NiMH Rechargeable Battery Comparison Using Kitronik Inventor's Kit and Adafruit CLUE

This article looks at how nickel metal-hydride (NiMH) rechargeable and non-rechargeable batteries can be tested to measure their capacity using the constant current tester made in Battery Capacity Measurement Using Kitronik Inventor's Kit and Adafruit CLUE. For many uses in the home rechargeable batteries are more cost-efficient and less wasteful than non-rechargeable ones.

The tester's limitations for its constant current mode are explored and the article shows the replacement of the load resistors enabling testing at higher currents. The tester's breadboard implementation is far from perfect but can be used with care for a reasonable analysis of battery capacity and a good relative comparison. It can be used to compare different brands of non-rechargeable battery, to determine the health of an aging rechargeable battery and to verify claims about capacity.

Newly purchased AAA size batteries from Varta, Energizer, Panasonic, Carrefour, IKEA Ladda, Ansmann, EBL, HiQuick and MaximalPower were tested. The capacity results for rechargeable batteries are compared against those from an inexpensive commercial charger with a test function. Capacity is commonly quoted in mAh, a measure of charge. A small minority of brands place a single prominent number on their batteries in mWh which happens to be a larger number.

The testing yielded some important observations.

1. Some less well-known brands of batteries are marketed with inaccurate, overly high capacity claims.
2. Pre-charged batteries will not arrive fully-charged and more importantly the charge level can sometimes vary considerably. It is best to top-up the charge just before use to obtain an even charge level across the set to prevent damage from some being over-discharged in use. If your charger has a test function with a charge/discharge/charge sequence, this is an even better way to first charge them.
3. A charger with a (low) 200mA charge setting may be unable to determine when to terminate the charging process. This is likely to result in overcharging of the batteries which will reduce their lifetime. A 200mA rate is likely to only be suitable for the low-capacity 500-550mAh AAA NiMH batteries sometimes sold as "phone batteries". One third of the actual capacity is a useful minimum for a slow charge rate, i.e. 300mA for a 900mAh battery.

A future article will cover the non-technical issues surrounding the capacity claims.

• See Instructables: Battery Capacity Measurement Using Kitronik Inventor's Kit and Adafruit CLUE and in addition to that
• A battery holder with leads.
• Some reputable and perhaps some less reputable batteries.

This is a test of a Varta AAA alkaline (also known as LR03, the IEC name) battery in a room at 24-26 Celsius. The test ran to a cut-off voltage of 1.0V but as primary cells can safely be used to exhaustion the test was continued in two steps to 0.9V and then to 0.1V. These were purchased from Amazon UK. They are labelled as made in Germany.

The discharge test has a target current of 40mA. This is a fairly low current for an AAA battery test, the value is due to the use of the pair of 47 ohm resistors from the Kitronik kit. The load resistance presents a cap on the current which can make it impossible to maintain the target current at low voltages. The first chart shows how the 40mA current cannot be maintained from around 0.94V and the tester becomes a constant resistance load below that voltage. The voltage drops steadily throughout the discharge. The voltage drop is a mixed blessing, it's useful as an indicator of remaining capacity but many devices will have a minimum voltage to operate correctly, preventing the use of the remaining charge/energy.

The calculations on the final chart show that there's an extra 26% energy in the battery between 1.0V and 0.1V. Few devices will be able to make use of a battery at very low voltages but it's interesting to see how much remains in this region. The capacity of a battery is typically quoted as a charge value in mAh but sometimes this is supplemented with the energy value in mWh.

Chemical Leakage

Alkaline batteries were marketed as not leaking (relative to zinc carbon batteries) but when completely discharged they can eventually leak caustic potassium hydroxide which will quickly "scrub" the carbon dioxide in the air to form white crystals of potassium carbonate. This is certainly not the best way to perform carbon capture! A long-term test seeking leakage is shown in EEVblog #1274 - Long Term Alkaline Battery Leakage Testing with a follow-up in EEVblog 1508 - We FINALLY Got Alkaline Battery LEAKAGE!.

Varta has several tips on avoiding leaks listed below.

