32 day self-discharge test on Powerex 2700 AA cells

I've done a lot of testing on the PowerEx 2700mAH AA cells as I have twelve of them. Four are in use and the other 8 are "matched" cells sitting quietly in a 8 cell holder. After running forming charges on a Cadex (they call it extended prime) and other various forms of testing they have about 7 cycles on them.

The big test was to see self discharge rates on the entire pack of 8 cells. The Cadex will do a 24 hour self discharge test which I did on one of the other cells in use and it returned a 6% self-discharge in one day. Since I am not using the cells presently, I decided that I could do a self discharge test to see how well they performed after a month of laying around.

The PowerEx batteries were treated to a two-cycle charge/discharge rate of C/10 charge and C/5 discharge set up as a 9.6V series pack. The discharge limit under load was 1.00V and charge cutoff was set at .08V per cell of voltage depression. The temp cutoff was 45C and the test was done with a room temp of 22C.

32 days later, I set up the Cadex 7400ER analyzer to discharge at a C/5 rate (540mA) with voltage cutoff of 1.00V per cell. The test returned a reading of 81% which is a drop of 18% from the initial 99% capacity reading. Or, to put it in mAH readings; intially it gave a reading of 2673mAH and dropped to 2187mAH.

Overall, I am very pleased with the PowerEx 2700's as they have a much lower self discharge rate than the PowerEx, Energizer and Sony 2500's (Sanyo HR stamp on all of them) Those 2500's would be very close to dead after 32 days so the 2700's are a big improvement.

Remember that those 8 cells were charged / discharged in a 8 pack in series so the worse battery would lower the results. Initially they were done seperately but now live and die together as a battery pack. Since the only thing these batteries have done is testing and cycling, it is probably the best case scenario for them but I'll use them eventually in a 8AA to 2D Mag mod.

Any questions?

Hello BentHeadTX,

That is right in line (actually just a little better) than what is expected with a 0.7% per day self discharge. It is good to see some base line confirmation of the self discharge rate. It will be interesting to see if it changes with use.

Tom

Thanks for the tests.
However even more interesting would be to perform these tests after about say 50ish cycles of real use. Say have a pair used in a low current application such as mp3 player, and another pair used for high current application such as a flash light, then test them after 50ish cycles of use. I'd be interested to see how well they manage real life use and if self discharge becomes worse and by how much.

I'll try such a test on my 2K cells that I really abuse (as I gave up on 2.5K Energizers, already killed a couple of them).

bob,
I do use the other four out of the twelve in Peak Mediterranean 2AA work flashlights. They have approx. 800mA load on them during use. Not exactly a very high load but they have about 20 cycles on them so far. Maybe in a few months I can see how well they operate.
I do have a MillerMods 1.7 watt L1P that pulls 1.73 amps off a single AA NiMH cell, maybe torture one of those cells for two months and see what it does.

Thanks a lot for the tests BentHeadTX
This type of real life information makes the forum a great place.

I have a mix of 2Amp AA NiMHs I use in a 6.5-8Amp application, which is a real abuse. I'll put aside 3 pairs for a 1 month test.
When we look at SilveFox and other tests of brand new batteries they usually look great. However, I am finding that in actual use some (such as high capacity AAs) may degrade very fast or simply die outright. Hence the question that is being missed by most of the tests is "How tough is a battery?"
SilverFox shootout kind of hints at that with the 5-10Amp tests. I am going to try Eneloops next as they show one of the best results. From the standard 2Amp AAs at the moment I seem to get around 1.2Amp useful power (before device performance becomes too low).

Oooops,
How do I test remaining capacity after a month using BC900????? Hmmm.
I might have to improvise somewhat.

Hello Bob,

Unfortunately, the only measure of how tough a cell is involves cycle testing. That is a big job that takes a lot of time.

My tests at high currents reveal the internal resistance of the cell. Those with low internal resistance test to higher currents.

You can do a discharge test in the BC-900 by selecting the discharge mode when you stick the cells in. You have to read the mAh from the discharge before they complete the charge after the discharge or you will loose that information.

Tom

I was reading the 160 page Cadex book and it has a selection for destructive testing. It will charge/discharge at 1C until it reaches the percentage of capacity you select. If I wanted to see how many 1C cycles a battery will go to say 70% of new capacity I can program it to do so. Program it and press enter and wait a few weeks to get the cycle count.

Maybe I should grab one of my PowerEx 2500's and see how long it will go until the battery is toast. The holidays are coming so the analyzer will be freed up from it's normal duties.

Hello BentHeadTX,

Do you have the option of tracking mid point voltage during these tests?

I have found that the mid point voltage drops long before the capacity drops off. You may still have 90% capacity, but a mid point voltage of only 1.1 volts at a 1C discharge.

Tom

Tom,
Right, the tricky part is seeing discharge value before it start charging. I suppose I could try ....

I was wondering if self-discharge rate could be interpreted as a rough approximation of a "battery toughness" or conversely amount of damage to internals sustained. So I would measure remaining capacity (vs original rating) and self-discharge to try to estimate battery "toughness" after a certain usage pattern.

Certainly more sophisticated equipment combined with more sophisticated test plan would provide more accurate results. For myself I don't need precise values; approximate results would be fine too.

This simple test using BC900 is a compromise. No?

BentHeadTX,
Actually for myself I'd prefer to rely on a real use. I have a motor that pulls 6.5-8 Amps. Clearly its power draw is variable, based on my usage, so different than a steady 1C or any other rate test equipment would use. For instance, how well do batteries handle temporary 8-10A peak draws that this motor may do? I suppose flashlights would have a steady current draw, so much easier. That is why I want this real abuse test as an extreme test (as opposed to typical usage that is much less demanding).

