Beehive Monitor

So I've had the project in-place and operating for about 2 months now. After fixing the wiring issues I had early on, it's continued to work quite well. For the last few weeks it's shown a steady (but also rather sudden) increase in weight, which I HOPED was accurate but couldn't be sure of until recently when we opened up the hive to check on things. The first box was getting close to full, which anecdotally matches up with the weight sensor readings. Added a second box, which hopefully they will fill up in the coming months!

While the strain gauges do still have a bit of noise, it's manageable for my needs. I've found that simply using a 12-hour rolling average smooths out the noise quite well. While it'd be cool to see more subtle changes in weight as the bees come and go throughout the day, tracking the weight longer term is the primary goal of things anyway. I also subtract out the known weights of the boxes, so that as I add new ones it doesn't skew the readings. Additionally, I've found using a daily average and plotting over the last month to be the most useful way of digesting the data:

A few days ago I finally had some free time to rewire the strain gauges and remove the splitter phone plug connections; I ended up simply stripping the wires and soldering them together using heat-shrink solder connectors. This meant I went from 5 phone plugs to just the single one on the PCB. I figured it would improve things, but only somewhat and that I would still need to ultimately either solder the wires to my board (or in a future PCB design, use screw terminals). Surprisingly, the readings immediately fell in line with what I would expect based on my previous calibrations (obviously I can't truly calibrate things, because there's an active beehive on the scale now). Even better, I played around with the phone plug and didn't get any wild changes in readings like I had previously with the other connections. It currently shows ~15lbs of bees, comb, honey, etc. which seems reasonable and when I added a known weight it increased by this amount. In the days since, it's also stayed relatively consistent; while there is a fair bit of noise in the data over time, I've found that if I use a healthy average it stays quite consistent too (found a 12 hour average works well, but am now trying smaller averages to see how low I can go). Time will tell if it continues to work well; I'm still planning to begin designing and testing a new scale for any future hives (and possibly for this one if my existing scale appears to fail again) using bathroom scale-style strain gauges I already have.