Batteries Charging Routine

[Nealh]

The guys on ES are experimenters and love all these facts and figures which gives the likes of us ammunition to use to decide what to use.
If you read the threads they build larger batteries then most so often we see 100+ cell packs being used and often more then 60v, so some of them have vested interest in charging to lower voltage to try an extend the pack life.
Unless one is very knowledgeable then 4.2v is best.

Downsides as mentioned are less capacity which also means less range, unless you have a big battery to counter the range. For most folks 50/60 miles is more then enough range, if you need more for along ride or touring just carry spare batteries .
For my 100 milers I carried approx. 1500wh but would return home with 40% total, extra wh just means one can use higer asssit and not have to be concerned about range anxiety.

 

[guerney]

Dont forget that there are good reasons for using the full voltage range for cells. Restricting the voltage range to multiply the battery life can considerably reduce the actual available battery capacity and range of the battery. For example, when endless sphere tested the Samsung 50E at 4.1 volt and higher lvc, they measured the capacity at them voltages and it was reduced down from 4.954 to 2.878 Ah.
So it it appears that if you want to built a battery to the 'Tesla' spec (2000 cycles), you will need about 50% more cells which means about 50% more weight and cost.
This appears to be the case, whether you compare Samsung 50E to Samsung 33j long life or whether you compare cells used at full voltage range (4.2 )or reduced voltage range (4.1 & higher lvc ) depending on actual cell used.

View attachment 52842

So it it appears that if you want to built a battery to the 'Tesla' spec (2000 cycles), you will need about 50% more cells which means about 50% more weight and cost.

Very interesting. Personally, because I have an ebike, I really wouldn't mind a 50% bigger battery.

 

[StuartsProjects]

Very interesting. Personally, because I have an ebike, I really wouldn't mind a 50% bigger battery.

Good point.

I have a 36V 10Ahr that weighs 2.6Kg.

So if the battery 'life' would increase by 400% for a measly extra 1.3Kg, then I suspect a lot of people would go for it.

But I can understand why eBike battery manufacturers would not be so keen to follow that development route.

 

[Sturmey]

Good point.

I have a 36V 10Ahr that weighs 2.6Kg.

So if the battery 'life' would increase by 400% for a measly extra 1.3Kg, then I suspect a lot of people would go for it.

But I can understand why eBike battery manufacturers would not be so keen to follow that development route.

The original Raleigh all steel Bikes were considered a bike for life. They came with chainguards, 40 spoke rear wheel, long mudguards, steel frame, leather saddle and they did last a lifetime. But I am not sure if people and perhaps sellers are really interested in long life stuff any more. They are probably more interested in the lighter weight and performance of modern bikes. Many bikes seem to spend most of there time in sheds and rarely used.
I have bought many bikes secondhand for 20 euro or less and many were on their original chains and tyres, which shows that there was little usage.
But I personally do high mileage and hence my interest in hub motors and long life batteries. You may see a turn back in the future towards more practical and utility type bikes. Who knows.
So coming back specifically to batteries, there are reasons why long life cells may not be worth the extra weight and cost. For example, many batteries fail early because of mechanical damage or drenched by rain or stolen or the bike/battery being stored away for a long time with exhausted battery. Also. many people only use their bike once or twice a week so hence a standard 500 cycle battery is sufficient.
But there are people who use their bike daily and they would possibly benefit from a bettery quality battery with long life cells.

 

[StuartsProjects]

But there are people who use their bike daily and they would possibly benefit from a bettery quality battery with long life cells.

Indeed.

And it would be nice for battery manufacturers to provide consumers with a choice.

 

[StuartsProjects]

On the 10S1P of Samsung 21700 I have for eBrompton when the battery is charged to 4.1V per cell, the capacity is reduced by 24% in comparison to charging to 4.2V.

So possibly double charge\service life for 24% range reduction.

 

[Woosh]

That doesn't seem right.

 

[StuartsProjects]

That doesn't seem right.

What about it does not seem right ?

 

[Woosh]

Your battery gives out what your charger pushed in. There can't be 24% difference in energy between 4.1V and 4.2V.

 

[StuartsProjects]

Your battery gives out what your charger pushed in. There can't be 24% difference in energy between 4.1V and 4.2V.

Well there is, as thats what was measured, and its in agreement with the best of the figures on the chart in post #19. One battery in those results, Samsung 50E, showed a 40% degradation.

And the mAhr in when the 4.1V point was reached was also down by around 1Ahr.

 

[Woosh]

Well there is, as thats what was measured, and its in agreement with the best of the figures on the chart in post #19. One battery in those results, Samsung 50E, showed a 40% degradation.

And the mAhr in when the 4.1V point was reached was also down by around 1Ahr.

can you give me the URL of the cell loss table in post #19?

Samsung 21700 50E discharge curve showing less than 10% difference in capacity between 4.1V and 4.2V
Source: https://lygte-info.dk/review/batteries2012/Samsung INR21700-50E 5000mAh (Cyan) UK.html

 

[StuartsProjects]

can you give me the URL of the cell loss table in post #19?

Nope, maybe @Sturmey does. .

 

[Woosh]

I am intrigued by the 24%.
I'll try to make a time lapse video about charging time / battery voltage/energy consumption over the weekend.

 

[Nealh]

On the 10S1P of Samsung 21700 I have for eBrompton when the battery is charged to 4.1V per cell, the capacity is reduced by 24% in comparison to charging to 4.2V.

So possibly double charge\service life for 24% range reduction.

If that is the info from the ES testing done by Pajda, Docware and the likes then this is after 1000 static charge /discharge cycles. It is theorectical and doesn't take in to account normal user usage from day to day , year to year.
It is though an indication of which cells may perform best under similar cycling/testing conditions and is good independant info.
1c/2c & 3c cycle testing originally was on;y 70% DoD 4.1v - 3.4v.