Wait a second, it is easy. I just need to record time when started discharge, then substract charge time (after the switch to charge mode) times the discharge rate. Simple.

SilverFox said:

Hello BentHeadTX,

Do you have the option of tracking mid point voltage during these tests?

I have found that the mid point voltage drops long before the capacity drops off. You may still have 90% capacity, but a mid point voltage of only 1.1 volts at a 1C discharge.

Tom

Click to expand...

Tom,
The way our Cadex 7400ER is setup to print battery labels with capacity and date. It can be connected to a computer to give endless graphs and vomit information if need be. If I pick up the Cadex book I get this look from the guys at work They always wonder what I am messing with at that point. The analyzer is not presently setup to do charts and graphs as there is not computer in the area that it lives. It generally tests medical batteries and charges cell phones on occasion.
Time to check the book to see what information it will provide with destructive testing. My hope is the programming CDs and cables are still around so I can see what it will do. I do know that a constant 1C charge and 1C discharge with no rest time will really keep the cells cooking. In theory, I should get 12 complete cycles per day and they should show heavy damage after 10 days to two weeks. The farther down the capacity goes, the faster the cycles. Maybe set it up on Friday and run it till after new years should get me 250 cycles or more.
I can also program the max discharge level to 1.1 volts which will give an indication when the mid-point is being reached. Since I am in the mood to destroy one of the PowerEx 2500's, should I go with a 1.1V max discharge level to determine wear and tear?

Hello BentHeadTX,

Most electronic equipment shuts off at around the 1.1 volt range, so your test would have more real world application by going that route.

On the other hand, I believe the industry standard is to discharge to 1.0 volts.

Tom

SilverFox said:

Hello BentHeadTX,

Most electronic equipment shuts off at around the 1.1 volt range, so your test would have more real world application by going that route.

On the other hand, I believe the industry standard is to discharge to 1.0 volts.

Tom

Click to expand...

I'll go with the 1.1V discharge termination and 1C charge/discharge rate and get the stats. This is going to take awhile

I use the refresh mode on the BC-900 for discharge tests. It will hold the highest discharge capacity measured in that mode. I generally try to check it fairly early in the following charge cycle and then manually cycle the BC-900 into another discharge test cycle or charge mode.

All the manufacturers seem to spec their NiMH cells for a discharge to .9V under load.

Duracell and Sanyo both have data in their tech docs on "memory effect" aka voltage depression for NiMH cells only discharged to 1.15/1.20 V. The data shows up to a 15-20% capacity loss to that discharge voltage after 20 cycles. They show it being fully reversible after a discharge to .9V under load and a full recharge. They state the effect is larger at higher termination voltages. At 1.1V cut off, I suspect it will still be present. Energizer and Maha state their NiMH show no "memory effect"...

I'd post a graph here but I can't figure out a way to attach a jpg image.

Mike

If you keep your NiMH's in the 'fridge, will it extend their shelf-life ?

I attempted to do a "life cycle" test with the Cadex to determine how well my one-year old PowerEx 2500's (Sanyo HR) hold up to a 1C charge and 1C discharge 2.5 amp rate. The first test I selected a 1.08V max discharge voltage level and a 50% capacity limit to hault the test. Figured it would take about a week or so to get results. I was wrong!

Capacity 2% and test terminated. Hmmmm, so I watched it as it cranked 2.5 amps into the cell for charge and switched to discharge after hitting 36C and 1.53V. The voltage plummeted as it was hit with the 2.5 amp discharge level. After a little more than a minute the voltage sagged to 1.08V and the test terminated.

Tried again and set the discharge limit to 1.00V and it was back to the races. A little while later the analyzer started to alarm and it gave me the results. Capacity 8% I ran it again... another 8% popped up and test terminated. One more time! Yep... it dropped to 7% and terminated. Tried another 2500 and it terminated at 9% so I came to a conclusion.

So there you have it weeks earlier than I anticipated, the total number of cycles until one year old HR stamped 2500's drop below 50% capacity is zero, nada, nothing and zip. The 2500's had about 20 to 25 cycles on them and generally charged at 1000mA and were loaded at around 800mA for discharge into a LED flashlight.

Figured that for CPF, that would be a very easy result... I went with 1250mA charge/discharge rate and let it rip. It was cycling the rest of the day without alarming that it dropped below 50% of capacity. I think it was showing 78% or something on the first cycle so those 2500's will work but not at loads of 2.5 amps. They have a battery label on them for March 06 that shows 91% capacity at a 1.75 amp discharge rate (My MillerMods L1P 1.7 watt pulls about that from a single AA cell)

I have 8 PowerEx HR stamped 2500's and two of them out of the sample give very close results. The HR stamped PowerExes are the best performing out of the Sony and Energizer 2500's but that ain't saying much! The Sony cells would discharge in a week or two so they landed in the recycle bin.

Although the 0.5C life cycle test continues on the HR 2500's, I do know that my MillerMods L1P should not be fed those cells. The one cell I have been using is a 2500 which gave the result of 93% at a 1.75 amp discharge rate the 28th of Sept 06 so I will test that one again (battery label printers rock!) I want to see how far down it has gone in the last four months and report back. I think I will switch to Titanium 2400's that I have laying around for my helmet-mounted 1.7W modded L1P.

I'll go into work tomorrow to see how well the 2500's are fairing at 0.5C, they might be fine.

BentHeadTX said:

(battery label printers rock!)

Click to expand...

Don't mean to change the thread, but is there a printer you'd recommend? I'm expecting the new Maha analyzer/charger, so this is timely.

Thanks,
Jack

bob_ninja said:

Wait a second, it is easy. I just need to record time when started discharge, then substract charge time (after the switch to charge mode) times the discharge rate. Simple.

Click to expand...

Or,you could just charge them up. The BC-900 will tell you how many MAs were required to get to full.