nRF9160 battery life

[eyss5]

Hi, I'm looking to develop a GPS tracker over LTE-M/NB-IOT using nRF9160. The user needs to be able to reach the device to check its location on demand (20s delay is OK). Product size should be small, likely 100mAh LiPo battery. I'm hoping for battery life of months - 1 year.

Do I understand correctly that since the device needs to be reachable at all times, PSM is out of the question? 

I tried using the online power profiler with DRX modes (20s iDRX interval). The other variables don't seem to make a big difference. Comes out to 0.15mA. On a 110mAh battery, that gives me about 30 days. Would you expect a on-demand GPS tracker on a small battery to last 30 days? Or is something misconfigured/misunderstood?

[Stian Røed Hafskjold]

Hi, in PSM mode the UE is not reachable by the network, so if you need a 20s maximum delay, you have to use eDRX mode. Based on the power profiler settings I can see that you configured a wakeup interval of 2.5 hours, and it's correct that with such a long wakeup interval the current will be dominated by the eDRX current, and the other network parameters will not have much impact. The current will therefore mostly be decided by the minimum accepted latency, which will be the eDRX interval, and in this case 20 seconds. 0.15 mA in average sounds reasonable with a 20 sec eDRX interval.

When it comes to the choice of battery, the LTE radio will most likely not run on such a small battery (100mAh). The TX peak current in poor signal conditions can reach up to 400 mA. Batteries that can handle such a high peak current are usually in the 2-3Ah range and above. 

https://infocenter.nordicsemi.com/topic/ps_nrf9160/_tmp/alta.nRF9160/autodita/CURRENT/parameters.id_current_modem.html?cp=2_0_0_4_1_0_14

Please note that the NBIoT standard defines a maximum repeat mode (CE = coverage extension) that can last up to about 12.0s with pattern: [256ms TX ON] – [40ms transmission gap, RX ON in practice]  – [256ms TX ON] - … and so on. In LTE-M mode the max total CE time is shorter ~3s, but still with the same TX pattern.

Also, generally the battery AC resistance is recommended to be no more than 200-300mOhm. Less is better when if you want to get the most out of the battery.

[eyss5]

Thanks for the reply Stian Røed Hafskjold. I wonder how companies like this one, also using an nRF9160 + small LiPo battery, claim 1 year battery life. The math seems off by quite a bit. They claim the alert button sends the location immediately. I assume they also cannot be using PSM? Is there something that I'm missing regarding increasing battery life? 

[Stian Røed Hafskjold]

The key is a bigger battery.

The 150uA average current previously calculated includes SIM card current, if you get a low power eSIM you can save maybe 20uA. It also included 32 PDCCH repetitions each wakeup which is network dependent, but fairly conservative. If you set PDCCH reps to 1 and SIM clock stop current to 0, you get about 70uA in average. Many network parameters are set by the network, and therefore battery life time is network dependent.

Using a 800mAh battery and having a 70uA current in average gives you 476 days of battery lifetime (very simplified of course), which is in the 1 year ballpark.

[eyss5]

Hi, I'm revisiting battery life and a bigger battery sounds like a good idea. One thing i wanted to clarify. The UE needs to wake up periodically (every 20s) to check if it has a message to send data. I'm assuming that yes/no data (1 byte) doesn't count as "data transfer" in the online power profiler, shown below. If so, the power becomes unfeasible (hundreds of uA). 

If the answer to the above question is no, how does the UE get the data? I was watching a webinar and it says the network can store data for device. Would this be the right way to do it?

[Stian Røed Hafskjold]

The UE uses the paging occasion (eDRX interval) to listen for incoming data. It doesn't actually receive anything unless the network has data to transmit. The network will store the data and wait to transmit the data until the scheduled UE paging occasion happens.

So while the UE stays in RRC idle mode, it wakes up periodically (eDRX interval), turns on the radio receiver, listens to the PDCCH channel, and if no information is received (no paging), it assumes that the network has nothing to transmit and will go back to sleep. If the UE receives a paging event and there is data to be received, it will go from RRC idle to RRC connected mode, receive the data, respond to the message (send data), and go back to RRC idle again.

If you want to simulate this in the online power profiler, turn off PSM, set the required iDRX interval (this will be the latency) and then adjust the data transfer interval to how often the tracker is being used, e.g. a day, or week. You will see that the added average current by waking up is starting to get negligible when you get to a wakeup interval of a day or more.

[eyss5]

Perfect, thanks!