Ni-MH battery life

[ColorMaker]

Is there a way to test a battery for cycle life without acually cycling them ? I am testing some D cell 9500 mah that specs for 1000 cycles. Using my current battery anylizer it will take 3 years to complete test.

JM

 

[SilverFox]

Hello JM,

Yes, but it is a little bit involved...

Normal NiMh cells will actually increase in capacity over around the first 150 cycles. After that, there is almost a linear decline in capacity. If you watch the capacity peak and then start to fall, you may be able to do a regression and predict the end point.

As a cell ages, its internal resistance increases. You can also use this fact to make a formula for the voltage drop off of the cell. I use 80% as an end point, so you could take a voltage discharge curve and drop the line down to where you run into 80% of the initial capacity. This would give you the voltage drop, which you should be able to work back to the cells internal resistance increase. Once again, tracking the change in internal resistance over some cycles may give you enough information to predict the end point.

If time is money and you are trying to accelerate the process because of target goals, there are battery analyzers that can be programed to automatically cycle the cell for you. I believe you can get into one of these for around $400.

Tom

 

[VidPro]

i would say "you cant test them LIKE they will be used without using them" no mater what you do.

because you could put them on a charger/discharger for 500 times, and it wouldnt be the same as when you use it, so what good will it do you to know that?

tests often dont involve any:
reverse charge <-- that'll kill it quicker than anything
time between, resting
temperature variances other than the charge temps
discharge capacity variations, depth or rate
Dropping the battery or unit it is in

variations can also be huge between:
charge alogrythm types
depths of discharge
rate of discharge
charge rates
time between
charge style (slow vrses smart)

and it might not test other important things, really important ones:
when the self discharge went to heck
when it would have fallen out of balance with a set
(again) how it will actually be used
similar current Styles, say like from drivers that INCREASE in current as voltage drops, or drivers that DECREASE in current as voltage drops, or standard restistive loads like incans.

and charge/discharge testing, keeps the chemicals cycled, so enthropy (or chaos) dont get a chance to occur.

so in reality, it would not only take 5 years to do the test, but 50 different types of test besides, otherwise what is learned in the test, might not translate well to reality. the reality that translates, is your own reality from use.

better to "Socket" the battery instead

 

[Mr Happy]

In reality, isn't the whole question of cycle life a bit academic?

Suppose the cycle life is 500. You use a set of cells for work, discharge them during the day, recharge them overnight. There are 250 working days in a year, so it will be 2 years before you hit 500 cycles. If you discard the cells at that point and replace them with new ones, the annualized cost is negligible.

Cells I own don't get cycled anywhere close to once per day. They are more likely to expire from age than cycles.

If I cycle a set once per month say, then after five years that will be 60 cycles. What is more important to how much the batteries degrade? Five years of aging, or 60 cycles of use?

 

[ColorMaker]

I am using a Cadex C7400ER to test I can cycle 1 time each day.

Anyone use TMK 9500 D cells ? I am getting good capacity results but its a low cost battery and I am not sure of the cycle life.

SilverFox in the Charger IC thread you directed me to another forum that was testing chargers. What tester were they using ? I liked the graphs the PC generated.

JM

 

[SilverFox]

Hello JM,

The were just charging their RC battery packs. The interest was on which chargers were using -dV, what values they used to terminate, and if they ended the charge differently, how they did that.

The software they used is available from Radio Shack and is used with a Radio Shack volt meter that has a RS232 output.

Tom

 

[ColorMaker]

Thank you