Why is there no electronic optical pyrometer ?

Discussion in 'Foundry tools and flasks' started by metallab, Jan 3, 2021.

  1. metallab

    metallab Silver

    An optical pyrometer is a device with an incandescent filament observed by the user against the background incandescense color. The voltage can be regulated until the filament becomes invisible and then it has the correct temperature.

    Such a device can also be used through a window of a transparent medium, like glass or quartz, e.g. a furnace peephole.

    But the very convenient infrared pyrometers don't work that way, they measure the temperature of the glass in between. That is why I cannot measure temperature through the peephole of my Kanthal furnace or the window of my wood stove.

    Why is there no device with a small camera in it which senses the color of the target object and compares it to the known RGB values of color temperatures ? Obviously, when the color deviates too much (e.g. green or magenta) it cannot measure.
    Just a color comparison, so when such a hypothetical device is pointed at a tomato (where it is not meant to be used for) would show about 700 C as a tomato color is very close to the color temperature of about 700 C, despite being not the actual temperature.
     
  2. rocco

    rocco Silver

    With recent advances in open source CV, that should be doable, have we got any coding nerds here?
     
  3. Additionally, infrared pyrometers have to be adjusted to the actual emissivity of the surface being measured. Your hand, tomato, and a chrome bumper all have different emissivities so you have to know what you are measuring to get the temperature. similarly, slag, oxides, and liquid metal have different emissivities. The emissivity of the surface may also be dependent on the liquid so even if you get it calibrated for clean liquid aluminum you can get different results with coppers and iron.

    Fortunately visible light appears to be emitted based solely on the temperature, unlike infrared.
     
  4. metallab

    metallab Silver

    Well, I later thought it should be possible to make a smartphone app which uses the camera. This app should detect when the user touches the target on the screen and then determines the color of the target. When close to a 'valid' (i.e. no e.g. green or purple) incandescense color, it can convert the average RGB value of the spot to a temperature.

    Despite having Android Studio installed on my computer, I never succeeded making an app.

    rocco: Indeed that is another issue, although most substances have an emission value around 0.95, but that is not valid for a clean metal surface (solid or liquid). When I look at a crucible filled with cast iron, I see a nice mirror with a mirror image of the exhaust hole in it (which is the clean iron), which have a much lower emissivity with much brighter patches (high emissivity, close to 0.95) which is the slag. For copper it looks the same. Brass however has too much dross.
     

Share This Page