Fast response air temperature measurement

Hi,

I am looking for a small temperature probe with a fast response in still air, ideally 10ms or better.

~0-100 degrees C is fine, accuracy not that critical say +/- 1 degree.

I am not having much success with google etc...

Any suggestions appreciated.

Thanks,

--

John Devereux
Reply to
John Devereux
Loading thread data ...

Do you really want to measure air temperature or something else?

If air, perhaps a very fine Pt RTD or thermocouple.

formatting link

You could also consider IR or fiber optic measurement of something thin and low heat capacity.

formatting link

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
 Click to see the full signature
Reply to
Spehro Pefhany

The only thing that will give you response like that in still air is an acoustic thermometer, which measures the speed of sound in air.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs

Possibly, rare earth phosphor thermometry.

Technically, if you can arrange the optics or use an appropriately thin fiber optic cable (less than 10 micron?) you may be able to achieve that as a physical sensor matter. I routinely do 100ms measurements with rather thicker fibers (200 micron.) Obviously, your source characteristics are also important in deciding just what is good enough and what isn't. Repeatability using phosphors depends on the instrument and method but roughly about 30mK. Accuracy can easily be had to 500mK over that range. (There are at least four methods I know about, each with differing specs to them.) One nice thing is that you can literally _paint_ the surface you want to observe so that even the fiber doesn't touch the surface and affect the temperature. If the paint has minimal mass, it will track VERY FAST. The method is good enough that it can be applied (and is, in fact) to IC wafers and used to monitor their temperatures, in situ. (And you know that IC manufacturers cannot tolerate much variation over the wafer and certainly nothing touching it that might make a cool spot. Lamp heating is used and the opposite side can be used to avoid changing the emissivity of the heated side.) Usable temp ranges from maybe

-200C to maybe +500C for two or three of the methods used. Your range is well inside this and, in fact, is just about the perfect range for it since most everything is 'easy' there. 100 samples per second is not impossible.

Do a search on these methods. You should see at least two companies involved, readily. Fibers can be pretty thin. I've used 10 micron, but I gather down to 2 micron is available.. maybe less. No real impact on the method, so those should work. I stay active in these areas, but not with those companies now. You may contact me privately, if you want. I can say more, that way.

Jon

Reply to
Jon Kirwan

l

The other technique that comes to mind would be to use infra-red or Raman spectroscopy to probe the population of the rotational excited states.

formatting link
ETRY=3D0

It might be sensitive enough to resolve 1 degree Celcius/Kelvin though there's always the risk that you might have to send so many photons through your sample that you'd warm it up enough to perturb the temperature.

-- Bill Sloman, Nijmegen

Reply to
Bill Sloman

Have you looked at Ir types?

Reply to
Jamie

Varies with wavelength, pressure, gas constituents (altitude affects H20 significantly), temperature, ... well, it varies.

Not to mention that 0-100C is where most everything around is also emitting.

Jon

Reply to
Jon Kirwan

What's the emissivity of air?

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
 Click to see the full signature
Reply to
Spehro Pefhany

Hi Spehro,

Yes, it is for actual air. The air will likely not be *completely* still, but I don't really know what movements there will be.

formatting link

That is the sort of thing I have been looking for, but all I could find commercially (RS and Farnell etc) were bead thermocouples/thermistors with thermal time constants in the ~1s region.

That would be hard due to the location, I really need a probe-style device of some sort.

Thanks,

--

John Devereux
Reply to
John Devereux

Thermal conduction is amazingly slow, but does speed up quadratically as you reduce the size of the object. If it's thermal-mass limited, it speeds up more slowly since the conduction goes as surface area but mass goes as volume, so the time constant goes as the length.

The thermal conductivity of air is very small, only about 0.025 W/m/k, and its volume heat capacity is also very small, so still air is a very unpromising heat transfer medium. Doing the actual calculations and seeing how very slow and inaccurate any solid-material heat sensor is in still air is very educational, and not at all difficult.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
 Click to see the full signature
Reply to
Phil Hobbs

formatting link

You might want to go trolling around for hot wire anemometry probes, and run them at a low enough current that they don't heat up too much.

They can have surprisingly fast response under wind tunnel conditions, so they might be fast enough for your application with air almost still.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
 Click to see the full signature
Reply to
Spehro Pefhany

But, still air doesn't have a single well-defined temperature if there's any fast temperature change going on, UNLESS it's a matter of adiabatic heating due to compression. So, use a fast barometer and any old slow temperature probe.

Reply to
whit3rd

I'm there. I just failed to notice this was about measuring 'air' when I posted. I've never done this, but temperature dependent fluorescent properties are probably found in some gases and aerosols. The oxygen in the air may quench it, though. Air is a tough target for those rates.

I suppose a good thing is that it is still air. If it had a velocity, there'd be a need a way to calibrate out that effect on most sensors that were not moving with it at the same speed and direction.

Yes, I gathered that just after I posted. My mistake. It may yet be possible in the situation. But we just don't know enough details (or I don't, at the moment of writing this.)

Jon

Reply to
Jon Kirwan

If it wasn't clear, your point about heat transfer in still air and mine regarding net velocity against the sensor provide a dilemma. To get better transfer, circulate lots of air over the sensor implying relative velocities; but to get accuracy, don't do that as the velocity cannot be discerned vs the temperature by some sensor which is affected by the statistics of molecular/atomic bombardment. I remember that this presented a problem in measuring exhaust gases, some time back, trying to discern which part was which part in what the sensor 'observed.'

Jon

Reply to
Jon Kirwan

Hi,

Yes, one of the reasons for the fast temperature change is due to a pressure change, which I am already measuring. Other reasons are the temperature of the enclosure and of the incoming air.

Perhaps I can use the pressure changes to "predict" the temperature changes as I think you are suggesting. I still need a fairly fast single temperature measurement, less than 1 second. (But better than 10ms in this scenario).

--

John Devereux
Reply to
John Devereux

I'm pretty sure that the time constant will change with air speed -- dunno, but it's something to check before you commit.

--
www.wescottdesign.com
Reply to
Tim Wescott

Perhaps a dumb idea, but since thermistors and such have a log time response, can you take multiple measurements and curve fit to get the temperature faster, though with less accuracy, at least until things settle?

Reply to
miso

Sure, good idea, that might help.

Thanks,

--

John Devereux
Reply to
John Devereux

l

This means differentiating the output of the thermistor, and using the rate of change of the temperature of the thermistor to tell you how much hotter or colder the thermistor is than the surrounding air.

It requires a very low noise amplifier, because the the gain of the differentiator rises in proportion to frequency, as does the the Johnson noise generated by the thermister and the input stage of the amplifier.

If the air is still enough that heat transfer rates conform to Newton's law of cooling - proportional to delta-T squared - life gets really complicated, but there is usually enough circulation that self- excited convection isn't significant.

-- Bill Sloman, Nijmegen

Reply to
Bill Sloman

l

This sounds like a promising idea. I've used an acoustic resonator to measure the He3/He4 ratio in sample gas. We had to stabilze the temperature to get good readings.

As pointed out in later posts, even with a really small detector you still have to get the heat from the air and into the detector. Perhaps better to measure the air directly.

George H.

Reply to
George Herold

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.