Recommend a long-lasting battery

For my project, I want a battery that can power my Particle Electron for at least 2-3 years or so. Powering the Electron through an outlet or using a solar panel for various reasons isn’t an option. The Electron will be in deep sleep for most of the day, only coming online twice a day for 1 minute to gather some data and publish it.

There’s no size constraints for the battery; even a D-cell would be appropriate. Obviously the LiPo sold in the Particle store isn’t going to cut it (may squeeze 4-5 months out of it but that’s it). I looked into alkalines, LiFePo4 and LiSOCl2 batteries but I’m simply unsure which one is going to have the most reliable performance in the long run. Also wondering what voltage to aim for; would 3.6V batteries cut it?

Right now I already have a D-cell LiSOCl2 that can power the Electron but of course can’t handle the peaks of the cellular module; was wondering if it’s worth trying to make it work however.

it's kind-of sad that there is no sun nearby...

A battery would be able to handle the requirements of the small amount of current needed to keep the LiPo battery charged...

put a massive battery next to the Electron, if necessary step down the voltage to the input requirement for Vin (5V) and that battery can maintain the charge of the LiPo, particularly if you are only broadcasting once a day.

car batteries can be recharged and have upwards of 70 Amp Hours. You'd get some life out of that, I'd guess.

Well, a car battery seems a little bit overkill for a simple IOT project… I read about the LiFePo4 somewhere, and they seem to have a good lifespan/able to handle the peaks if im not mistaken? Also read about using a LiSOCl2 together with a capacitor.

We are talking about 2-3 years, so quite a bit of capacity is likely needed.
What you really should do is some experimentation to find out approximately what your Amp-hours are for a week or suitable interval and then use that to estimate your capacity for 2-3 years plus any extra capacity for just in case situations. After all, we do not know all the parts in your project that will consume power.

Do not forget that temperatures can affect battery life. For example, -40 deg Celcius is not friendly if you get winter conditions like that (I do).

A car battery may not quite be the best choice, as it was designed for lots of amps for a short duration to start an engine. A deep-cycle battery for an RV (recreational vehicle, sometimes called caravan in some parts of the world) is very similar but is designed for lower current draw over a longer period.

Cost can be a factor for a large capacity. Sometimes a clunky old-tech battery like lead-acid can be less expensive in these sizes.

whatever, but you don’t need to worry about the draw-down during broadcast events, because the little LiPo can handle that very easily…

The current out to keep the charge is very minimal…

A car battery is cheaper than LiFePo4 apples to apples, I’d guess. Motorcycle battery maybe? Upwards of 15 AH and a nicer footprint. I’d guess you can buy one for $30.00 or so…

Your going to have to do what cyclin_al suggested, as far as actually knowing what your board uses for power consumption. Otherwise your just taking a stab in the dark.

If you dont have a Wh meter to show you the total your board has consumed, then take something like a 2000mA Li-Po battery and time how long your board runs off that. That will give you some clue how many mA total you need.
Bug again like cyclin_al said, you have to look at the temperature where it will be used. Li-Po’s are not industrial temp rated so they will freeze when it gets to cold.

One other thing to consider when talking about years of battery life is self-discharge which will take a considerable share of any extrapolation of a weeks usage to that kind of time span.


I have had an Electron with a LiPo running some old firmware code (but consumption should be the same), and after about 3.5 months it has gotten to about 30% (for the record, it’s the default 2000mAh one from Particle store). The lowest temperature I want to use my battery in is about minus 25 degrees; highest is around plus 40 degrees.

@ScruffR yeah, and the discharge rate of the LiPo seems quite considerable at 1-2% a month…judging by my experiences so far with my 2000mAh battery, to extend the lifespan of my Electron to 2 years would already require at least a 12000mAh battery it seems.

I read the LiFePo4 can have a self-discharge rate of almost 5% a month, so that’s a no-go I guess. LiSOCl2 seems to self-discharge at less than 1% a year, but doesn’t handle the peaks of the Electron. I have the 3G Electron (800mA peaks), and D-cells according to vendors handle between 200-300mA (tested with my own battery; indeed the Electron resets when it tries to connect). Is it worth trying a capacitor with this?

