Ultrafast temperature logging using thermocouple

[johnkl]

hi nepu

thanks for the clarification.

we are trying to learn the thermoacoustic property of different materials. for example,a gel is heated by burst microwave (1us every 20ms). we want to know the temperature fluctuation (vibrates between 10-50 degree c). ideally, we want to sample the temperature at 1MHZ(1/1us). do you know any solution?



thanks a lot

[Nepu]

Hi John,

unfortunately, I am not familiar with the physics of gels. I would have some solutions for metals or semiconductors.

If you want to stick with a T/C or other sensors that have physical mass (PT100, etc), make them as small as possible. You might build your own T/C typ K: get some bare Ni and NiCr wires and point-weld them together (This small ball is your sensor. The temperature of the wires does not matter). This gives the smallest mass possible and hence the fastest time response. Get an appropriate Typ K connector for the other end of the wires. Take in mind that the gel should not react chemically with the wires. In addition, the metal wires of your probe (and its sourounding) will be affected by the microwave field. So it is likely that the temperature of the T/C is not the same as the gel would have without the sensor in place.

In any case you should measure the time response of your T/C before interpreting the data of your experiment.

In any case you should be aware that any metal probe (like T/Cs) will change the microwave field around your sensor. Consequently the temperature close to the sensor is changed and heat from far away (100 micrometers) will not reach the T/C surface within a millisecond (rough guess from me. don't know the thermal properties of your gel). A good idea would be to solve the heat diffusion equation to double check.

You won't be faster than about 1 kS/s with a T/C. Remember that the fastets T/C you found had a 3 ms time constant. Especially because gels don't conduct heat too well, the heat in the gel will not spread fast enough into the T/C. In other words, you are for sure not limited by the speed of your digitizer (your soundcard might even do the job in principle) - you might be limited by the T/C (in the kS/s range) - but you are probably further limited to some 100 Hz or even below by the heat spreading in your gel.

For temperature sensing in the MS/s regime you have to look for contactless (hence massless) detection.
Find out which physical property of your gel depends on the temperature and build a sensor for this quantity.
Such quantities might be optical absorption, refractive index, magnetic susceptibility, electric resistivity, etc.
Just take care that such a quantity should only change with temperature and not with any other experimental parameter.
To find the right parameter and to calibrate the detection is a demanding, not so easy project.

Sorry that I do not have a definitie solution for your problem.
Maybe build your own T/C and be happy with some 100 Hz.

Nepu