ntt@dciem.UUCP (Mark Brader) (02/08/84)
The original question by ihuxl!seifert (D.A.Seifert), i.e., why do you have to turn the thermostat up on a cold day, has been answered in similar ways by alice!jhc (JHCondon), heurikon!jeff (Jeffrey Mattox), and ihuxf!parnass (Bob Parnass). All of them raise the point of the anticipator and its setting. I have no experience in control engineering, but living in Toronto, I do have some experience as a user of thermostats. And it seems to me that another reason why thermostats don't work is that they are located in one place. After all, heat is lost from a building through all its exterior surface, and particularly through thin places like windows; and it is added, usually, through a limited number of hot-air outlets or radiators. And the places where people normally are are NEITHER of these, because those places are too hot or too cold. There will be a heat gradient between the hot places and the cold places, and what you want is that most of the living space is on the comfortable part of that gradient. But clearly, the colder the weather, the steeper the gradient. Therefore the cold areas, which already cover a larger area, will seem to grow larger still, and you raise the thermostat to force them to retreat. QED. If the thermostat is placed too close to a heat outlet, this effect will be amplified -- and I have lived in apartments where the thermostat was right next to the heat outlet. If it is placed too close to the walls, the effect will be reversed, and you get too warm on cold days. Since the latter effect is counter-intuitive, I think architects tend to err in the other direction. All of this applies in reverse in the air-conditioning season. In the apartments I mentioned, I used to put my largest floor lamp (the largest single source of heat in the apartment in the summer) next to the thermostat and air outlet during the summer, and I could tune by moving it around. Mark Brader, Toronto, Canada P.S. Queries should not be posted to net.misc and another topic. Some of the referenced replies were in net.physics only. I have posted to both places because the original query was. If there's more to say, let's keep it in net.physics though.