|
|
|
new temperature sensor for the solar panel
Sunday, January 26 2025
Last night while we were watching the School of Rock kids performing Kool and the Gang and Earth, Wind, and Fire, Gretchen was exchanging text messages with the new tenant at our Rochester rental. Apparently the sparker was no longer igniting the burners on her gas stove top. Most people would know an easy (if temporary) fix for this: use a fucking lighter. But for some reason this thought hadn't occurred to the tenant. This morning Gretchen and I were trying to zero in on what the problem in the stove might be. Since the problem seemed to be intermittant and always affected all the burners simultaneously, it sounded to me like the problem was probably in the step-up transformer that takes normal lower voltages (such as 120VAC and steps it up to several thousand volts, enough for a spark to jump across a macroscropic distance filled with gas, thereby igniting it. Determining exactly what replacement part wasn't easy; first we had to get the model number of the stove, which our tenant eventually found. Then I had to converse with ChatGPT to get a sense of what the terminology was for the various parts of the stove that initiate a flame via electricity. It turns out that the oven uses a different system (a "bake igniter") to get gas burning. The stove top instead uses a "spark module," which includes a high voltage transformer and an electric manifold (or, if you will, "outlet strip"). I eventually found the correct device for sale online and bought it. With shipping, it cost about $70 and will be arriving on Tuesday.
While Gretchen was doing the initial (non-ChatGPT-aided) research on that issue, I was tackling another household chore. After over nineteen years of use, the temperature sensor in the homebrewed hydronic solar panel on our roof had failed, and now the controller in the basement had no idea what temperature the panel was reaching. These days when I use microcontrollers to measure temperatures, I often use a tiny device that determines temperature and then sends the information to the microcontroller via the I2C bus. This works great, because it provides accuracy and precision that I do not need to figure out experimentally, and often provides other information as well (such as humidity and barometric pressure). But the environment inside a solar panel is very different from the ones we normally inhabit. Temperatures routinely rise above the boiling point of water, and the panel is about 80 feet from the sensor down a long cable. A normal I2C sensor wouldn't last anywhere close to 19 years inside a solar panel, and even if it did, it wouldn't be able to send data down an 80 foot long wire. Back when I built the solar hydonic system, I used much cruder technology to read temperatures. I used a simple device that provided a variable resistance depending on the temperature, and I used that as one leg of a dual-resistor voltage divider whose voltage was applied to an analog input on an Arduino pin. I then had to figure out what voltages corresponded to what temperatures. But the great thing about this system was that it worked fine at the end of an 80 foot cable, and the temperature sensors seemed to work okay even in the sauna inside the solar panel. It turned out that I still had four of these sensors left from when I initially bought them from Mouser Electronics over 19 years ago. I attached a new one to a pair of wires and insulated everything with shrink tubing. Then I climbed up onto the solar deck (using a shovel to get rid of snow) and then used a drill to make a hole into the solar panel to place the new sensor close to where the old one had been. It started working right away and even seemed to be producing the correct temperatures.
This afternoon was basically a repeat of what I'd done yesterday, though this time without pseudoephedrine. I eventually found a fairly serious bug in the FRAM dump code that seemed to be the source of much of my garbled data. And I celebrated that by painting a few small areas on the laboratory floor that were bothering me. This cycle repeated several times, and before long it was dark outside.
This evening, Gretchen and I drove to the Garden Café in Woodstock to meet Lynne and Gregg for dinner. They'd rescheduled for fifteen minutes later at the last minute, so we went to the Hurley Ridge Hannaford for important staples along the way. We would end up going back there on the way home as well to buy supplies for making a gluten-free chocolate cake after we learned over dinner that tomorrow is Lynne's birthday.
Other things we talked about over dinner included the story of a patient of Gregg's who committed suicide while Gregg was his psychotherapist. Then there was the story of one of their kids almost being scammed by a Nigerian. (I assume the scammer was Nigerian, since the scam was to obtain Apple gift cards.) It seems this daughter got an email from her boss from an unexpected address, which that email acknowledged and said was temporary. Then another email from that same address said that the boss was planning on surprising everyone in the company with Apple gift cards as a Christmas bonus. Before Lynne & Gregg's daughter acted on the request to buy those gift cards, she felt the need to clarify what exactly was being asked. So she pasted the conversation into ChatGPT, which immediately told her that this had the hallmarks of a very common scam. She'd been saved from being scammed by ChatGPT!
For linking purposes this article's URL is:
http://asecular.com/blog.php?250126 feedback
previous | next |
