Measuring water temperature with 0.001 deg resolution

[shantanu@india]

Hi,
Thinking of building a thermodynamic pump efficiency measuring device where temperature and pressure needs to be measured across the pump.
https://www.robertson.technology/thermodynamic-method
My question is can I use a standard Pt100/K type thermocouple and use a 16/24 bit ADC to get the required resolution or do I need some special kind of temperature sensors?
Any pointers would be appreciated.

[TimB]

Personally I think unless you have many years of experience in very high precision temperature measurement you would be deluding yourself to think you can get near that precision.

I would recommend you use the LTC2986 and an RTD sensor. It will pretend to have the resolution but in reality you will never prove it.

To get it calibrated to that resolution will cost more than the board cost to develop and build.

Here is a good article to read on it.

https://andybrown.me.uk/2017/09/02/temperature/

[GDeSantis]

You might want to check out this link:

https://www.eevblog.com/forum/beginners/measuring-temp-with-0-01c-resolution/

[shantanu@india]

Thanks Tim, GdeSantis.
Let me start educating myself first.
Yes Tim, I know calibration would be big issue.

[shantanu@india]

Maybe the key really lies in not attempting to measure the absolute temperature but the temperature differential. Thermodynamic efficiency measurement only needs the temperature differential. Temperature calibration also may not be required... you need to compare the compare the thermodynamically derived pump efficiency with the conventionally derived value.

[TimB]

Getting to 2 probes to match to 0.1 is hard. 0.01 is really hard 0.001 nigh on impossible.

Look into using triple point references. Water is cheapest. The issue is that just moving the probe a mm away from the other probe is enough to give a 0.2c difference.

Here is a really quick video note the instructions at the end.

[JonW]

Have you done the math to see what accuracy you need to acheive your targets?  Accuracy and resolution are completely different and am with Tim on this 0.01 accuracy is really hard to meet over time.

[Fanie]

I agree with Tom, such fine measurements is difficult to achive, if at all.
I have a friend who wanted to build himself a milling machine that would measure/mill in angstrom's since mm was too course for his liking.  Reminds me of that.
I'm curious though Shantanu, do you also wear the white lab overcoat ? 
Usually using a bridge setup is supposed to offer the finest measurements, but physically it may not be possible.

[Fanie]

I've been thinking (dangerous  )
If one assume the temperature to be measured is within the range of thermistors, you can use such a bridge.
You said water, which will be between roughly 0oC, below will ice and beyond 100oC will become steam unless the water is pressurised of course.
The bridge should be from thermistors with an accurate tolerance so the values would be as close as possible.
You will have to add a low value multi-turn pot in one leg to balance the bridge for 0.000V output while all are at the same regulated temperature.
The 3 thermistors will have to remain in the temperature controlled enclosure when measuring, because the thermistors are not linear, keeping them all at a specific temperature should give a 0.000V reference temperature value.
The measuring thermistor's curve should be plotted as accurate as possible so you can compensate for non linearity during the measuring.
Use wiring that will have as little effect as possible when exposed to temperature changes.
You should end up with a repeatable very accurate measuring setup, but could still end up with a temperature with a certain error offset (although minute) because I doubt there is such an accurate reference temperature available.

Measuring water is another problem, since warm water rises up and cold water flow down, there is always turbulences which could influence the measuring at such accuracy, and will depend al lot on the water volume and the position in the container.  One can probably only hope to achieve an average over many measurements.

There is something else about water and that is it changes it's molular structure when exposed to turbulance, vibrations and even when exposed to imagery.
Here is one of many that indicate this
https://mywaterfilter.com.au/blogs/learning/how-water-responds-to-thoughts-and-emotions#:~:text=When%20we%20consume%20water%20with,of%20stress%20and%20dis-ease.

Water is actually a fascinating topic...