What are good AA/AAA batteries for our IoT devices

Hardware
1

Hey!

This is a topic about discussing different AA and AAA type batteries.

So far I have been using mostly basic rechargeable batteries (as I was annoyed by buying tons of non-rechargeable ones).

Li-Ion

I wanted to share a small finding I am happy about right now and potentially hear if I am missing anything. I have a couple of AA/AAA battery-powered devices, mostly ZigBee TRVS, NUKI lock and a few IKEA light controllers.

Many of these devices recommend using non-rechargeable batteries. AFAIK the only reason is that they provide higher voltage (1.5V), while most of the rechargeable batteries (NiMH) provide only 1.2-1.3, and it’s not as stable as for non-rechargeable. It’s not the “rechargeability” per-se that is the problem, but the low voltage caused by using a specific design: NiMH.

Well, enter Li-Ion AA/AAA batteries (that are somewhat new, actually)! They provide stable 1.5V (so e.g. TRVs that need more voltage work just fine), are rechargeable, are lighter, have bigger capacity, have higher cycle count (at least compared to my normal mid-range NiMh rechargeable batteries), charge faster, and provide a longer shelf-life.

I bought these: https://www.amazon.de/-/en/Battery-Rechargeable-Lithium-Batteries-Charger/dp/B08PCL5B7Y (US Amazon), and I am satisfied (although some people report bad batches and stuff). There don’t seem to be that many good alternatives, the only one I found was Hixon (but I suspect it’s secretly Kratax as well).

2 Likes
2

Interesting! Thank’s for sharing your findings. :+1:

I kind of doing quite the opposite and mostly migrating away from Li-Ion (and Li-Po) in favor of the (good old) NiMh-LSD for my low power scenarios. :battery:

I don’t have/had any AA or AAA sized Li-Ion batteries but as written in the description from your amazon link they “suffer” from the same disadvantages as all other sized/packed Li-Ion (and Li-Po) based batteries. Additional I expect the efficiency even be lower as “standard” sizes which output the nominal ~3.7V.

Specially the quite high self discharge of Li-* batteries (compared to NiMH-LSD) and the relatively quick degradation of capacity (again compared with NiMH-LSD) is something that bothered me a lot in practical usage scenarios. Another thing I also exclusively experienced with Li-* type batteries is that they sometimes (long before the prolonged EOL) just do some “spontaneous” chemical reaction and for example accumulate gas inside and get fat or just can’t hold the output voltage at the designed level. :chart_with_downwards_trend:

Which could be counted as the same - Li-Ion (and Li-Po) have a higher energy density (by mass & volume) compared to NiMH-LSD.

But while Li-* have a higher density this doesn’t look to be true for the AA/AAA types you linked (and probably all other with this form factor). Probably the extra circuit to get the voltage down from the cell voltage (nominal ~3.7V) to the desired 1.5V is to blame. This circuit does not only lower the overall efficiency but takes valuable space for the “juicy” layers :grapes:

One buyer claims for the Li-* batteries you bought that the capacity for the AA type is only 2300 mAh and for the AAA 700 mAh. This is nothing astonishing and you find NiMH-LSD batteries out there that have a higher capacity than that. :bar_chart:

Not sure if the “old” type (without Low Self Discharge) are still (widely) available nowadays? Even our supermarkets sell rechargeable batteries as “ready to use” - which is a great indicator that they are actually NiMH-LSD

NiMH-LSD can reach about the double cycle-count as Li-Ion and Li-Po and from my experience they really last much longer. It’s even possible to easily double the cycle count of NiMH-LSD batteries by not fully dis- and charging them. I own more than 10 year old NiMH-LSD batteries and they still work perfectly while at the same time having still more than 80% of the original capacity (for the few I tested). :chart_with_upwards_trend:

Fast charging is typically a bad idea - for Li-* and for NiMH type of batteries as they will degrade significantly faster. I even switched to “extra” slow/trickle charging for my NiMH-LSD batteries which should put around :zero: stress on them when charging. I have no problem having the batteries in the charger for one day when they will be back in the device for another year again after that. :man_shrugging:

I can’t even remember when I discarded my last NiMH-LSD batteries (did I ever? :thinking:) - but just a few weeks ago I had real troubles getting rid of another three broken Li-* cells (two of them were really bad smelling) :do_not_litter:

NiMH-LSD batteries have a low self discharge (it’s in the name :wink: ) - Some claim down to only 3% per year :warning: This is very far away from any Li-* type can handle - on the other hand I even experienced that (very) small Li-* can even “die” after only 1 year on the shelf :put_litter_in_its_place: Your amazon listing even says for your batteries to do a full dis+charge cycle every 4-6 months - I expect this type of Li-Ion battery really can be “killed” by just not using it for a longer period :frowning:

Do you mind sharing how long do you use this Li-Ion AA/AAA type batteries already? Like did you have them in a IR remote for more than a year already for example or in another (ultra) low power devices? :spiral_calendar:

I expect them to really last significantly shorter for low power scenarios then NiMH-LSD rechargeable batteries or even non-rechargeable leakage prone alkaline types :battery:

1 Like
3

Hey! Thanks for the extensive feedback and review, this is all useful and I am likely updating towards not using them that extensively / buying fewer of them at once. You make a lot of good points and I agree.

