1% of temperature.

So, at 0°C, no error?

How about at 32°F?

What about at 273.15°K, or 491.67°R?

 

Generally you want a thermistor that has about 1kΩ to 10kΩ resistance at the center of your temperature range.
Pick one that is rated for 1% accuracy of value over the range.

If you connect the thermistor in the feedback circuit of an op amp in a quasi-bridge connection, then the output is a linear function of the thermistor resistance.
The LTspice of such a circuit is shown below for a typical thermistor:

The values shown give an output of zero volts at 20°C.
For your circuit you would likely want zero output at -5°C.

The gain is a function of the bridge reference voltage, Vb, which should be a stable voltage from a reference circuit, such as a TL431.

If you need more gain to match the micros input scale, than add a non-inverting op amp with gain at the bridge output.

Thanks a lot for your support.
That's what I wanted to know.

Is there any specific type NTC needed for water application.
What is your opinion about schematic posted in post #19

Thanks again !!!

 

So, at 0°C, no error?

How about at 32°F?

What about at 273.15°K, or 491.67°R?

Thanks for your reply !!!

My temperature range -5degC to 35degC. so I have to measure temperature in these temperature range.

 

Thanks for your reply !!!

My temperature range -5degC to 35degC. so I have to measure temperature in these temperature range.

And I gave you 4 values within that range. What is the error budget for each?

 

And I gave you 4 values within that range. What is the error budget for each?

Thanks for your reply.

Actually I did not get purpose of those 4 values.
Anyway temperature accuracy is 1%.

Regards

 

Generally you want a thermistor that has about 1kΩ to 10kΩ resistance at the center of your temperature range.
Pick one that is rated for 1% accuracy of value over the range.

If you connect the thermistor in the feedback circuit of an op amp in a quasi-bridge connection which maintains a constant current through the thermistor, then the output voltage is a linear function of the thermistor resistance.
The LTspice of such a circuit is shown below for a typical thermistor:

The values shown give an output of zero volts at 20°C.
For your circuit you would likely want zero output at -5°C.

The gain is a function of the bridge reference voltage, Vb, which should be a stable voltage from a reference circuit, such as a TL431.

If you need more gain to match the micros input scale, than add a non-inverting op amp with gain at the bridge output.

Could you share LT spice file I don't find in LT spice library .
How to look resistance vs voltage graph in lt spice

thanks

 

Anyway temperature accuracy is 1%.

So at 1°C you want to measure to ±0.01°C? Seems unreasonably precise.

 

Generally you want a thermistor that has about 1kΩ to 10kΩ resistance at the center of your temperature range.
Pick one that is rated for 1% accuracy of value over the range.

If you connect the thermistor in the feedback circuit of an op amp in a quasi-bridge connection which maintains a constant current through the thermistor, then the output voltage is a linear function of the thermistor resistance.
The LTspice of such a circuit is shown below for a typical thermistor:

The values shown give an output of zero volts at 20°C.
For your circuit you would likely want zero output at -5°C.

The gain is a function of the bridge reference voltage, Vb, which should be a stable voltage from a reference circuit, such as a TL431.

If you need more gain to match the micros input scale, than add a non-inverting op amp with gain at the bridge output.

Hello,

Could you send me simulation file of LT spice for temperature range -5degC to 35degC.
Pl explain how did you plot this graph in LT spice and How do you calculate temperature.

Thanks !!!

 

So at 1°C you want to measure to ±0.01°C? Seems unreasonably precise.

This is what he keeps repeating. If the actual temperature is 1°, then his circuit must return something in the range of 0.99° to 1.01°!

Which is ridiculous... Most components won’t be able to support that accuracy. Add the resistor accuracy to the supply accuracy to the op amp accuracy, and he’ll require component ratings which don’t exist.

Unless he means something else?

 

Is there any specific type NTC needed for water application.

If you are going to immerse it in the water, then you need one that's waterproof.

What is your opinion about schematic posted in post #19

It's not as linear as the one I posted.

Anyway temperature accuracy is 1%.