1. Do not remove a cell or battery from its original packaging until required for use.
2. Do not mix old and new batteries.
3. Insert the batteries properly (correct orientation to achieve the correct polarity).
4. Remove the batteries from the device if they are not used for several months.
5. Immediately remove exhausted batteries from your device and dispose of properly.
6. Do not attempt to recharge a primary (non-rechargeable) battery.

The guideline on not mixing batteries also applies to not mixing non-rechargeable ones with rechargeable ones.

This is the same as the previous Varta test but in one continuous discharge to 0.1V at 24-26 Celsius. These were purchased from a large Carrefour supermarket in France. They are labelled as made in China.

The results are remarkably similar to the Varta ones with a capacity of 1546mAh vs the Varta's 1498mAh. The energy values are even closer, 1598mWh vs Varta's 1575mWh.

This is a test of two Panasonic zinc carbon AAA batteries (also known as R03) at 21-23 Celsius. One package was purchased in a small local shop in the UK, the other from a market stall - the latter had just expired based on the date printed on the batteries! Both are labelled as made in Poland.

The voltage profile is very similar to the alkaline batteries but the capacity is significantly lower at around 35% of the alkaline capacity. The expired battery has a reduced capacity compared to the fresh one, 5% lower in both charge and energy terms.

A few decades ago it was common to have a choice of zinc carbon or alkaline batteries but zinc carbon are rare now in many countries. In the UK, the red Panasonic zinc carbon range can be found in small convenience shops. Varta also manufacture them, selling them as "Super Heavy Duty" batteries and recommending them for "low-energy devices" meaning ones with a low current draw.

The spider chart above is from GP Batteries NiMH Rechargeable Batteries Catalog & Selection Guide illustrating how properties vary in their NiMH range.

NiCd

Rechargeable batteries have been available for many decades for use in the home. The first ones using NiCd (nickel cadmium) battery chemistry:

• have a lower nominal voltage of 1.2V per cell;
• suffer from a capacity drop commonly referred to as memory effect if the discharging and charging are not carefully controlled per cell;
• and quickly lose charge once removed from the charger (self-discharge).

The lower voltage is more significant when multiple batteries are used in series, for example 6 alkaline AA make 9V but 6 NiCd batteries make 7.2V. This can make them incompatible with devices designed originally to use zinc carbon or alkaline batteries but this should be rare because the voltage of non-rechargeable batteries progressively drops as shown in previous charts. They are likely to confuse a battery level indicator if there's no rechargeable battery setting or automatic guess on battery type. These factors left the batteries more suited for industrial/professional use and perhaps tarnished the image of rechargeable batteries in the eyes of consumers for a few years.

NiMH

The advent of practical mass-market mobile phones like the Nokia 2110 was enabled by a new battery chemistry, NiMH (nickel metal hydride), also used in the first models of the Toyota Prius hybrid vehicle. The voltage is the same as NiCd but the

• energy density is better,
• Cadmium-related environmental issues are no longer present,
• the charging regime is more flexible,
• and smart chargers are now more common and more affordable.

The majority of modern rechargeable batteries sold to consumers are now NiMH. There are a few NiZn batteries and some exotic Lithium-based batteries with embedded electronic circuits and USB (only) charging.

The Energizer Nickel Metal Hydride (NiMH) Handbook and Application Manual is a good read for more information on these batteries. For important properties, it mentions:

• rated voltage and discharge profile,
• capacity and variation with drain level,
• self-discharge rate,
• lifetime in recharge cycles,
• lifetime in years,
• performance variation with temperature,
• weight,
• recycling.

Eneloop (was Sanyo, now Panasonic) includes some statements about their AAA batteries in their marketing material, a subset of this is shown below which suggests increased capacity is linked with lower cycle life:

• eneloop lite: min 550mAh (or 680mAh), 3000 cycles, self-discharge to 85% in 1 year, 70% in 5 years.
• eneloop: min 750mAh (or 800 mAh), 2100 cycles, self-discharge to 70% in 10 years.
• eneloop pro: min 900mAh (or 930mAh), 500 cycles, self-discharge to 85% in 1 year.

