I’m still not entirely sure why they didn’t just stay up on Tuesday – it could have been the electrical connection wasn’t steady enough, but there was something about what we left that the radios didn’t like.
We’ll have to wheel it out (carefully!) on the 9th and have a play to find out what’s going on with it. John claims it was reacting quickly every time I walked by, but I’m not convinced I was particularly radiant that night, so we need to think about why! One thing we may want to think about is radiation shielding for the Arduino and battery – we have some aluminium foil we got in on purpose for this. The plastic globe is also ready to go, except we need to find a place to hang it, not too close to the copper one. We may try to do that ahead of the 9th to increase the science and decrease the time spent checking connections.
We were getting readings, and differences between air temperature and black globe temperature – right up until we started the pi logging to the web, set the rig in position for the night and quick cleared up so the caretaker could lock up. I’ll check the rig (and take pictures!) when I’m next passing, and then I think the first test is what happens when the heating goes off for the night and starts up in the morning, assuming it does. In these changes, air temperature should respond faster than the walls and show up as different patterns on the log.
The very observant at Queensferry Parish Church might notice the HeatHack logo dotted around the premises – if you do, and you’re wondering what those funny devices are, they’re taking temperature measurements. You’ll get your results in about three weeks. Meanwhile, you can get some sense of what they’re doing and why by looking at our previous data collection on our sister website, Science of Church.
Here’s a simple test of our monitoring equipment, operating in the worship space at Christ Church Morningside. It’s not even soldered together – the sensor location is just a mini-breadboard, Jeenode (Arduino plus radio transceiver) and battery pack, with a Pi plus Jeenode picking up the data and shipping it out to the web. We’ll replace it with a soldered version if it doesn’t stay up, but the main question for us is: do the batteries last long even with our brain-dead approach to power management, or does deployment need to wait for us to understand that better?
You can find other views of the data at the live feeds page. We’ll add one from City of Edinburgh Methodist as soon as we’ve figured out how to get past their wifi security!
The black globe thermometer challenge was fun, and we learned a lot by running it. One of the things we hadn’t quite realized was how long some of the parts would take – especially getting the paint to dry and the temperature readings in an enclosed space to stabilize. That means we aren’t quite finished. On Tuesday the 25th at 19:30, we’ll be cracking open the Sugru so we can robustly assemble at least the reference globe (made from a copper WC float) and a couple of the cheap alternatives, so that we can start logging the readings to a web service and find out how they perform.
Testing itself creates a whole new set of challenges. We want to know if the material and size affect the inertia of the thermometer. The most natural test would be to check competing models side by side, but they would influence each other’s readings by occluding other surfaces. Our proposed test rig is a cardboard box surrounded by bubble-wrap, but we still haven’t talked about whether we need to add the radiation shields (otherwise known as aluminium foil). We can get surface temperature readings for it either using the same temperature sensors as are in the globe or from infrared devices we can calibrate; what’s harder is controlling what those temperatures are. We’ll talk about this on Tuesday, but perhaps we want some simpler tests first that don’t use the rig:
(1) can we get readings that are different from the air temperature? If not, then there’s some kind of problem with the globe enclosure.
(2) can we observe changes in the difference between globe and air temperature readings that make sense based on environmental changes? A quick version of the test might make use of the angle factor, and look for changes in readings from moving closer and further away from the window, as long as we remember to take air temperature and draughts into account. We can also make use of heating changes. For instance, if the heating has been off for a few days, and goes on, we can be sure that the air temperature will go up faster than the surfaces. The Methodist Church is heated every day, but there are other churches we can try where the difference between air temperature and mean radiant temperature will be more pronounced.
We’re just finishing off our preparations for the black globe thermometer challenge on Friday – drilling holes in WC floats, looking for non-conductive hollow tubes, and making sure our software works. Meanwhile, for anyone who couldn’t make it and wants to know what was in the optional lecture, you can find the slides here.
At Tuesday night’s hackerspace (7:30, City of Edinburgh Methodist Church) I expect we’ll be exchanging experiences of getting data from our test sensor setups to the emoncms.org servers, and working on how to authenticate for the wireless network on site even though the usual procedure requires a human in the loop.
Meanwhile, we’re gearing up for Friday’s black globe thermometer hacking challenge – and pleased that we’ve been able to source the traditional option for a casing (a copper WC float), but modern equivalents in a range of shapes, materials, and sizes. It’s going to be an unusual mix of physics, aesthetics, and practical skills where all of us have something to learn.