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
The only reason I have suggested using tadiran/saft is because; I know British manufactured battery powered RTU’s use these with a HLC combo. And they advertise 5+ years…
I have no experience or knowledge of any other
Edit/ They use Sierra Wireless modems instead of UBlox but I should imagine they have similar power consumption, without knowing the exact part number and seeing the data sheet etc
If that is the case, then you could look at having a HLC in parallel with a Tadiran connected to VIN and don’t use the LiPO. The HLC should be able to cope with the current Spike from the modem and then get topped up again by the Tadiran… I’m sure there is more to it, than just that though.
I’m not sure I would trust a Chinese alternative battery, after so many reports of them in the news catching fire/exploding etc especially if you are using them in an enclosed area… they are cheaper for a reason.
This is worth a read to understand the difference between bobbin and spiral wound batteries…
Thinking about this more, I would buy a SAFT LSH 20 and connect it to the Li+ PIN
LI+
This pin is internally tied to the positive terminal of the LiPo battery connector. It is intentionally left unpopulated. Please note that an incorrect usage of this pin can render the Electron unusable.
Li+ pin serves two purposes. You can use this pin to connect a LiPo battery directly without having to use a JST connector or it can be used to connect an external DC power source (and this is where one needs to take extra precautions). When powering it from an external regulated DC source, the recommended input voltage range on this pin is between 3.6V to 4.4VDC. Make sure that the supply can handle currents of at least 2 Amps.
This is the most efficient way of powering the Electron since the PMIC by-passes the regulator and supplies power to the Electron via an internal FET leading to lower quiescent current.
The reason behind this is because, using VIN the Voltage range is 3.9 - 12VDC and using the LiPO Connector the voltage is 3.7VDC... where the SAFT is 3.6VDC.
Thanks for the help, yeah. It does seem like LiSOCl2 batteries with a support for higher peak currents (such as the LSH 20) are my best bet. They do generally seem quite a bit more expensive, but at least the solution is straightforward with a single battery juicing up the Electron through two wires.
I had a chat with someone from Tadiran a ways back and he gave me this PDF which explains how a primary (non-rechargeable) + secondary (rechargeable) option is the way to go when it comes to wireless sensors.
He provided me with this PDF sheet that sort of explains how they work. I asked him if he had a schematic that could help my application to switch between the primary/secondary batteries when the current demand from the load (the Electron UBLOX in this case) increases, but the conversation ended there. Pretty sure it’s because he didn’t see value in following through because of my low-volume.
Would be great if someone else had some thoughts on this as well. I have a low-peak LiSOCl2 battery ready to be used, just not sure how I should set it up and how to monitor it to check if indeed the LiPo is being recharged by the LiSOCl2…
You got me curious again @Vitesze. I did some additional research on a chip solution to this primary/secondary, and maybe…just maybe…you could utilize the LTC4415 from Linear.
My thought is, you could use the EN1 and EN2 (enable) pins on the IC, controlled by digital outputs on the Electron, to switch to the Primary (non-rechargeable) battery when the Ublox module is transmitting, and then switch to the Secondary (rechargeable) battery for the rest of the time when the Electron is finishing up code and sleeping / deep-sleeping.
It’s not a perfect solution because it’s not actually detecting the current from the load and switching to the primary battery, but maybe it can be used in conjunction with that.
The output of the LTC4415 would go to Li+.
Unknown: I haven’t read the whole datasheet but don’t know if the switching between batteries has a delay that would shutdown the Electron because it’s not receiving power.
Interesting. I already have a 3.6V 14Ah LiSOCl2 battery on the way, that supports 1.8Ah peaks. If it doesn’t work as I hope (or if I just feel adventurous ) I will try what you and Alph suggested and see how far I get. The thing is though, would the LiSOCl2 be recharging the LiPo? As in my code whenever my sensor isn’t in deep sleep, it spends more than half of its time retrieving geodata (making call to cell tower), so I imagine that would drain the LiPo quite a bit as well.
You mean, the battery can support a peak current of 1.8A?
Yes, running code continuously means that the microprocessor is always on, which also drains the battery. Unless you are in a deep sleep mode (both cellular and microprocessor modules off), your battery will not last 2-3 years, even if it’s 14Ah.
What you should probably do is use semi-automatic mode:
Turn off cellular modem.
Take a measurement, then go to normal sleep.
Wake up, take a another measurement, go back to sleep.
At the end of the day, turn on the cellular modem, output the data to the cloud, rinse, repeat.
My Electron actually will be in deep sleep mode for 99.9% of the day, only taking measurements for 1-2 minutes per day, before going back to sleep. That’s why I ultimately went with LiSOCl2-batteries. They have super low self-discharge rates of only 1% per year.
The new battery arrived by mail yesterday actually and it works like a charm - enough voltage and can handle the peak currents well.