Some brands have marketing which heavily emphasises a capacity value on their products which may lead consumers to strongly differentiate products using this value. This is easy to test using a constant current load to verify these claimed values. The test conditions for these values aren't always cited but a discharge rate around 0.1C (c-rate of 1/10) to a cut-off voltage of 1.0V is a good starting point. The IEC 61951-2 discharge tests to determine capacity are performed at 0.2C which will give a slightly lower capacity:

The battery shall be charged, according to Section 5.1 of this test methodology. After charging, the battery shall be stored in an ambient temperature of 20 °C ± 5 °C for not less than 1 hour and not more than 4 hours. The battery shall then be discharged in an ambient temperature of 20 °C ± 5 °C at a rate of 0.2C, where C is the rated Ampere-hour capacity of the battery. The test shall continue until the battery pack reaches its end of discharge voltage, according to Table 1 [1.0V for NiMH).

This is a test of IKEA's own-brand Ladda NiMH AAA batteries (also known as HR03). These were tested to a cut-off voltage of 1.0V between 22-25 degrees. These were purchased from an IKEA store in the UK. The claimed capacity of 750mAh equates to 0.053C for 40mA.

The new batteries had a near-identical pre-charge capacity and these immediately met their claimed capacity.

These were tested to a cut-off voltage of 1.0V between 22-25 degrees. These were purchased from the Amazon UK marketplace seller HiQuick. The claimed capacity of 1100mAh equates to 0.036C for 40mA.

The new batteries had a close pre-charge capacity. They failed to meet their claimed capacity by a wide margin.

These were tested to a cut-off voltage of 1.0V between 22-25 degrees. These were purchased from the Amazon UK marketplacer seller EBL UK Online-retailer. The claimed capacity of 1100mAh equates to 0.036C for 40mA.

The new batteries had a very similar pre-charge capacity. They failed to meet their claimed capacity by a wide margin.

These were tested to a cut-off voltage of 1.0V between 22-25 degrees. These were purchased from the Amazon UK marketplace seller ANSMANN UK. There's no prominent country of manufacture on the minimalist, shrink-wrap packaging, the advertising images have "German Quality Check" on them but they appear to be made in China. The claimed capacity of 1050mAh equates to 0.038C for 40mA.

The new batteries had a huge variation in pre-charge capacity. They were close to the claimed capacity but did not meet it.

One test abruptly dropped in voltage but this was due to the measurement of a rogue voltage of 0V which caused a premature test termination. The software should be more cautious on termination and only terminate after a few seconds below the cut-off voltage.

The 40mA discharge results did not match the expected values with the exception of the IKEA Ladda batteries. The Nordic Power Profiler Kit II excels at looking at patterns of power use and relative differences but it is not described as a high accuracy device for absolute values. It was enlisted for a second opinion on the current passing through the battery tester in case something had gone (very) wrong.

The photograph of the screen shows the summary of 40.82mA over a 7 second period at 1.250V - the actual plot is cropped out. The current calculated by the battery tested for the last 8 seconds is shown and is averaging 40.02mA plus 0.125mA for potential dividers. This is reassuring as there's just under a 2% difference. If this was 30mA or 50mA it would be a concerning difference and worth investigating.

The two chargers used in this article are shown above. Both are powered by 12V supplies, probably based on an original design intended to allow powering from 12V lighter sockets in cars.

• Charger 1, left, silver: Uniross Smart 2-4H, purchased from Amazon in 2008, independent connectors charge AA at 700mA and AAA at 350mA, only charges 2 or 4 batteries at a time selectable via switch, clear bi-colour LED indication of charging, charged or battery fault.
• Charger 2, right, black: Uniross Ultimate Battery Charger UCQ001, purchased via Ebay in 2023, four independent channels but cannot detect battery size, charges at 200, 500, 700 or 1000mAh, has discharge testing at half aforementioned rates.

The first charger's limitation of charging 2 or 4 batteries suggests it is charging batteries in series which isn't ideal but will be okay for batteries of the same brand, capacity and age.

The second charger was purchased for its test functionality and to see if there was a difference in the ability to fully and uniformly charge the batteries. The non-backlit LCD display isn't particularly easy to read. It's fiddly to select the different modes and currents but it has some useful functionality once you get used to its interface.

