precision thermistor temp. sensor circuit project

Dear All About Circuits er's

i started a project to make a egg incubator machine
the most important thing is the temperature accuracy whuch must be 0.1 degree Fahrenheit

i selected the precision thermistor from ussensor which is 30k ohm and generated the look-up tables for temp. Vs. resistance with the target accuracy

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=615-1004-ND

i want you to suggest the circuit that i will integrate with the microcontroller
how to connect the thermistor and the ADC should be used to get that accuracy with reduced self heating effect and long term stability

should i reduce the ohms of the sensor ? selecting for example 10k ohms ??

and an important thing is the calibration ability through cermet variable resistance

any recommendation will be much appreciated

thanks in advance

You might benefit from this project.

I can tell you that 0.1°F is relatively challenging, but the circuit I used in that project achieves that level of precision. Calibrating your system to that level of accuracy is not trivial, either. Do you have a reference thermometer of some sort?

As a really old person, I'm with wayneh on this one. I did a one chip analog controller 30 years ago that was accurate to 1/4 degree F, and it still works. The worst problem was that it put the triac into partial cycles and the heater hummed. I didn't need that much accuracy, so I increased the hysteresis, but I could have added another op-amp to get hard switching, much like the schematic that wayneh offered.

What's up with this, "put a microprocessor in everything" mentality?

Excuse me. It's time to feed the mule.

Thanks guys for your help
but
the sensor LM35 is not accurate at all regarding incubator controller
and we have to much to control rather than the temp." humidity ,alarms,doors,fans, so that i went for microcontroller

you don't think that ADC with 10bit can work precisely ?

f_elngar said:

the sensor LM35 is not accurate at all regarding incubator controller

Click to expand...

Are you claiming your thermistor circuit can give higher accuracy and precision? Pardon me if I'm skeptical. At the least you'll have to calibrate against a reference. And for that level of accuracy, you'll need an NIST traceable reference.

But I do understand your need for a smart controller. If you have a spare computer around, you could use a LabJack to acquire and log all the data, and to perform the control functions.

Now that you mention several other jobs that need to be done I can understand the use of a microcontroller.

As a repairman, I just get tired of seeing microcontrollers in jobs that can be done much more simply. They are so much more difficult to repair than a thermistor and op-amp circuit, and more expensive!

For instance, a couple of weeks ago I repaired a refrigerator with a defrost problem. When I lived in a cave, a defrost timer was a mechanical clock and the part was $35. Now that small microprocessors can be had for less than $2, the repair costs $85 for a controller board. The way things are going, I won't be surprised to find flashlights with microprocessors in them before long.

f_elngar said:

...
... the sensor LM35 is not accurate at all regarding incubator controller ...
...

Click to expand...

Huh? Who said that? The LM35 is a precision temp sensor that outputs linearly voltage per 'C. The datasheets and appnotes have many example of doing precision temp control and sensing using the LM35.

I would think it is a much better sensor for your incubator than a thermistor due to the linearity. Both sensors will need to be calibrated.

I wouldn't trust anyone who planned to use "The LM358 quad op-amp". No such part!

I'm another one who's skeptical about highly accurate temperature control using a thermistor. The LM35 seems like a better bet, as it's linear and therefore much easier to calibrate. But I'd look at a platinum Resistance Temperature Detector too. There should be some kind of benefit to be gained from the fact that this system only operates over a very narrow temperature range.

Do you expect to use a PID (or other) control algorithm, or will simple on-off control be good enough? Better plan for it from the start.

simple on-off control

If you have a lab grade temperature reference you can calibrate a off the shelf thermistor to the same degree as a PRTD sensor. The difference is a year from that point the PRTD sensor will still be in spec and the thermistor based sensor was out of cal in a week unless it was preaged for at least a year at 100C.
http://www.omega.com/temperature/z/pdf/z036-040.pdf
Calibration in this case is deriving a table of the resistance to temperature changes.

John P said:

I wouldn't trust anyone who planned to use "The LM358 quad op-amp". No such part!

Click to expand...

Thanks for pointing out a typo, which I've fixed. I got it right - "dual" - on the parts list but this error was in the sentence after talking about the quad comparator, so ... boo boos happen.

Boo boos happen--oh indeed they do. And look at the good side, people are reading your stuff!