A number of factors need to be taken into account when considering sources for calibrating radiation thermometers:

• Firstly, the calibration source needs to have a high and accurately known emissivity, to ensure that measured temperature will accurately reflect the true temperature of the surface.
• Secondly, the source needs to be large enough to fill the optical field-of-view of the radiation thermometer, since under-filling the field-of-view will result in measurement errors.
• Thirdly, the temperature of the source needs to be measured by some means, for example by using a contact sensor inserted close to the radiating surface. A blackbody source with an aperture of a suitable size meets these requirements and should therefore usually be used to calibrate radiation thermometers.

The use of an anodised aluminium plate is not generally recommended for checking or calibrating radiation thermometers. Firstly, the emissivity of anodised aluminium is quite low and depends on the thickness of the anodised layer. Many radiation thermometers operate at wavelengths in the infrared, and in this region the emissivity can be anywhere between 0.9 and 0.4. The actual value must be known if measurement errors are to be avoided. Secondly, because the emissivity is low, the radiation measured by the thermometer will be a combination of radiation emitted by the plate and radiation reflected from the plate from other objects in the room. This will, again, lead to potentially significant errors in the reading. Thirdly, there must be some other means, such as a contact probe, for determining the temperature of the plate. If this probe is not in good thermal contact with the plate, or if there are temperature gradients across the plate, it will not give a good indication of the true plate temperature. It will therefore not be possible to accurately check the calibration of the radiation thermometer.

Tungsten ribbon lamps can be used for calibrating some types of radiation thermometer. However, they suffer from a number of potential drawbacks. Firstly, the filament is very narrow: typically 1.5 mm, although lamps with 3 mm wide filaments are also available. This is generally too small to fill the field-of-view of most modern, commercial radiation thermometers and will lead to significant measurement errors. In addition, the emissivity of the tungsten filament is low (about 0.4), and varies with both temperature and wavelength, and the transmission of the glass envelope also varies with wavelength. Since the lamp is calibrated in terms of radiance (apparent, or brightness) temperature at a precise wavelength (usually around 650 nm - 660 nm) these effects can be neglected provided that it is only used to calibrate radiation thermometers that operate at that wavelength. However, radiation thermometers which do not operate over such a narrow and precise wavelength band cannot be calibrated using lamps.