Both chargers use a pulsing method for charging where a high current and then no current is applied every few seconds - the average over that period is the desired current. This is a form of pulse-width modulation. The discharging feature uses the same technique to discharge. The discharge rates are fixed at half the selected charging rate, i.e. a test at 500mAh charge discharges at 250mA.

The second charger's manual says it uses the "-dV" method to determine full charge. It has a temperature cut-off for safety using two thermistors that sit very close to the (plastic) battery compartment in thermal paste - this is a common feature in chargers for Nickel-based batteries. It may apply a simple time limit as an ultimate precaution to avoid gross overcharging in the event of failing to detect a full charge. After charging it switches to a 25mA trickle charge. Some NiMH charge control methods are likely to fail to detect the battery is full at low (slow) charge rates, see the Voltage Drop (-ΔV) section in the Charging Methods step lower down for more detail on this.

The first use of second charger at the default 200mA charge rate on some new AAA batteries did not appear to terminate the charging at the expected time based on a duration equivalent to 1.5x the capacity and it left the batteries rather hot. This wasn't investigated further. All of the subsequent testing was performed with batteries charged at the 500mA rate which appeared to work well for the AAA batteries. This would be between 0.45C to 0.75C based on the actual measured capacities.

The second charger's case, display, buttons and functionality have a remarkable resemblance to many other chargers suggesting varying degrees of commonality.

The lygate web site includes a very detailed review of a Charger EverActive NC1000 plus which looks identical and has identical charge modes but has a default of 500mA. The (expert) user notes:

"At 0.2A the current is too low to do -dv/dt termination, but the charger stops anyway, maybe on a timer. The battery is not fully charged, but that is basically my own fault for selected way to low charge current for that battery."

With the exception of the IKEA Ladda, the 40mA test results did not match the brands' claimed capacity. The batteries in the first batch of tests were were charged on Charger 1 (fixed 350mA for AAA) for around 11 hours and then removed. The green light comes on indicating the charge is complete after 3-4 hours but they were left a little longer to see if the continued trickle charge topped them up. The time from coming off the charger to being tested wasn't noted and was quite variable partly due to the different brands being intentionally interleaved for testing.

The second batch of tests at 100mA were performed using Charger 2 at 500mA with the batteries left in the charger for 24 hours, meaning they received an additional ~21 hours of trickle charge. They were all removed from the charger and tested in sequence at 12 hour intervals to look at the immediate self-discharge rate. The second test explored discharge down to 0.9V as the Ansmann 1050 batteries have a different style of curve and this facilitated comparison with Charger 2's test function which discharges to 0.90V. The Energizer Nickel Metal Hydride (NiMH) Handbook and Application Manual confirms this is a reasonable choice of cut-off voltage and a safe one from a battery health point of view.

Normally discharge cutoff is based on voltage drops with a value of 0.9 volts per battery (75 percent of the 1.2 volt per battery nominal mid-point voltage) often being used.

The image at the top of the page is a lygte test of a charger that looks very similar to Charger 2 charging an AAA at 500mA. The voltage, shown as a band of red, is showing both the high and low voltage due to the pulsed charging method.

NiMH (and NiCd) batteries are one of the more complex battery chemistries to charge. The charge stored in a battery can vary according to how it is charged or mischarged. The charge will drop after removal from the charger due to self-discharge which will vary based on the battery chemistry and construction. Battery charging is not a 100% efficient process, a commonly quoted coulometric charging efficiency for NiMH batteries is 66% which equates to needing a 1.5x charge relative to the actual capacity to fully charge a battery.

The constant current charge rates for NiMH batteries may be described as:

• trickle-charge (0.03-0.05C),
• low (0.1C),
• quick (0.1-0.3C) and
• fast (0.5-1C).

The various charge control techniques are summarised briefly below based on descriptions in The Handbook of Batteries, a charger may use a combination of these methods.

Timed Charge

A fixed timer, has to assume battery is fully discharged, may vary by battery size as a proxy for capacity, .

Voltage Drop (-ΔV)

The voltage drops slightly. The book notes an issue with this technique.

