jhc@alice.UUCP (JHCondon) (02/03/84)
Thermostats are more complex than most people realize. A simple bi-metalic strip sits there making and breaking, sensitive to vibrations, creating radio and tv interference. You avoid that by putting in some hysterisis; two forms are popular, a little magnet on the contact so that there has to be several degrees of heating before the contact opens again, or if it is a sealed Hg switch you mount the switch on top of a spiral of bi-metal so that once the switch falls to one side it wants to stay there. BUT the hysterisis now causes the temperaure in the house to have big swings since the furnace must overheat the house in order for the stat to turn it off, and additionally there is still more heat in the furnace to come out even if it's turned off. Well, humm, what to do. AH HA the ANTICIPATOR. Believe it or not there is a resistor under the bi-metal that is in series with the thermostat contacts. It "knows" the furnace is on and supplies a little heat to the bimetal in anticipation of the heat from the furnace. Well different furnace realys draw different amounts of current, and houses are different so that resistor is usually a variable one. Now if the anticipator is has to high a resistance it overanticipates and cuts the furnace off to soon, making the behaviour ihuxl!seifert describes. OK you software types get your little screwdrivers out. What's that. What's a screwdriver? Oh well I should know better. The above explanationdoes not apply to houses with heating coils cast into a slab floor. Those are bizarre.
rpw3@fortune.UUCP (02/05/84)
#R:alice:-256700:fortune:8600011:000:1901
fortune!rpw3 Feb 5 02:20:00 1984
The use of "anticipator" resistors also has the effect of transforming
a simple first order feedback loop into a "proportional controller"
(assuming the anticipator is set right), also known as a pulse width
modulator (often seen these days in power supplies). The net result
(which also very much applies to floor mounted glass panels, b t w)
is a MUCH more closely controlled temperature.
The thermostat (and therefore the heating system) just sits there cycling
on and off at a period (hopefully) shorter than the time constant of the
room or house. What you feel is then the long-term average of a bunch
of "heat pulses" rather than the surge of the heat pulses themselves.
Oviously, this has got to be adjusted correctly to work right, for it
has to interact (in forced-air furnaces) with the fan turn-on/turn-off
threshold temperatures (which delay the fan starting until the air in the
duct is above a certain temparature, etc.). This puts yet another completely
independent term inside the feedback loop, so adjusting it is tricky for
the average handyperson. A rule of thumb I have used is: On a "normally"
cold day (one that doesn't threaten to overload the heating system), the
fan should come on about half way (or later!) through the gas "burn".
That way, the big pulses the gas makes get smoothed out a little by the
fan/duct system before hitting the house. Another way of saying it is
that the demand for heat (thermostat) should be about "90 degrees out
of phase" (imprecise term) with the response (hot air from the duct)
when the system is delivering about half it's maximum output.
As I noted before, not something the average bloke gets right without
calling the repair crew (who may not know how either!).
Rob Warnock
UUCP: {sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3
DDD: (415)595-8444
USPS: Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065