Measuring temperature of a certain body without contact

Temperature measurement

is there a way to measure the temperature of a certain body from without contact
im trying to make a robot that measure temperature of certain objects and tries to follow a certain temperatue
any help?

Re: Temperature measurement

There are many "IR thermometers" on the market. They are usually hand held units with a LCD display. Fluke at least at one time made one that plugged into a multimeter. This type could be adapted to your use. These units have different acceptance angles. You could if necessary use a lens or parabolic reflector (shaving mirror) arrangement to narrow the beam width.

Temperature measurement

i need some more details about this subject does any one have some schematics about this sujbect and some useful links

Re: Temperature measurement

do a google search on bolometric sensor...
h**p://w*w.ulis-ir.com/default.html

hope this help,
//a

Re: Temperature measurement

ZeleC said:

i need some more details about this subject does any one have some schematics about this sujbect and some useful links

Click to expand...

Every material at temperature above absilute zero emmit elctromagnetic radiation with a spectral density described by the Planck radiation law for a Black Body:

r(L,T) = C1/((exp((C2/L*T))-1)*L^5) [W/(cm^2*uM)]

where L is the wavelength in [uM], the T is the object temperature in [K],
C1 = 3.74E4 [W*uM/cm^2] and C2 = 1.438E4[uM*K].

The maxima of the spectral curve happens at C2/(L*T) = 4.965, or for room temperature of 300K the maxima of the spectral curve happens at (2896/300) = 9.65 uM. This is why the best measurements of the room temperature ranges are made in the mid infrared wavelength range, and all the passive motion allarm systems implement infrared pyroelectric sensors.

The cpectral density at the peak of the curve is

r(Lmax,T) = 1.315*(T/1000)^5 [W/(cm^2*uM)]

For 300K body the spectral density at peak of the curve is 3.2 mW/(cm^2*uM)

If some measures the radiation over the complete spectral range, then the total emmited power is

R(T) = s*T^4,

where s = 5.67E-12 [W/(cm^2*K^4)] is the Stefan-Bolzman constant.

For instance a 300K Black Body will emmit appr 46 mW from each cm^2 of it's surface, and the derivative vs T change is appr 600 uWatt/K.

The real materials aren't completly black, they don't absorb and don't emmit like the abstract Black Body.

Each real object have it's oun characteristic emmisivity curve, where the metals tend to have low emmisivity (high reflectivity), while the various isolators tend to have high emmisivity. The water surface emmits almost like a BB, so do the human body. There are other quirks like the transparency of the material whose temperature U'd like to measure.

So in general some have to know the composition of the material in order to estimate it's temperature by the measured thermal emission. There are methods to minimize the error of temperature measurement by measuring the complete emission spectral curve and by taking ratios of measured radiation in two or more spectral bands. There also are methods to estimate the composition of the measured material by limiting the influence of it's temperature, but that's another topik.

What U need to get started is a simple cheap motion sensor.

There are two catches: first of all, the used IR sensor is of pyroelectric bolometer type and it produces signal only while the teperature of the bolometer changes, hence some have to use optical choper in order to provide the necessary thermal derivative, but then the bolometer measures the thermal diference between the object and the chopper, so the chopper temperature have to be measured and used as reference.

Second, the motion sensors actually contain two bolometers connected in a way the common radiation changes to be nulled, so the detector senses only the derivative of the difference between those two., so U have to tape half of the sensor window with aluminium tape to leave just one of the bolometers active.

Naturally U'll need some optics to define the field of view of the sensor. The motion detector contains an aray of Frenel lences which produce a funy multy-lobe pattern. U may consider to use only one central of those with penalty of reduced collection area (less energy accepted by the system).