Latterly new cycling has been undertaken to real world cycling of 95% DoD 4.15v - 3v also using 1c/2c & 3c rates. I haven't looked clearly or properly at the results though it appears that some of the better quality cells show not much loss of capacity by doing so.

 

[Sturmey]

can you give me the URL of the cell loss table in post #19?

There is a long running discussion on 'Li-ion cells cycle ageing' on endless sphere. Although they have demonstrate that cell life can be extended considerably by reducing the 'Depth of Discharge' by keeping the cell working in the range 3.4 to 4.1 volts, they have considerably reduced the available capacity. It does appear that high density cells e.g Samsung 18650 35E or 21700 50E are particularly affected. I think the losses are at both ends (>4.1v & below <3.4v) and work out to about15% at each end.
One contributor (Pa***) reckons that good quality cells do not require or benefit from this much and that a .5v reduction to 41.5 volt alone is sufficient .(leave LVC at 3.1 volt I presume)
The discussion is long and ongoing. The table is at below about a quarter way down the page 12.

Li-ion cells cycle ageing

The principle of this issue was already formally described in this topic. So yes HP cells have significantly worse cycle life than most HE cells under low loads up to ca 1-2C continuous discharge. Generally(Although I do not like generalizations) HP cells cycle life is almost independent of the...

endless-sphere.com

I suppose a point that I wonder about is that Samsung do already make cells (e.g.21700 33j for Tesla) specifically designed and tested for 2000 cycles at 4.1 volt which I have used so it appears possible to make a long life battery. There may be a possible issue with mechanical robustness on some 'powerwall cells' but there is no mention of this in the data sheet . (I have seen the warning 'not for automotive use' given by nkon on some prismatic cells.)
At present, these cell are in 'surplus' and can be bought very cheaply with nkon. The negative outer case is light and can only be laser welded so there is the option to buy them with copper tags.

 

[saneagle]

There is a long running discussion on 'Li-ion cells cycle ageing' on endless sphere. Although they have demonstrate that cell life can be extended considerably by reducing the 'Depth of Discharge' by keeping the cell working in the range 3.4 to 4.1 volts, they have considerably reduced the available capacity. It does appear that high density cells e.g Samsung 18650 35E or 21700 50E are particularly affected. I think the losses are at both ends (>4.1v & below <3.4v) and work out to about15% at each end.
One contributor (Pa***) reckons that good quality cells do not require or benefit from this much and that a .5v reduction to 41.5 volt alone is sufficient .(leave LVC at 3.1 volt I presume)
The discussion is long and ongoing. The table is at below about a quarter way down the page 12.

Li-ion cells cycle ageing

The principle of this issue was already formally described in this topic. So yes HP cells have significantly worse cycle life than most HE cells under low loads up to ca 1-2C continuous discharge. Generally(Although I do not like generalizations) HP cells cycle life is almost independent of the...

endless-sphere.com

I suppose a point that I wonder about is that Samsung do already make cells (e.g.21700 33j for Tesla) specifically designed and tested for 2000 cycles at 4.1 volt which I have used so it appears possible to make a long life battery. There may be a possible issue with mechanical robustness on some 'powerwall cells' but there is no mention of this in the data sheet . (I have seen the warning 'not for automotive use' given by nkon on some prismatic cells.)
At present, these cell are in 'surplus' and can be bought very cheaply with nkon. The negative outer case is light and can only be laser welded so there is the option to buy them with copper tags.

Somebody is misunderstanding something somewhere. The difference in charge between 4.2v and 4.1v is about 1% because the voltage drops very quickly between those two values. Try it on your own bike. Charge up to 42v, then ride your bike off your driveway onto the road and check again. It'll already be at 41v!

 

[Woosh]

Somebody is misunderstanding something somewhere. The difference in charge between 4.2v and 4.1v is about 1% because the voltage drops very quickly between those two values. Try it on your own bike. Charge up to 42v, then ride your bike off your driveway onto the road and check again. It'll already be at 41v!

it can be up to 10% but I agree, the discharge curve is much steeper between 4.2V and 4.1V than between 4.1V and 3.5V where it's pretty linear.

 

[Sturmey]

The action and explanation really takes place on the charge curve and not the discharge curve. A battery spends longer on the charger when charged to 4.2 volts and gets saturated with more electrons during the constant voltage saturation phase. More ions or whatever (I am not a chemist) are stuffed into the battery at 4.2 volts. The problem as such is not the voltage but the saturation of the anode and cathode. The voltage is really only the indicator for a particular chemistry so reading the voltage on the discharge curve to give an indication of what has happens during charging is not reliable. There is also a lag when reading battery voltages so batteries need time to recover in order to use the voltage to indicate the state of charge.
The is a link below that might also help.

BU-808: How to Prolong Lithium-based Batteries

BU meta description needed...

batteryuniversity.com

 

[Woosh]

A battery spends longer on the charger when charged to 4.2 volts

that is correct but it's because it's more difficult to push the same amount of Lithium into an electrode that is already 90%-95% full. Its internal resistance to charging goes up when it is getting full.
If you watch the amount of energy usage of a typical 36V 2A charger, you will see that it takes about 80W-90W from the mains until about 40V then gradually decreases to about 30W-40W when the output gets to 41V then continues to decrease but more rapidly until practically zero when the battery is full.

 

[StuartsProjects]

I am intrigued by the 24%.
I'll try to make a time lapse video about charging time / battery voltage/energy consumption over the weekend.

The amount of charge that goes into my battery from the 41V point to end of charge at 42V is 1.12Ahr, so even if the charge current was the full 2A all the way to the end, it tails off of course, it would take 34 minutes to go from 41V to end of charge.