I built a digital thermometer 15 years ago with a transistor as the probe. I used Celsius to calibrate it. Ice water for freezing and boiling water for 100 C. It worked very well. On Sat, 2006-03-18 at 20:20 +0000, Martin Vlietstra wrote: > I did an interesting experiment today. I have a digital thermometer with a > probe on the end of a three-metre cable. The thermometer also has a C/F > switch. > > I set the thermometer to °C and held the probe between my thumb and finger. > It registered a temperature rise. Once the temperature moved above 30°C, > the temperature continued to increase in 0.1°C increments, the time between > each increment becoming larger as the temperature approached body > temperature. > > I allowed the probe to cool and then repeated the experiment, this time with > the thermometer switched to °F mode. Once the temperature moved above 80°F, > it continued to rise, but the increments had what to the outsider would > appear to be an illogical sequence - there were four increments of 0.2°F, > followed by one increment of 0.1°F, another four increments of 0.2°F, > followed by one increment of 0.1°F and so on. The result was that one would > get five consecutive different readings in which the tenths of a degree > digit was even, followed by five where it was odd. > > I can see those who understand the technology itching to tell me the reason. > It is quite simple - the thermometer's electronics is designed around a > counter (within an ADC) whose value increments whenever the temperature > increments by 0.1°C. The C/F switch merely converts the Celsius display > (already rounded to the nearest 0.1°C) to the closest Fahrenheit value > thereby giving a further level of approximation. > > I suspect that most digital measuring instruments are designed around metric > units and that a Customary or Imperial Unit switch merely converts the > output rather than switching in alternative electronics. Any comments > anybody? >