... the peak with the metal hydride cell is not as prominent and may be absent in charge currents below the 0.3C rate, particularly at elevated temperatures. The voltage signal must be sensitive enough to terminate the charge when the voltage drops, but not so sensitive that it will terminate the charge prematurely due to noise or other normal voltage fluctuations. A 10mV per cell drop is generally used for the nickel-metal hydride battery.

This is echoed by Isidor Bachmann on Battery University:

It is difficult, if not impossible, to slow charge a NiMH battery. At a C rate of 0.1C to 0.3C, the voltage and temperature profiles do not exhibit defined characteristics to trigger full-charge detection, and the charger must depend on a timer. Harmful overcharge can occur when charging partially or fully charged batteries, even if the battery remains cold.

Voltage Plateau (0ΔV)

The voltage peaks with a slope of zero (a horizontal line).

Temperature Cutoff  (T)

The batteries reach a prescribed high temperature. This requires temperature sensor(s) which can monitor each battery possibly of varying size or the whole battery compartment.

Delta Temperature Cutoff (ΔT)

The batteries rise a certain amount above the ambient/initial temperature.

Rate of Temperature Increase (ΔT/Δt)

The temperature reaches a certain level.

Coulomb Counting

This is another technique not mentioned in the portable NiMH chapter. In general, batteries that are part of an advanced ("smart") battery pack or ones built into a device with an integrated charger may have a battery management system (BMS). This can continuously monitor the charge going in and out of the battery to aid determining the state of charge. This is called coulomb counting (some refer to this as gas gauging) and is used in devices like mobile phones and electric vehicles. This works best for smart batteries or ones that aren't usually removed.

Swapping the pair of 47 ohm resistors for 12 ohm increases the maximum current while staying within the power limit of the new resistors for this type of battery. The new resistors can be seen above. The colour coding indicates they are 1200 ohm 2% but a multimeter confirms they are 12.0-12.1 - the fourth multiplier band should be gold for (0.1x) but looks brown (10x).

A slightly more sophisticated tester could duplicate both the load resistors and transistor. These could then provide a high current and a low current path by using different resistors to give good precision over a wide range of currents.

This is a repeat of the tests at a constant current of 100mA to 0.9V at 23-26 Celsius with the previously tested rechargeable batteries plus some new additions.

• 2 Ansmann 1100mAh purchased from RS Components (UK), labelled made in China.
• 4 Energizer Recharge Power Plus (700mAh) purchased from Amazon UK, labelled made in Japan.
• 4 MaximalPower 1200mAh purchased from www.maximalpower.com in California, US, labelled made in China - they also sell through the marketplace on Amazon US.

For a 700mAh the rate is 0.14C, for a 1200mAh the rate is 0.08C. These tests were performed sequentially to look at the immediate self-discharge rate. A "24hour self-discharge rate" is sometimes cited for NiMH batteries reflecting their property of having an initially high drop in capacity after charging. The mean and maximum values are shown below for all the tests bar the pre-charged ones.

1. Energizer Recharge Power Plus 700mAh: mean 709mAh, max 728mAh.
2. IKEA Ladda 750mAh: mean 747mAh, max 763mAh.
3. Ansmann 1100mAh: mean 1011mAh, max 1036mAh.
4. Ansmann 1050mAh: mean 951mAh , max 980mAh.
5. EBL 1100mAh: mean 789mAh, max 823mAh.
6. HiQuick 1100mAh: 799mAh, max 818mAh.
7. MaximalPower 1200mAh: mean 779mAh, max 808mAh.

One extra test was run for a single MaximalPower battery to give it another chance at reaching its most atypical claimed capacity.

Observations:

• The results are consistent with the 40mA tests across brands.
• The Ansmann 1100 batteries again have huge variation in pre-charge capacity whereas the Energizer ones are near-identical. The Ansmann 1050 in 40mA tests had similar huge variation, IKEA Ladda ones were near-identical.
• EBL, HiQuick, and MaximalPower appear to be 800mAh batteries mis-labelled/mis-sold as 1100-1200mAh ones.
• For comparison, the lygte site features
• similar results for Ansmann 1100 and silver/green IKEA Ladda 750mAh batteries (modern IKEA Ladda ones sold in the UK tested here are 750mAh but grey-coloured);
• and Sofirn 1100mAh tests mirroring the EBL, HiQuick, and MaximalPower results here in terms of a battery with a high capacity claim that turns out to really be a ~800mAh battery.

