Ultra small temperature sensor

[_maxim_]

I need to read out the temperature of ZrO2 devices which rotate along their axis between 10-100 kHz by using dry compressed air.
The external diameters of each device are: 1mm, 3mm or 4mm, and they rotate inside a special stator

The actual rotation speed is read by an optical fiber close to the bottom where a black mark is placed.
Up to now the temperature was read out by using a PT100 thermocouple (Cu-Costantan) placed somewhere close to the top of the device (exhaust air).

I want to realize a better measuring circuit: it should read directly above to the surface of the spinning device (which is white colored), should be well miniaturized and precise.

What is the state-of-art today? A laser ray? An InfraRed sensor? Or what else?

Thanks for any advice.
Max

 

[_maxim_]

Up to now the temperature was read out by using a PT100 thermocouple (Cu-Costantan) placed somewhere close to the top of the device (exhaust air).

Of course is NOT a PT100, is a thermocouple T-type (Cu-Co) as written.
I apologize for the mistake.

 

[dlgoff]

http://www.omega.com/ppt/pptsc.asp?ref=IRCO_CHAL_P13R_P10R&ttID=IRCO_CHAL_P13R_P10R&Nav=

Their sizes are similar to this Unsheathed Fine Gage Tungsten-Rhenium Microtemp Thermocouple:

 

[0xDEADBEEF]

It also depends on the temperature range. If the temperatures are not too low and the measurement speed is not crucial PT100s are good. They can be small and reasonably fast, and parts of the ITS-90 temperature scale are actually defined with PT100 thermometers, so you would actually be using the definition of temperature for your measurement.

 

[_maxim_]

Thanks.
The range should be between 260K to 330K. The point is that the temperature should be read on the rotor surface and not from the air flow. That's why I thought to an IR / laser approach.
The present in-air thermocouple is not capable to read out efficiently when the speed rate becomes (i.e. air flow -> friction) large.

 

[0xDEADBEEF]

Well there would be this: http://www.zdnet.com/blog/emergingt...emperature-sensor-opens-up-possibilities/3137 but the temperature accuracy of 1C is very poor, and you don't know how much air you measure. The only way to measure it with a laser would be that some property changes with temperature like a line width of some die that is applied to the rotor. Or some elongation, but with something spinning it is not that easy to handle either. I suppose that it is some kind of NMR setup? Because a common way to measure stuff like this is adding a substance with a known strong temperature dependence to the sample. I don't know enough about NMR, but if your measurement would not be magnetic one could use something like the susceptibility of a piece of Au[itex]_2[/itex]MnAl which seems to have a Curie temperature that is low enough.

 

[_maxim_]

Well there would be this: http://www.zdnet.com/blog/emergingt...emperature-sensor-opens-up-possibilities/3137 but the temperature accuracy of 1C is very poor, and you don't know how much air you measure. The only way to measure it with a laser would be that some property changes with temperature like a line width of some die that is applied to the rotor. Or some elongation, but with something spinning it is not that easy to handle either. I suppose that it is some kind of NMR setup? Because a common way to measure stuff like this is adding a substance with a known strong temperature dependence to the sample. I don't know enough about NMR, but if your measurement would not be magnetic one could use something like the susceptibility of a piece of Au[itex]_2[/itex]MnAl which seems to have a Curie temperature that is low enough.

Hi 0xDEADBEEF,

thanks: TMP006 from TI is a very interesting device. Due to its small area, maybe it can be inserted into the stator and close to (but not touching) the rotor. The accuracy stated as 1C is poor but not dramatic.
The job is for improving the reading temperature in Solid State NMR under UltraFast condition.
As said before, the thermocouple is no longer useful at high speed rates due to air friction, while a laser or IR approach is more promising from this point of view.
The "real" temperature of the rotor is important for biological contents which suffer at temperatures less than 0 or bigger than 25C.
Adding paramagnetic or diamagnetic drugs is almost impossible (or at least very problematic) for crystallized proteins, if one is interested to the structure determination.
And, of course, one needs a probe/RF capable of doing Au₂MnAl spectroscopy

 

[Mike_In_Plano]

I've worked with various temperature sensing schemes, including taking IR down to .1C. Getting IR to focus on a target as small as yours and be accurate below 1C is problematic. Generally you need germanium optics for something so small. To get the added accuracy, you can control the temperature of the sensor such that it approaches that of your target, or you can use a shutter with and RTD.

I don't know what off the off-the-shelf opportunities look like, but in the distant past I worked with some guys on a product where they doped the end of fiber optic cables such that they would fluoresce. Then they excited the fiber and watched the time decay response to ascertain the temperature. I'm pretty sure they held .1C and their product shipped.

It occurs to me that you might find such a material and have it sputtered on. Then you would need an optical filter over your photo-diode (something that doesn't fluoresce), a good trans-impedance amplifier, acquisition, and an algorithm to pull out the time constant.

 

[_maxim_]

Thank you MIP for having shared your experience!

About Ge-RTD from Lakeshore, AA type have a very compact design while unfortunately the range of use is limited to 100K. As said above, I need to keep the rotor spinning between 273-313 K, does not matter the accuracy is below 1C: that value could be considered satisfactory.

About doped fiber optics, are they available for sale?
Maybe the the doping process is rather easy to implement, even if I would prefer to use a commercial device to be operative as soon as possible.

On the other hand, if you can kindly share with me more details about the following points

you might find such a material and have it sputtered on. Then you would need an optical filter over your photo-diode (something that doesn't fluoresce), a good trans-impedance amplifier, acquisition, and an algorithm to pull out the time constant

it would bereally great!

Have a nice w/e
maxim

 

[M Quack]

Take a look at this thesis, they built a similar high-speed rotation device (pp 76).

http://www.esrf.eu/UsersAndScience/Experiments/DynExtrCond/ID18/publications/theses/Thomas_Roth_thesis/troth_thesis.pdf

 

[_maxim_]

Thanks MQ,

they have used a PT100 resistor inserted in the gas pipeline, but it is clear that the sample temperature differ from that revealed for almost 5K or more...

regards