Linearization of thermistor

[Nivii1406]

What is the need for linearization of thermistor? How does ADC help in linearization?(Output of thermistor is connected to a log amplifier which is then connected to a dual-slope ADC)

 

[FvM]

You didn't tell what "linearization of thermistor" means for you. In case of a temperature measurement, you would specify range and intended accuracy. Log amplifier seems to me as a possible but unwantedly complex method. I would go for higher adc resolution and numerical linearization.

 

[Nivii1406]

Here, it is linearization of thermistor for the purpose of temperature measurement.
I intend on constructing an ADC for the purpose of linearization using hardware. So the use of a log amplifier seemed more appropriate.

 

[FvM]

If you decided to make the design this way, what are you exactly asking?

 

[pratomer]

to measure temperature and get digital data, you must have uniform quantization and thermistor characteristics is exponential hence, to get the output voltage to vary linearly with temperature you need a log amplifier. Next ADC will do quantization to provide digital data.

 

[FvM]

Are you familiar with the design of log amplifiers? If you consider the effort to get a precise and stable characteristic, you'll at least think about other options.

 

[chuckey]

You do realise that you will need three points of temperature calibration? This is no mean task. It is better to start with a more suitable temperature sensitive device.
Frank

 

[FvM]

Depends on the application. There are thermistors with fairly accurate characteristic, e.g. the 2780 ohm@25°C standard type used in HVAC, which has 0.5 to 1 K initial accuracy over a wide temperature range. You'll take care that the measurement circuit doesn't add too much error.

 

[p11]

i think if you don't linearize the circuit then you will not connect any series resistance but this may increase the current through the thermistor to such a high amount that it can cause self heating error.so to avoid this error you have to connect a series resistance to it,and this series resistance will no doubt linearize the circuit,the only diff is that the inflection point will be shifted.

- - - Updated - - -

can anyone explain me the clear concept of inflection point and why at that point the 2nd derivative becomes 0????????...plz as soon as possible,i have to submit my project,please answer me as soon as possible...:|

 

[chuckey]

Look here :- https://en.wikipedia.org/wiki/Inflection_point
Frank

 

[p11]

OK ,that means inflection point is that point where the concavity or curvature changes sign,but in case of thermistor how the cancavity is changing sign at the inflection point i can't understand...

https://mathscinotes.wordpress.com/2011/07/22/thermistor-mathematics/


check figure 3 of the above link ,and please explain me that how 50 degree centigrade becomes an inflection point,i mean how at that point the concavity has changed from positive to negative or from negative to positive.

 

[D.A.(Tony)Stewart]

Inflection point is where the 2nd derivative = 0 or the change in slope of 1st derivative is linear when used in an R Ratio expression. as shown in figure 4

 

[p11]

Inflection point is where the 2nd derivative = 0 or the change in slope of 1st derivative is linear when used in an R Ratio expression. as shown in figure 4

can you please explain once again, i want to know the physical significance...i mean how at 50 degree centigrade the 2nd derivative is becoming 0.please help me...

- - - Updated - - -

can anyone give me some idea that what should i write as the scope of the project of temperature measurement using an NTC thermistor.

 

[D.A.(Tony)Stewart]

The slope of the ratio of the thermistor R curve (orange) is shown in purple.

The slope of the purple curve is constant (1st derivative) is the gain.

(2nd derivative shown hand scratched in Red (not to any scale) where linearity is best over this region shown by degree of flatness.

Shifting the midpoint would shift the 2nd derivative ( red ) left or right.