I have used them for ~1 year and have had better results than with the traditional NiMH I bought, mostly in TRVs. Sure, more data and longer time scales would be better, but they haven’t been available for that long.

I will try to get to a proper answer sooner, but one of the IMHO pieces I miss in your answer is the voltage. NiMH-* still have 1.2V, am I right? Isn’t that an issue for you in the type of devices I mentioned (TRVs, lock, …)? Specifically, I haven’t used them in ultra-low-power devices (e.g. IR remote). In TRVs or locks, regular NiMH batteries (the highest capacity I can find on Czech market) don’t last more than a couple of months and once the voltage drops, they just cannot deliver the force needed to do the thing. That’s where I use them. Maybe the path is then using Li-Ion in these devices, but NiMH-LSD in the low-power ones?

To support my claim about 1.5V being important, I am going to throw these random claims from FAQs of TRV companies:

Rechargeable batteries should not be used, as the power delivered by these batteries is not sufficient to operate the motor properly.

Most standard rechargeable batteries have a lower starting voltage than conventional disposable batteries, which means the Wiser Radiator Thermostat will not get the voltage required to drive the motor as designed, leading to a system potentially behaving at sub-optimal levels.

(you can google “why trv need non-recharable battery”)

I have also just experienced that e.g. with this radiator valve, using NiMH will just not work: the valve reports “low voltage” after a week or two.

Interestingly enough, though, NUKI developers actually recommend using Eneloop batteries (NiMH-LSD?).

My current take on what to use for what device (conditional on using rechargeable AA/AAA)

It depends on what the batteries are used for.

  1. Any device that needs to draw 1.5V (usually devices that need to develop high mechanical force), use Li-Ion batteries.
  2. In any other case (1.2V is fine), use Ni-MH-LSD.
4

I changed the topic name to be more generic about various types of batteries.

5

That is about correct. When topped up to the max my NiMH-LSD reach up to 1.4V but normally quickly decline to 1.2V for most of the charge level to just drop around to roughly 1V when almost empty. Classical (non rechargeable) alkaline batteries on the other hand start at about 1.5V but decline more steadily (not “holding” the voltage) down to also about 1V when about empty. :hourglass_flowing_sand:

The AA/AAA Li-Ion type on the other hand you own should output a steady 1.5V from the beginning to then end because internally they have a higher voltage and use a dedicated curcuit in each battery to always have a 1.5V output. :hourglass:

This graph shows the difference quite well (The pink line is the alkaline battery the blue and black ones are NiMH-LSD types with different capacity) :point_down:

image
image800×564 94.2 KB

Source: http://s3.media.squarespace.com/production/833805/9936504/wp-content/uploads/2010/03/eneloop-discharge-curve.gif

I do not own battery operated TRVs or locks. On the other hand all AA/AAA type battery operated devices I call my own can be operated with the NiMH-LSD types without any problems. :battery:

It just depends on the design/circuit of the device. I would expect that a manufacture can save a tiny bit on the production costs when limiting the voltage range of the device to 1.5V only - but that would also mean that the device can’t make use of the full battery charge of ordinary alkaline batteries - for me that would be a failed design. :thinking:

Are you certain they are the “LSD” (often also labeled “ready to use”) of NiMH and not the “old” (obsolete) ones which will self discharge in a couple of months already by themself? :page_facing_up:

The NiMH-LSD types of batteries can be discharged quite fast (also at lower capacity) - I think they easily allow a 3C discharge rate/current or more - that should be on par (or maybe even better) than Li-Ion batteries. :arrows_counterclockwise:

The “force” is not determent by the voltage (only) but would be counted as power which will be current multiplied with voltage. As said there should be no technical reason for a NiMH-LSD type of battery to deliver less power at a lower capacity (I on the other hand would maybe expect this from the non-rechargeable alkaline types of batteries to a degree). :arrow_down_small:

For the Li-Ion AA/AAA types you have the circuitry inside which also “drops” the voltage to 1.5V ant that will determent the maximum discharge current and therefor limiting the maximum possible power output. :arrow_backward:

I didn’t even do that but just openend the first of your links and looked at the comments. Like mention there other manufactures “addressed” this issue already and Tado for example is suggesting (eneloop) NiMH-LSD batteries for their TRVs :point_down:

Funny enough this company even advises against every type of rechargeable batteries but suggests disposable Li-Ion types :put_litter_in_its_place:

  • The Valves are designed to operate with 1.5V AA alkaline batteries. However, better performance can be obtained with 1.5V lithium batteries (non-rechargeable).

