Hello, I have built a monitor for a remote water holding tank. My system has been deployed for 5 days now and it’s working really well it would seem. Having said that, tt’s been raining and overcast (see picture) every day until today, when I got a few hours of sun. They system is running off of an 1800mAh LiPo battery and I have a 5v solar panel facing south with out much obstruction. My battery has been very slowly dropping and today it dropped below 75%. It seems to me that the solar panel has been helping considering I still have between 50% - 75% after 5 days of use. The radio is only sending updates 3 - 5x/day, although I’ve got some code cleanup to which I think will drop battery usage.
I have 2 more solar panels, and I’m thinking of adding another one in parallel to keep the voltage to 5V, but reading the tech. details on the RAK19007, I can’t find a rating for the max allowed current for the charging circuit. Is there anyone on there that has any idea of what the limit is? The nominal current for the board I have is 380mA. These are the 5" panels on Rokland’s website.
I decided not to add another panel and just see how long I could run. The system has been running since the 13th and we’ve had very little sun and then about a week of rain. Today we got a few hours of sun and the battery is recovering well:
I implemented some code to send battery percentage on the 18th, so I don’t have info prior, but man, I’m impressed with how little power this system uses!
My sketch is very simple. I have a float switch in my holding tank and in the main loop, I check the state of switch, determine if it’s changed, and if so I send the updated state to another radio. I only send when the state changes, which only happens twice a day (based on the metrics I’ve gathered from running for nearly a month).
I also send the battery percentage every 5 hours. The code to read the battery comes from the examples, but I added a bit of code to read the voltage multiple times and average it, and I also clamp the value from 1 - 100. That’s all I’m really doing. I’m not intentionally sleeping the device, and it spends almost all of its time spinning and checking the switch.
One area of improvement I’ve been working on is, this radio sends the switch state over and over ever 5 seconds until it gets an acknowledgement from the listening radio. Sometimes I notice that the message needs to be sent 5 - 7x before I get the ACK. I’m testing in my office with a mock set of radios, and I figured out yesterday, I’m basically interfering with the other radio with the way I’m sending. I knew this was a possibility based on the fact I’m using the raw TX/RX code. I’m updating that now and I suspect I’ll save a bit more battery once I don’t have to send so many times.
Anyway, this is all new to me and I’m taking it slowly, so I may end up adding intentional sleep code/etc as I progress, but right now I’ve been able to run the remote switch for 19 days off of solar and I currently have 70% battery, so it’s good enough for government work, as they say.