The Uniross UCQ001 test function was used for charging and this does a charge/discharge/charge cycle where the second phase measures the discharge time to obtain a capacity. The mean values across the batteries from each test cycle are shown below:

1. Energizer Recharge Power Plus 700mAh: UCQ001 774mAh.
2. IKEA Ladda 750mAh: UCQ001 793mAh, 808mAh, 805mAh
3. Ansmann 1100mAh: UCQ001 1124mAh
4. Ansmann 1050mAh: UCQ001 1059mAh, 1065mAh, 1079mAh
5. HiQuick 1100mAh: UCQ001 878mAh, 907mAh, 907mAh, 882mAh
6. EBL 1100mAh: UCQ001 864mAh, 880mAh, 867mAh
7. MaximalPower 1200mAh: UCQ001 854mAh, 865mAh

The discharge rate was 250mA (based on half of the selected 500mA charge rate) to a cut-off voltage of 0.9V. For 700mAh the discharge rate is 0.36C, for 1100mAh it's 0.23C. This is a mean rate obtained by pulsing a larger discharge current and then passing no current with a period of a few seconds. This may mean the capacity measurement varies slightly from a true constant current test.

The table below shows the batteries in order of claimed capacity from exceeding it to substantially failing to meet it, any difference below 10% is highlighted in bold. The Uniross UCQ001 (charger 2) values are the maximum of the mean values.

The IEC 61951-2 standard test would be likely to yield slightly lower values with its rate of 0.2C and 1.0V cut-off voltage.

The original article, Battery Capacity Measurement Using Kitronik Inventor's Kit and Adafruit CLUE, has some ideas at the end for areas to explore.

Here's some further reading and battery reviews for batteries, generally about NiMH ones:

