Battery Life: 23 charges of my Kalkhoff 18Ah suggest a live of 5+ years

[10mph]

Summary
The reduction in charge up time for my Kalkhoff 18Ah battery (~26V), which has occurred during the 23 charge ups in its first 11 months from new, suggests a reduction in capacity of 6% per year. If degradation continues at that rate I will have still have 12Ah capacity after 5 years.

Record of charge ups

In June 2011 when my Agattu was just 2 months old, I posted a plots of my first 9 charge up times,
(see: http://www.pedelecs.co.uk/forum/electric-bicycles/8976-kalkhoff-agattu-range-9-charges-18ah-battery.html#post109828)

The 18Ah battery is now nearly 11 months old and I have done 23 charge ups shown in the following plot:

The charger provided by Kalkhoff regulates the charging current to a constant 1.8 Amps until the voltage reaches about 29.35 Volts then it maintains a constant voltage with the current gradually dropping. The charger switches off when the charging current has been reduced from 1.8A down to about 0.8A. This is what I have called the ramp down, shown by the the light green bar at the fully charged end of each plot.

One can calculate the total Ah absorbed in each charge by multiplying the duration of each charge (in hours) by 1.8A, and making a small correction for the reduced current during the ramp down.

For charges numbers 9, 10, 14, 15, 16, 19 charging was terminated immediately LED 5 stopped flashing and became steady. The grey area on those bars is an estimate of the additional time the charging would have lasted had it been continued to the fully charged state.

On receipt of the new battery charge number 1 took place from the sleep state which judging from the charge duration must have been in a deeply discharged condition. The longer time for each stage of charge 1 could be due partly to additional current being taken to balance the cells. There is an animation showing cell balancing in the link I posted here.

Variation of charge up capacity

Charges numbers 3, 5, 13 and 21 were from the low voltage cutoff point, so on these charging cycles the battery had been completely discharged before charging commenced. I noted that the time taken to do these complete charges had fallen very slightly over the last 11 months. Fitting a straight line to these 4 points and extrapolating this rate of fall to 5 years leads to the prediction that the charge up capacity of the battery will then be reduced by one third to 12.6Ah. A linear extrapolation may not be correct, perhaps the rate of decline will change. I will know more at the end of the second year. The battery warranty is due to expire after 2 years. At the moment it looks as if the battery will easily outl;ife its warranty period.

Using just 4 data points to derive the rate of decline could be rather limited in accuracy, so I have plotted a second set of points, in red, for 13 charges from the moment when the 2nd LED becomes illuminated steadily up to fully charged. This is nominally 40% charged up to fully charged.

Extending both these lines to 5 years suggests a 29% reduction over 5 years red points with the blue points figure of 33%,which is a very reasonable agreement.

Coulombic or Faraday Effciency (see Faraday efficiency - Wikipedia, the free encyclopedia)

The Coulombic efficiency can be defined as Ah during discharge divided by Ah during charge. For Li-ion batteries it seems to be 99% or higher. Some authors quote lower values but invariably it seems that they are quoting energy efficiency which is indeed somewhat less because of the small voltage drops during charge and discharge.

It is of course discharge capacity that is important, but until I can make some direct measurement of the discharge capacity for comparison with the charge capacity, I will assume >99% Coulombic efficiency and so claim my measurements show that my Kalkhoff 18Ah battery may drop to 12Ah capacity after 5 years. This is a very much longer life than I had anticipated before purchase after reading many warnings to budget for a replacement in one to two years, even with low usage such as mine of 2 cycles per month.

 

[Scottyf]

Excellent Post and I love a good graph.
Useful information and what is the cell type in the Kalkhoff 18ah battery that you are describing?

5 years is then not a bad amount of time to have for the cost.

I still think you need to do more cycles to really take advantage of the cycle life / time of the batteries. 23 uses if hardly anything.

This is why I can live the the so called 300 cycles of Lipo becasue this equates to about a year of use.

Some good info either way.
Thanks for your informative post.

 

[flecc]

Thanks for the details 10mph, this is consistent with the claim of 1100 charges which for a weekday commuter charging every usage day means almost five years of battery life.

There's little doubt we are at last moving away from a two year life, for some e-bike batteries at least.
.

 

[Scottyf]

Here's a quick question which Flecc or 10mph maybe able to answer.

Once the so called 80% capacity is hit after say a 2 year cycle life. The battery must surely still work.
However does it just loose its capacity. Or does the load / voltage sag render the battery pretty useless in an Ebike application?

 

[morphix]

Blimey that's some good scientific work there 10mph. Mighty impressed with your data and charts. Want a job at CycleCharge? ;-)

 

[10mph]

It appears from some sources, that the internal resistance does increase a little due to ageing, particularly with number of cycles and probably also with just time. For example, there is a curve on page 12 of this http://www.eetimes.com/ContentEETimes/Documents/Schweber/C0792/C0792-1_15.pdf

The increase of internal resistances is manifested as voltage sag. So it will definitely make an old battery less useful, ie high peak currents will no longer be so successfully handled when the battery is not well charged up. I don't know exactly how significant this will prove to be practice. If we had values for the resistance increase and the required current one could try and calculate its effect on battery life

 

[Scottyf]

This is my point. Ive you buy a big enough battery with say an addtional 20 - 30% capacity in whats actually needed could you not just keep using the same battery for much longer anyway.