You aren't listening and still don't get it.
1% of what?
If the temperature is zero, then based upon your definition, the accuracy would also be zero, which is nonsensical.
So how about 1% of the range (40°C) which would be 0.4°C?

 

Could you send me simulation file of LT spice for temperature range -5degC to 35degC.

See attached.

Pl explain how did you plot this graph in LT spice

I plotted it as shown.
What do you not understand about that?

How do you calculate temperature.

The output is determined by the gain of the circuit.
The bridge is biased at 1/2 the bridge voltage by R1 and R2.
The output is then (Vb/2)/(R3+Ru2) * Rt.
Calculating Rt from the output voltage by plugging into that equation, you can then determine the temperature that caused that value of Rt.

 

If you are going to immerse it in the water, then you need one that's waterproof.
It's not as linear as the one I posted.
You aren't listening and still don't get it.
1% of what?
If the temperature is zero, then based upon your definition, the accuracy would also be zero, which is nonsensical.
So how about 1% of the range (40°C) which would be 0.4°C?

I beg a pardon , if there is gap in my understanding.

Temperature range: -5°C to +35°C
I realize that temperature accuracy is +/- 0.5degC and not 1%.
You people were correct.

Sorry again.

Could you post LT spice simulation file and explain how did you plot the graph.
Pl do simulation for -5degC to +35degC.

And calculation for temperature.

It will really help me.

 

Where/what is this water whose temperature you want to measure? If it's turbulent or subject to rapid temperature fluctuations then an accuracy of even ±0.5° may not be achievable.

 

See attached.
I plotted it as shown.
What do you not understand about that.

The output is determined by the gain of the circuit.
The bridge is biased at 1/2 the bridge voltage by R1 and R2.
The output is then (Vb/2)/(R3+Ru2) * Rt.
Calculating Rt from the output voltage by plugging into that equation, you can then determine the temperature that caused that value of Rt.

Thanks for sharing simulation file.
Could you let me know which analysis you have done to get linear graph.

 

Where/what is this water whose temperature you want to measure? If it's turbulent or subject to rapid temperature fluctuations then an accuracy of even 1% of the temperature range (0.4°) may not be achievable.

I beg a pardon.

Accuracy is +/- 0.5degC not 1%.

I am really sorry . Actually misunderstood and later I realized.

Regards,

 

hi m87,
Have run this LTS sim, you can create your own NTC, but you will need a circuit as posted by crutschow to linearise the response.
E

 

Could you let me know which analysis you have done to get linear graph.

The output voltage is a linear function of resistance if the thermistor has a constant current through it.
The bridge circuit does that, since the current through the pot and R3 does not vary (there is a constant voltage across them), and the thermistor current has to equal that current to keep the bridge circuit balanced, so thus the output is a linear function of resistance.

Remember that, with negative feedback, an op amp will do whatever it can to keep the voltage across its two inputs very close to zero volts.

 

Where/what is this water whose temperature you want to measure? If it's turbulent or subject to rapid temperature fluctuations then an accuracy of even ±0.5° may not be achievable.

Its distilled water and carbonated water for which we have to measure temperature.

example : Hot and Cold Electric Water Cooler

So assume you have to measure temperature of cold water and based om temperature you take your decision for refrigeration.

What is achievable accuracy in your opinion.

Regards,

 

The output voltage is a linear function of resistance if the thermistor has a constant current through it.
The bridge circuit does that, since the current through the pot and R3 does not vary (there is a constant voltage across them), and the thermistor current has to equal that current to keep the bridge circuit balanced, so thus the output is a linear function of resistance.

Remember that, with negative feedback, an op amp will do whatever it can to keep the voltage across its two inputs very close to zero volts.

How does these values comes from. What is calculation for that.
What is the Opamp gain to measure -5 degC to 35degC

Thank you so much.
This forum helped me a lot whenever I got stuck.

Regards,

 

How does these values comes from. What is calculation for that.

No calculation.
The resistor values vs. temperature are from the thermistor data sheet.

What is the Opamp gain to measure -5 degC to 35degC

Depends upon what output voltage range you want for that temperature change, and which thermistor you select.