My Ongoing Battle with Condensation (Part 2)

So after discovering my Night Sky Pi dome was turning into a mini rainforest every night, I bit the bullet and ordered a couple more BME 280 sensors. I already had one kicking around, but figured I’d try my luck with what I call the “sensor lottery” - hoping to get two from the same manufacturing batch so their readings would be more consistent. No idea if this actually matters for these cheap little sensors, but after years of PC building, I’m paranoid about component variations. Better safe than sorry, right?

Before sticking these things outside in the elements, I figured I should probably make sure they weren’t wildly different in the first place. I wired them up to the same board, which meant having to hack one of them by cutting the addr+1 pads so they could play nice together on the same I2C¹ bus. Then I grabbed two copies of my environment checker script, tweaked the sensor addresses and output files, and had them log readings every 15 minutes for a day. Here’s what I ended up with:

Looking at the graph, there are a few weird blips here and there, but nothing that would keep me up at night - the differences never went beyond half a degree. Close enough for what I’m trying to do! I just wanted to make sure I wasn’t going to get completely bogus readings from one sensor compared to the other.

Deployment

I stuck one sensor inside the Night Sky Pi dome (obviously, since that’s where all my moisture problems are happening). The other one went outside but protected inside a Stevenson Screen² that I 3D printed - found the file on printables if anyone’s interested. Like my initial test setup, I configured both with the environment checker scripts running every 15 minutes via cron.

I wasn’t satisfied with just two data points though. I’m pretty lazy when it comes to hardware maintenance, so I was wondering if I could get away without an external sensor at all. So I set up a third instance of my environment checker that pulls data from the OpenWeather API instead. This way I can compare the external sensor readings with the API data. If they’re close enough, maybe I can ditch the external sensor completely and just rely on the API. One less thing to worry about!

Data Collection

I’ve had all three systems up and running since May 1st, and I’m planning to keep them going for a full month. Thirty days should give me enough data to catch all sorts of weather patterns - rainy days, sunny spells, whatever England decides to throw at me this month. With my luck, we’ll probably get a freak heatwave or something now that I’ve set all this up.

The goal is to collect enough readings to actually understand what’s happening with temperature and humidity inside vs outside the dome. Once I have that, I can finally figure out whether I need to focus on heating, ventilation, both, or something else entirely to solve my fogging problem. I’ll post another update when the data collection is done and I’ve had time to make sense of all the numbers.