:man_facepalming: Simply a sh!tty design/manufacture to avoid I guess…

Which somewhat supports the claims I made so far I guess? :thinking:

That the device reports low voltage is due to the discharge curve it expects by (it’s bad) design which is for alkaline batteries. :bulb:

On the other hand if you use your AA/AAA Li-Ion type it will show 100% battery charge and suddenly turn empty without advising low voltage before (for the same reason as just mentioned). :point_up:

Again, I would just blame a bad manufacture design. If a company is throwing a device made for AA/AAA batteries and only takes non rechargeable alkalines into account for it that is nothing more than a dead end. Ether “accept” the form factor and make it work stable down to 1V or else just design it directly for Li-* without leveraging the AA/AAA form factor in the beginning. :battery:

One could argue high mechanical force needs a high power in the beginning. In theory that is something all NiMH-LSD should be able to provide. For example:

  • NiMH-LSD type with 1.2V and a current of 833mA equals 1W of power :muscle:
  • Li-Ion type with 1.5V and a current of 667mA equals 1W of power :muscle:

Like mentioned it just looks like there are manufactures out there which just want to save a penny or simply have a bad hardware design :man_shrugging:

6

You mention buying the highest capacity batteries you have found. They might be great for flashlight where you want to use them for few weeks*. However, if you want to use them for something like one year without charging, highest-capacity batteries are likely not the best choice. I have both an anecdotical evidence (cheap IKEA 750mAh batteries outperforming 950mAh GP) and theoretical explanation.

Anecdotical evidence: I bought few Styrbars. Most of them got new IKEA Ladda 750mAh AAA batteries. I didn’t charge them before use, so they have started at about 85 %. One Styrbar got 950mAh GP ReCyko+ 1000 AAA batteries, 100 % charged. I expected the GP (fully charged and higher capacity) to last longer than the Ladda (partially charged and lower capacity). At the moment, all Laddas show at least 75 % (with short temporary drops to 70 %), while the GP shows just 30 %. Note that the comparison is not scientific, the GPs weren’t new (but I don’t expect them to be that worn out after just few cycles) and the usage frequency also differs (but this probably rather favors the GPs over Laddas).

Theoretical explanation: Low self-discharge NiMH batteries need a stronger separator, which takes more space. For this reason, they can’t achieve as high capacity as high self-discharge batteries. So, when you buy the highest-capacity batteries, they will likely have rather high self-discharge.

Also, not all LSD batteries are equal. Eneloops have a good reputation for being ultra low self-discharge. However, there are some other brands (including IKEA) that seem to be just rebranded Eneloops, sometimes for much lower price: https://eneloop101.com/batteries/rewrapped-batteries/

And even not all Eneloops are the same. Some are from Japan, some are from China, and it seems to matter.. Also, they vary by capacity, and higher capacities have probably higher (though still reasonably low) self-discharge.

In Czech Republic, I’ve bought IKEA Laddas from Japan, both 750mAh AAA and 1900mAh AA. They are quite cheap (129CZK / ~$5 for 4 batteries). I haven’t tried the 2450mAh (more expensive and probably higher self-discharge), but they are reportedly exclusively from Japan. However, on some other markets, you might find batteries from China. Also, you can ignore sentences mentioning Sweden:

PXL_20221123_101018760
PXL_20221123_1010187601920×1440 151 KB

Using NiMH batteries when manufacturer recommends non-rechargable batteries can affect the estimate of battery percentage. This is fine for many devices, but not for those where accurate battery percentage is critical, e.g., smoke detectors. I have shared some more thoughts on battery status recalculation for NiMHs: NiMH battery percentage recalculation

*) And even with this use case, it might be a good idea to check (and potentially recharge) them even twice a year, even if you don’t use them, in order to prevent damage caused by deep discharge.

EDIT: Fixed the capacities of the GP batteries + a typo

1 Like
7

Wow, excellent resources and thoughts, thank you a lot for sharing!

This is terribly rabbit-holing btw :sweat_smile: .

8

I’ve noticed that Tado recommends Eneloop AA 1900 mAh batteries: What type of batteries should I use in my tado° devices? How do I replace them? | Help Center

I assume they are aware of Eneloop Pro with higher capacity (but probably higher self-discharge). I wasn’t sure where exactly the tradeoff between high capacity and low self-discharge is (I need it only in part of the year; charging once per year is OK), but Tado slightly suggest that my choice of the 1900mAh AAAs was a good choice even for TRVs.

1 Like
9

I’ve decided to test the capacity of the ReCyko batteries (designed capacity 950mAh). One of them has measured capacity 780mAh, the other has 880mAh. So, while they have some disadvantage in the comparison (they weren’t new), they still seem to have significantly higher self-discharge than the IKEA ones.

10

Looks like this ones here:

As I didn’t not found any “advertised” self-discharge for the IKEA ones I will just post the marketing claims from the ReCyko (950mAh) ones:

Pre-charged and ready to use, it can stay up to 50% charged after 12 months

And the marketing text from eneloop:

Retains 70% capacity after ten years in storage

This is a significant difference (in marketing?) and by the looks of it LSD ≠ LSD (low self discharge) :battery:

11

My ReCyko batteries have product number 1032114080. Your link doesn’t show such number, but it appears to be the same.

While IKEA doesn’t seem to promise anything specific about the quality of their LSD, there are some reasons they are the same as Eneloop (except label and price) or quite similar: https://eneloop101.com/batteries/rewrapped-batteries/

1 Like