Brawo Roland!
Kawał porządnej badawczej pracy! Jestem pod wrażeniem. Z niecierpliwością czekam na udoskonalony soft. Jestem dupa z matematyki więc wykresy jeden pod drugim będą dla mnie bardziej zrozumiałe. Z chęcią poanalizuje pozostałe wykresy ale odpowiedz mi na dwa pytania:
czy odczyt temperatury jest ciągły?
jakie jest opóźnienie pomiędzy zmianą temperatury a zarejestrowanym odczytem?
Pozdrawiam,
Mikołaj
ps.
znalazłem też trochę porad racjonalizatorskich odnośnie projektu TS2:
Use a thermistor rather than a diode as temperature sensor. These devices are made for the purpose. A value of 10 kOhm should be fine.
Create the reference voltage that is applied across the sensor with a REF02, rather than simply take the general supply.
The version 2 idea to measure the difference of identical two-sensors in different thermal environment seems good to me.
Put the sensors in a wheatsone bridge (
en.wikipedia.org/wiki/Wheatstone_bridge
). This minimizes the effect of variations in the reference voltage.
Use a resistor array for the reference resistors in the wheatstonebridge. Glue the array to a piece of aluminum. This minimizes errors due to temperature differences of the reference resistors.
Use a decent instrumentation amplifier to evaluate the wheatstone bridge. These devices are made for the purpose. LT1167 may be a reasonable choice. Using inverters for linear amplification is a hack.
No amplification stage should amplifier more than about 100. Use more stages if more amplification is needed.
Use low offset, low drift opamps for the amplification stages. The classical choice would be OP07, which affordable, too.
The largest resistor value in the circuit should be 100 kOhm. With multi MOhm values, the Johnson noise of the resistors becomes a problem.
Use comparators to decide whether to beep, or boop, or fall silent. Again, these devices are made for the purpose. I'd choose the LM393.
Put the circuit in a rugged aluminium enclosure. E.g. Hammond 1590BB