Change in data frequency and battery consumption

[Elliot Richardson]

Hi Team,

We have a case open - 285055 - where we raise issues about some data being rejected from your nrfcloud platform. As a secondary investigation we are looking at the changes in behaviour of our sensors after 19/02/22. I've attached our data for you to view below.

I've conducted a preliminary investigation around the data that we were collecting from our deployed devices. Looking at this it is clear that for the majority of sensors something happened around 19/02 and their battery consumption drastically increased. Calculating the gradients as an estimate their power consumption doubled but you can see that some of them stopped returning data all together.

Given these were all out in the field and had no firmware changes applied from 18/02 to 19/02 can you comment on what happens to these devices if the local mobile network suffers issues? I know that the V2.0 asset tracker firmware was deployed around then but storm Eunice battered the area, likely damaging a mast in the process.

As I said, can you comment on what might have caused this drastic change in behaviour and battery consumption?

Best regards,
Elliot

[Elfving]

Hello Elliot,

I don't think the storm necessarily had a lot to do with it, and it doesn't look too drastic to me. Version 2 of the asset tracker being deployed at the 19th could explain the change in data rate at that point, though it seems that sensor 22 didn't get updated along with the rest. After the 19th all the batteries are discharging rather linearly (as they should) until they reach about 3500mV, at which point they drop down quickly - which is normal in batteries, as the internal regulator tries to maintain an operating voltage. All the batteries essentially died. 

I had a talk with Claire about your case earlier, and I hear that you are also working on decreasing the power consumption of your devices?  I understood that getting a "live feed" isn't your main goal, so decreasing the data rate and taking advantage of the sleep modes is an idea. Could you expand on this a bit and what you have tried? Have you used the PPK2 to find out how much you are currently consuming? Have you had a look at this? 

Regards,

Elfving

[Elfving]

Hello again, 

I might've been a bit quick on Friday - I see now the change in consumption on the 19th, which I guess was what you were referring to. It seems that the gradient is about twice as steep. 

But the point remains that we should take a look at what is drawing power in the new FW and try to optimize it.

Regards,

Elfving