Temperature differential measurement

[Egy]

Hello.
I want to measure the temperature of flowing liquid at the beginning of the tube and compare the temperature at the end of the tube. The liquid heats as it goes through the tubing. The difficult part is that the tubing inner diameter is only 125 micrometers, and I want to be able to measure the temperature difference with 0.001 degrees C precision if possible.

Sorry, I am very new to this. If you could get me started with how this can be done, I'll research more into. I just couldn't find a good starting point, and my researching left me more confused.

One idea that I came across was to use a feedback circuit with thermopiles to measure the temperatures, which is used in calorimeters that use a reference cell and a sample cell. The feedback circuit is used to heat the cooler end and I can measure the differential power signal. However, this does not seem to be the best way, since I don't want to heat the other cooler end.

 

[crutschow]

The most accurate temperature measurements are usually done with Pt temperature sensors.
You might look into those.
You could use two sensors, each in one leg of a bridge circuit, which should give good sensitivity to the temperature difference.

 

[flatfootskier]

Hello.
I want to measure the temperature of flowing liquid at the beginning of the tube and compare the temperature at the end of the tube. The liquid heats as it goes through the tubing. The difficult part is that the tubing inner diameter is only 125 micrometers, and I want to be able to measure the temperature difference with 0.001 degrees C precision if possible.

So you're not fussed about accuracy, just precision & I guess you'll calibrate one against the other? Even so, with one thousandth of a degree target precision I'd think it'll be a challenge. Trying to get more than one probe to read the same temperature in a fluid isn't easy. At least you won't have convection!
What temperature range? I'm guessing your DeltaT might be multiplied by the flow rate? How quick a response rate do you need? What's the fluid? What pressure?
With that level of resolution I'd think you'd be experiencing a change in temperature due to obscure factors like changes in potential energy if not perfectly level!
I predict calibration hell!
Interesting. Tell us more!

 

[Egy]

I'm thinking the temperature change would probably be in the hundredths of a degree at around room temperature. A chemical reaction with a stationary support in the column, and since I set the flow rate and the tube inner diameter, I know the reaction time. So, depending on where I put the probe in the tube, I expect the temperature to be constant. Therefore the response time doesn't matter, I can can keep flowing until the probe equilibrates with the temperature of the liquid at that point in the tube (I hope). The liquid would be non-viscous like water or organic solvents (acetone, ethyl acetate, etc.). The pressure should be almost zero or at most 10's psi. I guess thousandths of a degree may be too much to ask for. Could I at least get hundredths of a degree? Do I sound to naive, or am I missing several other variables?

The most accurate temperature measurements are usually done with Pt temperature sensors.

I'll read more into the Pt temperature sensors. Thanks.

 

[flatfootskier]

Mechanical mounting of the probe in the fluid might present challenges - PT100 sensors are very accurate/precise but also quite big in comparison to your pipe diameter, so they then need to go in a larger vessel, where the flow rate would rapidly diminish & convection might set in, together with external influence (conduction through wall primarily). Room temperature will be all over the place. I expect you'll have to insulate the whole arrangement.
I assume you know how difficult it is to even get two identical reading from a glass of water, even with only 0.1 deg C resolution?
I think you can get that much precision, particularly with a bridge as suggested, but there will be many factors kicking around to skew the readings. Your fluid's thermal mass isn't the dominant factor. Set it up to circulate & have say a 5W resistor heating the (well insulated ) pipe. You may be able to get a reasonable calibration this way.

I only dabble occasionally in temperature measurement. There should be some people here who will now their stuff.