cannibalizing discrete goodies
Tuesday, November 15 2005
Nearly all of the solar heating system's control circuitry is in place, but one important component is still missing: the stuff that determines when it's sunny enough to circulate water. I've been having to make those decisions in my own wetware and then flip switches, either (assuming sufficient sun) to turn the circulation on at 9am or to turn it off at around 3pm. As you might imagine, I've been thinking a lot about how best to do this robotically. The simplest system would use a conventional aquastat to flip the existing basement "solar sufficiency" relays whenever there was enough heat in the panel. That would work fairly well in a system that could heat water at the rate my pump can pump it, but my sixty square foot panel can't keep up, and water tends to leave it before reaching the high temperatures that makes pumping it worthwhile. An ideal system would pulse the water through the panel, waiting for it to reach some suitably high temperature (say, 175 degrees Fahrenheit) and then pumping it until the water arriving in the basement fell below a certain temperature. That last part is important, because to turn off the pump before the water reaches the basement would leave a lot of perfectly good hot water in the pipes between the roof and where it needs to be. I'd been thinking about using a photocell as the method for initiating pumping, but that would ignore all the dynamic information related to the actual heat of the fluid in the panel. A better solution is to have temperature sensors up on the solar panel calling for circulation in sunny conditions and then waiting three and half minutes after the water falls below a certain temperature before requesting the pump to be turned off.
Before getting to all of that, though, I first had to have some circuitry in place that could read the weak signals coming from either photocells or temperature sensors and then generate a signal powerful enough to throw a mechanical relay, the 19th-Century-era logic of my solar controller. Yesterday I'd had success connecting a digital oven thermometer down in the basement to its probe up in the panel using a pair of wires in the telephone line that parallels the copper plumbing between them. (This indicates that the thermometer is measuring differences in a high ohmage probe.) If I could create a circuit that could react to those differences and throw a relay, I'd be partially in business. So tonight I built such a circuit using a LM339 quad comparator IC and a pair of A673 PNP transistors, all of which I cannibalized from an old color monitor. Today's electronics are so miniaturized and reliant on proprietary chips that it's hard to find circuitry containing any parts that can be scavenged, but thankfully color monitors still use plenty of big discrete components, usually on easily-desoldered circuit boards. I couldn't even find specs for the A673s, but I was still able to deduce their pinouts (they're TO-92s with pins in this order: Emitter-Collector-Base) and build a circuit that was sensitive enough to turn on a 12 volt fan whenever I touched my fingers to one of the input leads.
I haven't done a lot of electronic soldering since the early 1990s, and back then the rosin core flux smelled kind of good. The stuff used in this solder, on the other hand, smelled like melting plastic and was most unpleasant. I don't know if I developed a twinge of chemical sensitivity from my gasoline swallowing incident or what, but by the end of tonight's soldering I was feeling weak both in the head and in the knees, as if I was coming down with the avian flu. It made me nervous that perhaps I was suffering delayed effect from the gasoline, some of the chemicals of which could persisting in my liver and being converted into even worse species of toxins as if by some malevolent, but nevertheless intelligent, designer.
For linking purposes this article's URL is:feedback
previous | next