Battery Life Estimates

[tremmert81]

Hello,
Are there any battery life calculations available? We are preparing for a deployment and need to achieve 1 year of data. We are trying to determine the minimum battery capacity that will be required to take a sample every 1 minute.

Are there any estimates available or additional information that may help?

[amanda-matthes]

In most scenarios, your limiting factor will be storage not battery capacity (provided your battery is fully charged). SnapperGPS 1.0 has a 512 Mbit flash which can store just over 10900 snapshots. That is enough to take a snapshot every minute for a week, every 5 minutes for a month or every hour for a year.

If you are still interested in power consumption, you can check out the preprint of our hardware paper for standby currents and charge consumption per snapshot.

One thing to potentially watch out for is that the standard firmware flashes the LEDs when a board is set to record at a later date. This is usually negligible if you only pre-program the device a few minutes in advance, but can quickly add up if you do it too early. If this is really needed for your application, consider disabling the LED flashing in the firmware.

[JonasBchrt]

I agree with @amanda-matthes that the storage will be your limiting factor.
If you are still interested, there is information on the expected energy consumption during a deployment in our hardware paper in the Section Power Consumption:

The current should be 1 µA–2 µA if the board is sleeping [...]. The charge consumption for a single snapshot should be <0.3 µAh.

This is all at 3.3 V supply voltage (cf. Section Electronics). The 0.3 µAh is rather conservative. So, if you want an upper bound on the charge consumption, you can do

charge_consumption [µAh] < 2 µA * deployment_duration [h] + 0.3 µAh * number_of_attempted_fixes.

For a year-long deployment with hourly-fixes, this is around 20 mAh. (Again, this is a conservative estimate.)

[JonasBchrt]

But note that you need to account for the battery self-discharge on top of this. A rule of thumb is that LiPos self-discharge with roughly 5% per month. However, this significantly depends on the quality of the battery, its protection circuit, temperature, humidity, and more. I would recommend that you (A) check the datasheet of the battery that you bought / are intending to buy or any additional information provided by the supplier and/or (B) account for a considerable safety margin.

[gauteh]

Hi, you write 25 mA in your preprint. What does that mean?

[JonasBchrt]

Hi @gauteh ,

I understand that you're referring to the sentence "the maximum current should be around 25 mA when capturing a snapshot" in the paper. It means that the electrical current flowing from the battery into the SnapperGPS board isn't higher than 25 milli-Ampere = 0.025 Ampere. Usually, it's much lower. The current only gets close to its maximum during the very brief time when a GNSS signal snapshot is captured.
For your choice of battery, the 25 mA mean that your battery must be able to supply peak currents of 25 mA. But that's not a tough requirement. Really any LiPo battery with a few milli-Ampere-hours capacity should be able to deliver that.

Hope that helps.

[gauteh]

Thanks! That clarifies it. We're developing some low cost wave buoys (https://github.com/gauteh/sfy). But gps power consumption is limiting, we could process it offline. How much data is necessary to send for frequent sampling (10 Hz)? We have decent bandwidth over cellular.

[JonasBchrt]

The default firmware captures 6 kilobyte raw GNSS signal snapshots. 1 snapshot = 1 fix attempt. So, with 10 Hz sampling frequency, you're looking at whooping 60 KB/s without overhead. (Note that the raw data can't really be compressed.) Maybe Chapter 6 - A Low-Cost Snapshot GNSS Receiver with Cloud-Offloading via a Narrow-Band Cellular Connection - of my thesis is interesting for you. I built a SnapperGPS prototype with a Blues Notecard attached to it there and do a comparison with the on-board Blues GPS as well as some theoretical solution. I wasn't really convinced by either, the SnapperGPS solution and the Blues solution... (Although, Blues uses different GPS chips for different Notecards AFAIK.)