OT: Room AC update

Borrow a resistance decade box...experiment...repeat as necessary...purchase appropriate resistor...solder in place of decade box...return decade box to rightful owner...report results to group.

Vaughn

Too much common sense, I need a more eclectic answer.

#2 looks good. So, it is 13K at 68F. But, you probably want the evaporator running quite a bit cooler than that, maybe 40F or so. So, with the thermistor at 40F, it needs to tell the controller that it is 68F, say. So, say it actually reads about 25K Ohms, and we want the controller to think it is

13K Ohms. So, 25K in parallel with X gives 13K, and we need to find X. solving, I get 1/X = 1/13K - 1/25K , and X = 27K. So, parallel it with a 27K resistor, and you should get your desired temperature showing up somewhere in the settable range of the controller.

Jon

"Tom Gardner" wrote in news:DSzpc.8166$ snipped-for-privacy@newssvr28.news.prodigy.com:

You really need a few more points. From looking at the data: Resistance (k) Temperature (F) Rate Of Change (Resistance(k)/Degree (F)

37 34 13 68 0.705882353 5.4 104 0.211111111

The rate of change is different from 68 to 34 vs. 104 to 68, by a considerable amount. A few more points closer to your desired temperature would allow you to fairly accurately determine the proper resistance.

The thermister is a pretty good fit for an exponential curve (as in R^2=.9956 or as good as it gets!) Equation is k ohms =

90.608e-0.0275 *temp in degrees F

The parallel resistance required to raise the sensor reading will change depending on what the base point is, a good start for your applicaton would be about 70k ohms.

Have fun!

Tom Gardner wrote:

I would try #2, but keep in mind that if you drop the evaporator temp below

32F the coil will freeze up from condensation. You may not get much colder. Greg

If I was intent on doing this, I would probly use a potentiometer and do some experimentation.

Here is a parallel resistance calculator for any who dont care to memorize a formula to do repetetive calculations they might only have occasion to use once in a blue moon :

Tom;

Move the thermistor so it does not contact the coil. Add a little insulating material so the thermistor does not contact the coil.

JRW

Tried that. I get one loong cooling cycle then the thermistor stays cold longer in warmer room air. I even tried a warm glass of water with the same results. Sorry about being obcessed with something stupid but now I'm on a mission.

An approximate equation based on the Steinhart-Hart equation for NTC thermistor behavior is found at

I did a curve fit using this equation form on your three datapoints after converting your temps to degrees Kelvin. In the form of the equation found on that website (I hate typing equations in ASCII), the coefficients were:

R0 = 30512, B = 4347.8, To = 277.35. To probably should be 273, but this isn't a bad fit for three measured datapoints.

From that, I estimate that the resistance of your thermistor will be about 16.6K at 288,5 Kelvin or 60 deg F -- your 8 degree difference if the kickoff point is evaporator temp.

The controller trips at 13K if kickoff point is evap temp, so you need to parallel your thermistor with 62.3K. 62K is a standard 5% value. I'd use a 100K pot.