1 Like

The best Primary Batteries that I’ve used is Energizer L91. They have a flat discharge curve, an extremely low self-discharge, can handle the 2 amp Modem requirements, and are fine in your -25 degree temperature enviroment.

Each will provide 3,500 mAH in this application.
You will see 1.6v / battery “in circuit” at 100%, down to 1.3v / battery reaching discharge.
That’s a Voltage range of 4.8v to 3.9v using (3) AA L91 batteries in series.

You may be able to search and find reports for the Electron’s Sleep Current on this forum.
If not, you can use something like

Then you parallel as many sets of (3) that you require for your mAH rating (3500 mAH each set).
These cells are $1.50 each.

As @ScruffR said, the self discharge profile for the battery will be a major factor. I haven’t found a better primary battery without getting too exotic and expensive.


thank you guys for recommendations, been searching for these type of answers, thanks a lot for answering with facts and not just ‘because it’s the best’ as i usually find. i want to learn more about these while i’m sitting home under doctor prescription of taking saizen because of feeling bad in the meanwhile. so would you mind if i would ask some other questions later? thanks!

Thanks for the reply, it’s very interesting! I’m going to look further into this now.

For the record, I don’t have appropriate meters for microamps, but I did do a rough calculation ousing state of charge values, and found about 180uA for deep sleep current.

Looking at the specs of the Electron and the battery, it should be possible to put 6 of them in series for a supply of 9.6V (down to 7.8V)? This wouldn’t exceed the Electron’s 12V limit, and give it a 21000mAh capacity to draw from. My 2000mAh LiPo has lasted for about 4 months or so (but with a much higher discharge rate), so this setup should last around 4 years?

No, Series doesn’t add to the mAH capacity. And I don’t see any point supplying a higher voltage, you just loose efficiency.

The Operating Range for (3) L91’s in series is 4.8v to 3.9v with a 3,500 mAH rating.
If you parallel 2 sets of 3 batteries, you get the same voltage, but 7,000+ mAH rating.

Thanks. 3500mAh doesn’t seem like much though.Would that really juice it up for years?

That depends on your Sleep current, how long the Electron is awake taking samples, how long the Electron spends transmitting/receiving, and the self discharge rate of the battery.

I recommended the L91 because it has a very low self discharge rate and satisfies all the requirements you’ve mentioned.
A 3S2P may be a decent starting point, but you really need to take measurements if you want any degree of certainty.

I’m not saying it’s the best choice for your application, but it’s where I’d start…

My Deep sleep current seemed to be around 200uA or so, close to what Particle lists as the lowest current for the Electron. It’s powered for 1 minute a day, mostly gathering data and transmitting it. All in all not too high. The reason I’m asking is that I drained a 2000mAh LiPo battery for 80% within 4 months, so unless Im missing something (I probably am) a 3500mAh capacity would seem a bit lacking still to achieve 2+ years?

I would suggest using SAFT/Tadiran batteries, you will find these are used in almost all industrial battery powered applications; Energy Meters, Data Loggers etc…

You still need to use the LiPo for the additional current required when the modem is in use, you could look at swapping the LiPo out for a HLC instead too.

Without doing the calculations but from what you’ve said, with this combo I’d expect 5+ years.

Tadiran are LiSOCl2, correct? That was my initial plan but indeed the modem current is a barrier. So you’re saying I could simply combine a Tadiran and LiPo battery together? I remember reading somewhere its not recommended to use two batteries with different compositions so I just want to make sure.

In this case, does it matter if they’re in series or parallel?

I’m suggesting you use the LiPo plugged in as normal and connect the tadiran to VIN, I don’t have a data sheet at hand but, I believe the tadiran should then trickle charge the LiPo and the Electron is powered by the LiPo first before VIN… If someone could have a read and double check that would be great :+1:t2:

Is there a specific reason you recommend Tadiran? Or would any LiSOCl2 battery do? (e.g. I have one D-cell Xeno Energy)