I am currently designing a simple data logger that will be deployed in the field for (hopefully) months at a time. It will be an ATMega 328 with a real time clock and an SD Card for data storage. Obviously batteries are critical to this application and owing to the various voltages etc. from different battery chemistry I can't finalise the design until I have chosen a battery technology. I have done A LOT of Googling and reading around the subject but I thought I would ask to check that I had got all the bases covered.

So what do I need to consider when choosing rechargeable batteries that will run with very low current (I'm expecting quiescent current well under 100uA and peaks of not more than 150mA) for a long time. Obviously self-discharge is critical but is there anything else? Are some battery chemistries better suited to low discharge rates than others? Are some more rugged (wrt damp, temperature, etc.)? Do some cope better with deep discharge?

At the moment I can see pros and cons for low self-discharge NiMH, lead acid and lithium and I'm not sure which was to jump. The advantage of lsd NiMH is that I can nicely get 4.8V and therefore can run the ATMega without a regulator. I could do this with 3.7V lithium but the SD card needs 3.3V (regulated) and so there is very little overhead to play with in that approach. The advantage of lead acid is the cost and capacity – I can get 10AH for around £16, so I could afford to lose some to self-discharge and a regulator…

Thoughts and suggestions please!

 

I have built data loggers that have to run for a year or two in the field under severe weather conditions.

I would recommend that you look at not only the battery technology but also the entire power requirements. With today's ultra low power technology, 100μA standby is 100 times too much. Today's mcu will power down to less than 1μA.

Secondly, look at the programming and take advantage of low power modes as much as possible. Why run the clock at say 16MHz when 32kHz will work just as well? Keep the unit in power down mode as long as possible. For example, would 5-minute intervals between samples be sufficient?

For batteries, I have used 6V alkaline lantern batteries such as this one. The screw terminals ensure a solid reliable connection.

Also, the data loggers I designed and built had to be left in the field (1km down below the surface in caves). The researchers went down into the caves and swapped out compact flash memory cards in order to retrieve the data.

 

Thanks, I am aware of the low power techniques for microcontrollers, I was erring on the side of caution with that figure, the mcu may do <1uA but if a regualtor is need it could go up (please lets not get into a low quiescent current, low drop out voltage regulator debate). Besides a 4000mAH low self discharge NiMH could be loosing 500uA per day in self discharge and lead acid 10x that, so 1, 10 or 50uA quiecent current for the device is pretty much noise...

So I am really interested in what I need to think about regarding choosing a rechargable battery for this project.

 

Here is a nice link.

http://batteryuniversity.com/learn/article/whats_the_best_battery

Self discharge may dictate your choice.

 

I would go Li non-rechargeable for the best power/weight density. There are tons of sizes to choose from.

I could do this with 3.7V lithium but the SD card needs 3.3V (regulated) and so there is very little overhead to play with in that approach.

Not necessarily. At very low currents such as yours, you can get LDO regulators that will regulate with a couple hundred mV across the regulator.

 

Thanks guys, that battery university link is very helpful.