AA and AAA NiMH low self-discharge tests--Round 2

Low self-discharge battery tests--Round 2

In a previous thread, I provided test results and interpretative observations on a few brands of AA and one brand of AAA low self-discharge NiMH batteries.

http://forums.dpreview.com/forums/read.asp?forum=1023&message=22988426

People posting in that thread pointed out that there are some new brands since my first selection. Also I've gotten some encouragment from people finding those measurements and postings useful.

So I've purchased 4-sets of six different as yet untested brands of AA and sets of five untested brands of AAA, and intend to test them in very similar style to my first round of tests. I intend to present the results for all eleven sets of AA and all six sets of AAA together (rounds 1 and 2), to ease comparison.

I'll document my intended methods at the start here, and plan to myself initiate several subthreads:

1. Methods--as I can't edit my posts, this thread will document any changes or added detail to my plans from this post--also any further added brands.

2. Graphs--I'll make a post here for graphs which compare capacity and retention by readout by brand. These are intentionally not in the first post, so folks won't be waiting for them to load each time they visit the thread. I'll use GIF format for the images of the graphs, and expect download size on the order of 10kbytes

3. General discussion and interpretation--for overall comments about this sort of battery, the testing and methods, references to test results posted by others...

4. Initial Readout--Capacity measured on the fresh batteries as received.

5. One hour retention--the capacity measured after a round of breakin cycling, then a standard charge, with a one-hour delay from end of charge to start of discharge. The number logged at this readout is the reference for percent capacity retention at longer hold time readouts.

6. One Week retention--no further cycling, just a standard charge after the one-hour discharge. Hold time from end of charge to start of discharge very close to 168 hours.

7. Four Week retention--similar interpretation to one week, but without so much care given to matching the hour of the day.

8. Three month retention--again like One week, intended to start discharge 91 days after charge, but some errors up to three days have happened.

Schedule--rough estimates
Intial readouts, time zero, oneweek--December 18, 2008
Four week--January 25, 2008
Three month--May 1, 2008

Charge and Discharge Details.

All the charging and discharging for actual measurements is done by my Maha Powerex MH-C9000. This unit uses a pulse waveform on both charge and discharge. I don't know the exact time scale or duty cycle, but believe the time scale is roughly one second per pulse, and the duty cycle roughly 90%. The driven charge or discharge voltage is adjusted in real time to try to reach the specified charge or discharge current during the active part of the cycle. The cell voltage reported on the display (and used to control discharge termination) is measured at the end of the rest interval.

Termination of charge, as with most recent "smart" chargers is primarily by a negative delta-V criterion, with a safety backup thermal termination that in real life does not activate unless very high charging rates are used.

Termination of discharge is at 1.00 volt, measured after a few tens of milliseconds of rest. The compliance limit on discharge, the lowest voltage to which the unit will drive the battery, is 0.8V. If the prescribed rate is not attained, even with that level, then 0.8V is used, and the lower actual current (than set) is reported on the display, and used in the measured capacity calculation. This can lead to a rather long slow tailoff of discharge for batteries which exhibit high discharge impedance late in the discharge cycle. I've seen this behavior very strongly in many of my used conventional NiMH batteries. So far I have not observed it to any appreciable degree on the low self-discharge (and almost unused) subjects of this test.

For AA batteries the charge condition is 800 mA, and the discharge condition 200 mA.

For AAA batteries the charge condition is 300 mA, and the discharge condition 100 mA.

While the great majority of charge cycles are expected to give automatic termination, I'll terminate charge manually at approximately 130% of nameplate capacity where required.

As a defense against error caused by initial use cycle capacity changes seen on some NiMH batteries, I'm cycling the batteries between the initial readout and the first one-hour retention readout.

On the first round I performed this cycling in the MH-C9000, using three cycles at 1200 mA Charge/500 discharge for AA. For all the candidate batteries in the first round I saw negligible shifts in this cycling.

So to save a little time and charge capacity, I shifted this task in the second round to my LaCrosse BC-900. I've used its Discharge/Refresh cycle, with the discharge current set at 500 and charge at 1000 for AA, and discharge at 250 with charge at 500 for AAA. On the BC-900, the number of cycles is not fixed, as the unit does another cycle if the most recent cycle had higher capacity than the previous one. As I present the cells to the charger in a discharged state, this means that all cells get a minimum of three cycles from this exercise, quite a few get four (typically with capacity increase of one or two percent) and a few get more cycles.

