How about just measuring the difference in level rather than the set level.
IE: your float must have a pot on it to send its level to... whatever (arduino?). When you power it up take the first readings as the "cool" temp, and then average it with the last 5-10 cool reading levels (in case you warm-started, that running average would save your bacon).
Then your louver opening/fan on temps would just be a certain level above this... to be experimentally determined. You would only run into issues if your coolant bottle went completely empty, but you could even add a check for that in your ardy code if you wanted (if the initial value is less than ____, turn on an LED).
|