2V but it might not be as pronounced as it would be if the cells were discharged fully.

Offroader said:

Most likely there will not be any difference in any of the cells within 10 MAH capacity, and my charger only gives me a 10 MAH capacity resolution. New lithium cells are probably much better than the old stuff and just doesn't deteriorate all that fast. But we will see, that is just my hypothesis. So far these cells on my bike have shown very small capacity loss after 1 riding season.

Click to expand...

I want to believe this is true but based on my research, I do believe that the opposite will happen. The cells stored at 4.2V will have suffered a capacity loss of at least 300mAh. Nonetheless I am excited to see the results!

I did not fully discharge them to 2.5 volts with my test. I will never go below or above 3.3 volts and 4.1 volts on my bike so I decided not to do a full discharge cycle of 4.2 to 2.5 volts.

basically, I was only interested in the 3.3 to 4.1 because I never go beyond those numbers. Since this is for ebike use, I decided to stick with these numbers as this is also what most people here use. I do believe 4.2 volts is not a big deal because I once overcharged a cell to 4.4 volts and it is in my bike pack which is not in parallel. It still has held perfect voltage with all the other cells. This means that these cells can really hold much more than 4.2 volts. It took about 500 MAH to go from 4.2 to 4.4 volts.

My GT 500 charger can be hooked up to the computer through USB for discharge / charging graphing, but I'm just doing a basic test of capacity loss, and only need to know the amp hours that the cell has from 4.1 to 3.3 volts.

That will be a big test to see if the 4.1 volt stored cells had much higher discharge than the 3.3 volt cells. This is always a concern if I leave my bike topped off or keep it at a low voltage and then charge right before use. I believe there will be very little difference from the 3.3 volt stored cells and the 4.1 cells, but I could be wrong here. 300 MAH will be out of the question if anything it may be 30MAH or less. 300 MAH would be about 10% capacity loss, that would be impossible. Most likely 1% max loss from the 4.1 and 3.3 volt cells. More likey even less like .25%. The reason I think this is because I've forgotten lipo cells charged at high voltages only to find that they lost hardly any capacity at that high voltage.



The tests should be good, I used 280 of these same cells from the same lot in my bike and the ones in my bike have all held perfect voltage from each other and I have not paralleled the cells in my pack so each cell voltage moves independently.

These Sanyo cells are very consistent cells, so any differences in capacity will most likely be from the storage conditions.

Offroader said:

Since I have Sanyo 18650GA 3500 cells left over from my battery build I decided I want to do a capacity loss test.

Click to expand...

Big Kuddos Offroader we really need to start an awards program for people doing original firsthand data collection and research like this on their own time and dime and publishing the results here. Especially with battery ageing it's a lot of long-term planning that most people don't have the patience to do, but everyone (including me!) wants the answers and results. Staying tuned..

Just started to run some cells through my charger still have many more to test.

So far the ones stored in the basement are showing based on the stored charge voltage.

3.3 volts = about 0% loss
3.67 volts = about .75% loss
4.1 volts = about 1.5% loss

My resolution is only .01 Amp hours,

Still have lots to test but it seems the voltage you leave the cells at plays a big part in capacity loss. This is good because I always try to leave my bike at a low voltage when not using it. When not using it for a long time I like to keep it close to 3.3 volts per cell.

This may also explain why my lipo packs show almost no capacity loss also because I keep them stored close to empty.

Offroader said:

Just started to run some cells through my charger still have many more to test.

So far the ones stored in the basement are showing based on the stored charge voltage.

3.3 volts = about 0% loss
3.67 volts = about .75% loss
4.1 volts = about 1.5% loss

My resolution is only .01 Amp hours,

Still have lots to test but it seems the voltage you leave the cells at plays a big part in capacity loss. This is good because I always try to leave my bike at a low voltage when not using it. When not using it for a long time I like to keep it close to 3.3 volts per cell.

This may also explain why my lipo packs show almost no capacity loss also because I keep them stored close to empty.

Click to expand...

That's great - even if your degree of certainty was >100%, that's still very telling.

3.3 volts though - Glad you have no self discharge! the actual capacity between 3.3v and 2.5/2.8 (Depending on brand to empty) is usually very small.

great info