Summary Statistics

While in the first round I presented each individual battery readout, I've decided this time to present the data as set mean, and set range.

The range calculation is provided as a measure of within-set variation. It is the difference between the max and min in set, divided by the set mean.

Brands and Tags

I'm using a consistent set of alphanumeric tags to label the data. Here is the list, the capacities listed are nameplate capacities.

AA NiMH low self-discharge
First Round
ENL--Eneloop by Sanyo 2000 mAh
TPE--Titanium Power Enduro 2100 mAh
HYB--Hybrio by Ultralast 2100 mAh
ACC--Acculoop 2100 mAh
RHY--Hybrid by Ray-O-Vac 2100 mAh

Second Round
DPC--Duracell Precharged 2100 mAH
ENO--Nexcell EnergyON 2000 mAh
KPC--Kodak Pre-Charged 2100 mAh
AMX--Ansmann Max-e 2100 mAh
IME--MAHA IMEDION 2100 mAh
GRY--GP Recyko 2100 mAh

AAA NiMH low self-discharge
First Round
HY8--Hybrio by Ultralast 800 mAh

Second Round
IM8--MAHA IMEDION 800 mAh
AC8--Accupower Acculoop 800 mAh
EN8--Sanyo Eneloop 800 mAh
DP8--Duracell Precharged 800 mAh
GR8--GP Recyko 800 mAh

AA NiMH used conventional batteries (individual cells provided as a comparison reference)
First Round
KOD--Kodak 2100 mAh NiMH
RAY--Ray-O-Vac 1600 mAh NiMH
NEX--NEXcell 2200 mAh NiMH
TI24--Titanium 2400 mAh NiMH

Second Round
TI26--Titanium 2600 mAh NiMH
LC20--LaCrosse 2000 mAh NiMH
SY23--Sanyo 2300 mAh NiMH
PW20--Powerex 2000 mAh NiMH

Prices--in the first round I listed the purchase prices, with allocated shipping. However, at the moment the great majority of these seem to run around US $12.00 for a set of four, plus applicable shipping and taxes, so I'll just note the larger exceptions.

Low Price

Kodak Pre-Charged AA available at my local Wal-Mart for $7.99 for a set of four--also that low price at Thomas Distributing's web site.

High Prices as found

Duracell--found on the shelf at my local Staple's chain, at a price of almost $16.00 for a set of four--same for AA and AAA.

Gp Recyko

Found only on the net from Hong Kong. Price including shipping to US about $14.00 for set of four--same for AA and AAA.

Graphs

My intent is to update these graphs as new readouts are available by posting them to the same location with the same name.

That way the new content should just appear on this page.

Caution: In order to show the initial capacity as delivered, I've adopted the convention of calling that -500 hours. This readout is not available for the reference conventional batteries included in the test.

The planned readouts are:

-500 hours capacity as delivered
0 hours zero-time capacity after at least three cycles
168 hours--one-week capacity
672 hours--four-week capacity
2184 hours--three-month capacity

Excel will show the value for readouts not yet taken as zero, so expect to see a bunch of curves drop to the bottom for the next readout.

You will need to click on the graph for dpreview to provide the full width.

AA batteries:

AAA batteries

I may make additional graphs--if so I'll post links to them in this subthread.
--
new 350D user July 1, 2005

Perhaps the biggest reservation I should make regards sample size.

Generally I've only provided data for one set of fours batteries of each type, usually purchased in a single package.

I did buy several more sets of Hybrio AAs after their good performance in the first round, and they were pretty consistent, which gives hope.

Still, I think my results are more capable of confirming that this new category of low self-discharge NiMH batteries differ from their conventional cousins than pinpointing small differences of specific brands in most cases.
--
new 350D user July 1, 2005

Initial Readout

These readings are the capacity logged just as the batteries came out of their shipping packaging.

As such, they give some validation of the manufacturer's description of batteries in this category as "pre-charged", or as having useful capacity as delivered.

Capacity and variation

AA batteries

batt Init Var
ENL 1438 0.010
TPE 1009 0.582
HYB 1619 0.058
ACC 1737 0.006
RHY 1738 0.018

DPC 1515 0.013
ENO 1189 1.337
KPC 1723 0.019
AMX 1645 0.030
IME 1831 0.019
GRY 1687 0.009

AAA batteries

batt Init Var
HY8 593 0.062
IM8 727 0.048
AC8 543 0.015
EN8 600 0.017
DP8 622 0.008
GR8 723 0.014

The extremely large variation numbers for TPE and ENO signify that each of those sets contained an outlier battery as delivered.

