| Electronics > Beginners |
| Measuring light spectrum intensity |
| << < (4/4) |
| scatterandfocus:
--- Quote from: MosherIV on October 07, 2019, 06:19:31 pm --- --- Quote ---So how are any of these devices capable of measuring frequencies in the light spectrum? --- End quote --- I do not think the devices measure the freq. They respond to light/em waves between certain freqs. They can measure amount of said light within the freqs. Measuring system use some kind of filter for each specific freq and the device measures the amount of light that gets through. --- End quote --- So I have to wonder how it can be verified which range of the spectrum is being measured. Also, I am now seeing that LED's can be used as light sensors, and UV led's aren't expensive, where UV photodiodes are. But it isn't clear to me how to know which wavelengths a particular LED will be sensitive to when used as a light sensor. But if that can be known, it seems that an LED could be used as an already filtered light sensor. |
| CatalinaWOW:
There are probably other and better ways to know the frequency, but ones I have worked with use a hammer and tongs approach. First identify one frequency. Using a line emission from an ionized gas. The sodium line is a common choice. That sets the offset on the scale. The slope of the scale is built into the instrument and is based purely on geometry. Using the diffraction equation for units using a grating, or the dispersion equation for prism based instruments the relationship between angle and wavelength is known, and the gear train that moves the grating (prism) and the scale on the dial are chosen to match that. That linear relationship between wavelength and angle is why spectrometers are marked in wavelength or wavenumbers. Much easier to calibrate a linear scale than a reciprocal one. Recognize that due to the high frequencies involved even very narrow band filters are hundreds of MHz wide. Its just a scaling problem. A 1kHz bandwidth filter in the SW RF is roughly 0.1 percent bandwith. The same 0.1 percent bandwidth at optical frequencies is about 500 GHz. So your very good and accurate spectrometer might have three or four digits of wavelength accuracy but have dozens or hundreds of GHz frequency error. You might be able to use LEDs as your sensors, but it will be a tough road to follow. They are optimized for emission, not the other way around and so are relatively insensitive. In most cases in spectrometry there is relatively little light available and sensitivity is a prime requirement in the detection chain. It is why photomultipliers were so dominant in the tube era. Even with the extra noise of the multiplication process the gain provided was useful in getting above amplifier noise levels. |
| Navigation |
| Message Index |
| Previous page |