• Manufacturer Handbooks and Videos
• Panasonic Ni-MH Handbook (pdf)
• Energizer Nickel Metal Hydride (NiMH) Handbook and Application Manual (pdf)
• Tenergy Official: How to use your NiMH batteries so they last longer (YouTube)
• Adafruit
• Collin's Lab: History of the Battery (YouTube)
• Collin's Lab: Battery Basics (YouTube)
• Collin's Lab: Powerful Battery Usage with Ladyada (YouTube)
• MIT Electric Vehicles Team: A Guide to Understanding Battery Specifications (pdf)
• bigclivedotcom
• Charging NiMH cells - smart vs dumb chargers (YouTube)
• Bogus capacity NiMh cells on eBay. And a look inside one. (YouTube) - opening the case of a battery and touching the contents isn't wise, opening an NiMH battery is dangerous and best left to foolhardy YouTubers.
• Project Farm:
• Which AA Rechargeable Battery is Best after 1 Year? Let's find out! Eneloop, Duracell, Amazon, EBL (YouTube) - some commendable long-term and fast-talking testing - a capacity test after 1 year of continuous real world use to look at cycle lifetime, a test after 6 weeks and 1 year of storage to look at self-discharge rates and an internal (DC) resistance check using a BT-C3400 charger.
• Which "Lithium" AA Rechargeable Battery is Best? Let's find out! (YouTube) - tests on some rechargeable AA Lithium batteries plus a 2 year update on the long-term NiMH cycle testing at 11:35.
• Ben Krasnow Blog: Testing Tenergy 9V NiMH batteries (discharge curve) - a look at capacity and self-discharge for a Tenergy 7-cell PP3 battery which cannot meet its claimed capacity at 0.2C.
• Ed Nisley’s Blog: Eneloop AAA Cells: Six Years of Blinkiness - a look at capacity degradation (showing mWh) of Panasonic Eneloop AAA batteries over 6 years of real-world, typical use.
• EEVblog: EEVblog #35 2of2 - NiMH and NiCd Battery Charging Tutorial (YouTube)
• GreatScott! Are AA NiMH Batteries obsolete? Here is everything you need to know about them! (YouTube) - wide-ranging discussion of NiMH batteries including how they are charged and making a charger.
• Linus Tech Tips: Sanyo Eneloop AA Batteries & Charger Unboxing Linus Tech Tips (YouTube) - a young man opening a box and reading some words on the box.
• Isidor Buchmann's Battery University site:
• Summary Table of Nickel-based Batteries
• Nickel-based Batteries
• Charging Nickel-metal-hydride
• Cycling Performance - the gradual loss of capacity with use.
• Wirecutter at the New York Times
• Are Rechargeable Batteries Better Than Alkaline? Most of the Time. | When to Choose Rechargeable Batteries Over Disposable
• The Best Rechargeable AA and AAA Batteries
• The Best Rechargeable Battery Charger (for AA and AAA Batteries)
• TrustedReviews: Best rechargeable AA and AAA batteries 2023: More power for longer - the site has a slight feel of a clickbait site but it actually has a solid statement on independence and some real testing over many cycles by David Ludlow.
• Which: Best AA and AAA rechargeable batteries: Which? Best Buys and expert buying advice - the (paywalled) not-for-profit magazine of the (UK) Consumers' Association established in 1957.
• (US) Consumer Reports AA battery tests - (paywalled) testing, no recent tests, no rechargeable testing.
• lygte-info.dk: Battery test for AA and AAA (Alkaline, lithium, NiMH) batteries, individual tests - detailed and very thorough tests of many batteries.
• TylerTube: I Tested Rechargeable Batteries From Amazon! Which one is the best? (YouTube) - 7 AA NiMH tested from Amazon US in a tester and a real-world test with fans.
• Stefan Vorkoetter
• Review: Testing Sanyo’s Eneloop Low Self-Discharge Rechargeable Battery
• Review: Pre-Charged (Low Self-Discharge) Rechargeable Battery Comparison - the reviews from 2009-2010 are a little dated now but very thorough and interesting data on self-discharge rates.
• Gough Lui's 2016 AA Alkaline Battery testing: part 1 | part 2 | part 3 - extremely thorough and many brands tested.
• Underwater Arduino Data Loggers: Testing Cr2032 Coin Cell Batteries with μA to mA pulsed duty cycles - mentions finding counterfeit Sony (Murata) CR2032 batteries sold via Amazon US.
• Michael Bluejay's Rechargeable Batteries — compared and explained in detail - a good summary of technology but doesn't raise the critical issue of quality appraisal of Amazon reviews/reviewers.
• Fran Blanche: Can A Battery Flip Polarity? (YouTube) - battery voltage reversal which can be caused by different capacity batteries being discharged in series.
• Analog Devices: A Closer Look at State of Charge (SOC) and State of Health (SOH) Estimation Techniques for Batteries
• Analog Devices: An Accurate Battery Gas Gauge
• Linden's Handbook of Batteries by By Kirby W. Beard - a large, comprehensive and expensive book covering everything!
• Young, Kwo-hsiung; Yasuoka, Shigekazu (2016): Capacity Degradation Mechanisms in Nickel/Metal Hydride Batteries - an academic summary covering temporary and permanent loss of capacity.
• Nickel Metal Hydride Batteries 2017 - a free pdf book collecting together academic research papers.
• International Electrotechnical Commission standard IEC 61951-2:2017 Secondary cells and batteries containing alkaline or other non acid electrolytes - Secondary sealed cells and batteries for portable applications - Part 2: Nickel-metal hydride (pay to download) - the current industry standard for the batteries covered in this article.
• Energy Star: Test Methodology for Determining the Energy Performance of Battery Charging Systems (pdf) - includes some of the detail from IEC 61951-2.
• BBC Future Planet: The battery invented 120 years before its time - the story of the "battolyser", a nickel-iron battery chemistry which can produce hydrogen on overcharge without damage, discovered by Waldemar Jungner and refined by Thomas Edison.
• Chen Tech Electric: MCP Plus Series - battery testers used by industry and researchers.
• IET: The measure of Duracell battery elephant - the story behind the image above of Tony Diaz's sculpture.