In the TPE case, the outlier had higher capacity--1415 compared to about 850 for its three partners.

In the ENO case, the outlier had zero initial capacity--while its three partners were tightly clustered near 1580.

The ENO bad battery showed substantial recovery during my intial cycling before one-hour readout. The BC-900 reported it as delivering the following on successful cycles:

1 0
2 125
3 709
4 1947

One hour retention

Capacity and variation

AA low self-discharge

battery 1 hour var
ENL 1980 0.010
TPE 1932 0.074
HYB 2026 0.006
ACC 2022 0.002
RHY 2021 0.017

DPC 2010 0.011
ENO 2017 0.013
KPC 2124 0.018
AMX 2078 0.027
IME 2186 0.035
GRY 2116 0.017

AA conventional (single batteries for reference)

KOD 2072
RAY 1579
NEX 1239
TI24 2109
TI26 2001
LC20 1727
SY23 2091
PW20 1963

AAA low self-discharge

battery 1 hour var
HY8 799 0.016
IM8 870 0.017
AC8 779 0.015
EN8 838 0.016
DP8 833 0.006
GR8 912 0.027

Comments:

The IMEDION batteries were clear winners for capacity in the AA category for this readout. The GP Recyko batteries won for AAA

Also, I noticed that the initial voltage at zero load as delivered was much higher for the IMEDION batteries than others. While I did not log the data, I did put them all on my meter.

The IMEDION cells were about 1.35 volts at that stage, while many others were only about 1.28, and few were much over 1.30.

One Week retention

AA NiMH low self-discharge sets

batt mAh var
ENL 1847 0.012
TPE 1847 0.044
HYB 1918 0.005
ACC 1899 0.004
RHY 1867 0.020

DPC 1906 0.012
ENO 1904 0.011
KPC 2021 0.022
AMX 1936 0.033
IME 2053 0.030
GRY 1992 0.018

conventional AA reference cells

KOD 1917
RAY 1471
NEX 1211
TI24 1859
TI26 1745
LC20 1595
SY23 1899
PW20 1804

AAA NiMH low self-discharge sets

batt mAh var
HY8 738 0.009
IM8 810 0.022
AC8 724 0.007
EN8 783 0.015
DP8 780 0.006
GR8 846 0.025

The IMEDION and GP Recyko AAA cells exceeded their name-plate capacity after a week. It appears that in general the AAA market has had less of a capacity horsepower race than the AA market. Perhaps in consequence, the ratings are much more conservative.
--
new 350D user July 1, 2005

thanks for your work on this, most interesting, Merry Christmas, hope the new year is very good for you.
regards - tom

Yesterday, December 21, 2007, I visited a Radio Shack store in Christmas Shopping.

They had on display "pre-charged" AA and AAA NiMH batteries. The price was extremely high $19.95 US for four batteries. I am not eager to enter new brands at this stage, am not optimistic about Radio Shack battery quality after experience in recent years, but aware that this may be the most readily available option in this category for at least USA users. So I compromised and purchased the 4-pack of AA at $19.99 for this study, but not the AAA.

These will be called "RSK" here.
--
new 350D user July 1, 2005

adding in the Radio Shack AA to the AA results, we have:

battery initial var
ENL 1438 0.010
TPE 1009 0.582
HYB 1619 0.058
ACC 1737 0.006
RHY 1738 0.018

DPC 1515 0.013
ENO 1189 1.337
KPC 1723 0.019
AMX 1645 0.030
IME 1831 0.019
GRY 1687 0.009
RSK 1505 0.027

Given the very high purchase prices, the fact that Radio Shack badges their product at 2000 mAHr, while nearly all the others save the Eneloops claim 2100 is not encouraging. The higher price of $19.99 is also a matter of dismay, which the initial readout does nothing to dispel.
--
new 350D user July 1, 2005

Santa can come through for you to at least make up for the Radio Shack AA's you shelled out for. Great work you are doing and I can't thank you enough for taking the time to not only do the testing but also post it so we can benefit.

You and your family have a happy and safe holiday!