I have some Nicad / Ni-mh cells in telxon scanners for our stocktakes. Its apaarent there pretty useless and hold hardly any charge (Considering there about 6v 900mah packs) but we've been using them for stocktaking for about 10 years.
There tired and capacity is pretty awful but they do still work once charged up.

 

[flecc]

I have some experience of the resistance rise effect on some Phylion batteries. As 10mph says, they do become less useful than indicated by remaining capacity by the fall in voltage under load causing safety cutouts. To give an example on the very powerful (1000 watt peak) eZee SB setup, where using full throttle on an ageing battery caused cutouts, reducing to half throttle and assisting more extended the usefulness. Clearly the motor power factor is important, and the 400 watt peak Panasonic motor will be much better at getting extended battery life than many of the more powerful ones.

 

[Deleted member 4366]

When you've extrapolated forward on the graph " Capacity measured during charge up", you've made the assumption that the line of the graph is straight. The points you've plotted are not enough to make that assumption. My guess would be that it would be a parabola - accelerating over the years, which would mean a lot shorter life than predicted. It should be something like a laptop battery, which has the same cells, where, after a year. you can hardly notice any difference, after 2 years it's noticeable, and after 3 years, your lucky if you can get much use out of it.

 

[Morgann]

Hi Scott.

I'm one of those supporters but if you feel that your comments warrant an apology at all then, they must warrant a discrete one not tucked away like this at the bottom of a wall of text.

I guarantee that you will get grief for that.

Cut your last para and paste it in the removed post thread or start a thread.

Personally, I work in an A+E. you couldn't offend me with a chorus of 'Eskimo Nell'.

 

[banbury frank]

Hi Morgann

I must congratulate you on your job working in A+E Keep up the good work

we all take your department for Granted till we need It

Frank

 

[10mph]

When you've extrapolated forward on the graph " Capacity measured during charge up", you've made the assumption that the line of the graph is straight. The points you've plotted are not enough to make that assumption. My guess would be that it would be a parabola - accelerating over the years, which would mean a lot shorter life than predicted. It should be something like a laptop battery, which has the same cells, where, after a year. you can hardly notice any difference, after 2 years it's noticeable, and after 3 years, your lucky if you can get much use out of it.

The present straight line is unlikely to continue indefinitely. I will have a better handle on it after another year has elapsed. I had originally budgeted to buy a replacement battery just after the warranty expires after two years has elapsed. In fact my battery replacement fund stands at £275. Every time I do a measurement for a race I ask for a £25 contribution and I have done 11 measurements with the bike. I am now actually planning to reduce the charge as the battery is clearly going to be useful for quite sometime more than 2 years. Only time will tell if a linear extrapolation is appropriate. Perhaps the sharp drop off only occurs once the battery approaches 5 years.

I had planned to do my analysis of the charge ups in April after one year had elapsed. What prompted me to do it sooner was a post I read somewhere here a week or two ago, but can't now find (it must be on one of the quarantined threads), which seemed to imply that all lithium ion batteries decay at 16% per year. I looked at my data and it was obvious that mine is not decaying at that rate. So I thought it worth posting.

I must say that my search for good data test data on this topic reveals very little on the internet. It would be nice if I had information on the cell type and the life tests done by the cell manufacturer. Failing that all we can do is to read the marketing material and carry out our own tests.

 

[Morgann]

If I give up 1 Mars bar a day I will have the money for a new battery when this one dies and I will get better mileage... So there!

I have a graph.

 

[flecc]

What prompted me to do it sooner was a post I read somewhere here a week or two ago, but can't now find (it must be on one of the quarantined threads), which seemed to imply that all lithium ion batteries decay at 16% per year. I looked at my data and it was obvious that mine is not decaying at that rate. So I thought it worth posting.

That decay rate prediction has been falling like a stone over the last four years. Starting at around 35% but as much as 60% in some circumstances, it's very obviously been falling rapidly with each new generation of batteries, roughly in line with the number of charges predicted. The best ones seem to be well under 10% over the last year or so.

It would be nice if I had information on the cell type and the life tests done by the cell manufacturer. Failing that all we can do is to read the marketing material and carry out our own tests.

The Kalkhoff BMZ batteries generally use 18650 size cells made by Panasonic or Samsung, currently all Samsung. The 26 volt 18Ah battery will almost certainly use an array of 42 cells as 7s x 6p, each cell of 3000 mAh. I don't have any data on the cells, but you may find out something online.

 

[bode]

Finally I would like to thank the customers that expressed their support for 50cycles over the past few days.

Best regards

Scott

And a big thank you too, no doubt, to all those who pointed out the error of your ways...