Update to add the Radio Shack batteries to the AA entries

Capacity and variation
battery 1 hour var
ENL 1980 0.010
TPE 1932 0.074
HYB 2026 0.006
ACC 2022 0.002
RHY 2021 0.017

DPC 2010 0.011
ENO 2017 0.013
KPC 2124 0.018
AMX 2078 0.027
IME 2186 0.035
GRY 2116 0.017
RSK 2024 0.008

KOD 2072
RAY 1579
NEX 1239
TI24 2109
TI26 2001
LC20 1727
SY23 2091
PW20 1963

For this test, the Radio Shack performance is entirely respectable, though not at all justifying the price premium.

Update to add Radio Shack to AA 1-week results

battery 1 week var
ENL 1847 0.012
TPE 1847 0.044
HYB 1918 0.005
ACC 1899 0.004
RHY 1867 0.020

DPC 1906 0.012
ENO 1904 0.011
KPC 2021 0.022
AMX 1936 0.033
IME 2053 0.030
GRY 1992 0.018
RSK 1868 0.012

KOD 1917
RAY 1471
NEX 1211
TI24 1859
TI26 1745
LC20 1595
SY23 1899
PW20 1804

The Radio Shack batteries had a lower percent capacity retention at the 1-week readout compared to the 1-hour readout than most of the competitors:

92.3% mean across the set. Typical performance for this category is about 94%

While small errors in the readouts could drive this, it does not bode well for the Radio Shack being at all competitive.

Already the most expensive of the lot, it seems likely to turn out to be one of the poorest performers. Already at the 1-week readout these have the lowest capacity of the second-round competitors.

But they may be the most readily available for many USA users, so I'll continue reporting them out to three months.

Please keep us informed. (Shameless bump )

You deserve a medal for all this!

At some stage it would be useful if you could provide a non-technical non-brand specific summary of your findings. Whilst I am not put off by your straight forward relatively non highbrow reports there is a whole world out there that could benefit from knowing your findings but would be frightened of your figures.

Can I suggest that at some stage in another thread you exclude the best and worst performers both pre-charged and ordinary NI-MH and provide a short two paragraph summary of your findings to date comparing standard NI-MH and pre-charged and generalising your findings.

This is great timing, as I am thinking of going down this road to power my new SB-800
--
Thanks,
Glenn

http://KerenOwens.com

Chris Elliott wrote:

At some stage it would be useful if you could provide a non-technical
non-brand specific summary of your findings.
--
Chris Elliott

http://PlacidoD.Zenfolio.com/

Thanks for your kind words. I did write a "Conclusions" section to my three-month post in the previous thread.

http://forums.dpreview.com/forums/read.asp?forum=1023&message=24756026

I admit it is a bit buried there. Part of my thinking in setting up a "General Discussion and Interpretation" subthread for this post was to provide a more obvious place for such things.
--
new 350D user July 1, 2005

Kodiak wrote:

Please keep us informed. (Shameless bump )

I've avoided posting while only part of the data from a particular time-span is in hand.

I have updated the graphs (which does not require a new post) somewhat more frequently. Just before writing this post I updated the AAA graph and AA graph with all results to date.

I'll post comments in the one-month result thread when the GP Recycko AA reads out about January 16, and update them when the Radio Shack straggler reads out about February 1.

Three-month initial text result post will be about U.S. income tax day (April 15).

Four Week retention

repost--it appears the original was in the irretrievably lost posting period.

AA NiMH low self-discharge sets

batt mAh var

ENL 1811 0.013
TPE 1795 0.020
HYB 1870 0.007
ACC 1790 0.016
RHY 1795 0.018

DPC 1857 0.011
ENO 1854 0.012
KPC 1950 0.021
AMX 1856 0.031
IME 2008 0.030
GRY 1938 0.024

conventional AA reference cells

KOD 1673
RAY 1273
NEX 978
TI24 836
TI26 1495
LC20 1498
SY23 1615
PW20 1677

AAA NiMH low self-discharge sets

batt mAh var

HY8 696 0.007
IM8 791 0.024
AC8 697 0.007
EN8 765 0.017
DP8 757 0.003
GR8 824 0.028

Among the AA cells the Maha IMEDION have the best 4-week results. The Kodak are respectable second, and low price and high availability will make them the preferred choice for many.

The GP Recyko AAA cells exceeded their name-plate capacity after four weeks! It appears that in general the AAA market has had less of a capacity horsepower race than the AA market, leading to more concervative ratings.