Spectral response of the DIY Spectrometer & why I can’t detect red

The following is a series of tests performed tonight of the spectral response of the “DIY Spectrometer”.


The “DIY Spectrometer”:

Feed in to the DIY Spectrometer:

No telescope tonight – direct feed into the end of the device below that I made below using a beam splitter and modified Cheshire eyepiece.

Spectral Response:

Spectrum of Compact Fluorescent light bulb through above device attached to DIY Spectrometer (without a telescope) with and without the illuminated eyepiece on the above device switched on (see below).

Spectrum of Compact Fluorescent light bulb WITHOUT telescope and with illuminated eyepiece turned OFF (below):

Spectrum of Compact Fluorescent light bulb WITHOUT telescope and with illuminated eyepiece turned ON (below):

The above show that the fiber optic cable does not pick up the illumination from the red LED on the illuminated eyepiece.

I initially thought that this might be due to mis-alignment between to fiber optic cable and the LED light from the illuminator.

I tried covering the telescope aperture on the device and pointing the guiding aperture on the device at the compact fluorescent bulb and could replicate the spectrum above – hence light through the guiding aperture DOES get to the fiber optic cable.

Why I am not picking up the red LED light from the illuminator on the illuminated eyepiece in the guiding port on my ATM device:

I then worked out the reason for my inability to pick up the red LED.

The following spectrum was seen when I pointed the device at the blue LED on my printer (with all other lights switched off in room):

Notice the peak at about 465 pixels, consistent with detection of the blue LED.

The spectra above have not been calibrated. However, the following spectrum of a Compact Fluorescent Bulb was taken by myself using the same instrument on 9/8/2018 and I calibrated it at the time (see result of calibration below). This shows that the instrument’s response, with its diffraction grating as currently positioned, is 323-612nm.

The relevance of this can be seen below on the table of wavelengths of visible light from Wikipedia (https://en.wikipedia.org/wiki/Color, accessed 23/8/2018):

The colours of the visible light spectrum[5]
Colour Wavelength
Red ~ 700–635 nm ~ 430–480 THz
Orange ~ 635–590 nm ~ 480–510 THz
Yellow ~ 590–560 nm ~ 510–540 THz
Green ~ 560–520 nm ~ 540–580 THz
Cyan ~ 520–490 nm ~ 580–610 THz
Blue ~ 490–450 nm ~ 610–670 THz
Violet or Purple ~ 450–400 nm ~ 670–750 THz

The wavelength of red light is outside the wavelength response of my DIY Spectrometer’s diffraction grating as I currently have it set.

This is a potential advantage is it means the illuminated eyepiece will not interfere with the spectra produced in use. However, I may need to change the setting of the diffraction grating (not an easy task) as main colour in nebulae is red and infra-red. A possible job for the future…..

In addition, being able to pick up the LED in the spectrometer would allow me to align illuminated cross hairs to fiber optic cable by adjusting one or other’s position in my device until maximum intensity in red line achieved – which I can’t do if the spectrometer does not detect red colour,

Spectrum from DIY Spectrometer when pointed at white computer screen:

This shows two peaks (blue and green) – remember red missing as outside the range covered by the spectrometer currently.

Residual spectral response when all light excluded from DIY Spectrometer:

I have not been able to work out what this peak at 460 pixels is – it persists even when light excluded but is swamped when any reasonable source light enters the spectrometer – but does not completely disappear even then.

Any one got any ideas?

6 Responses

  1. Can you average the signal over several minutes and produce a ‘dark’ frame to subtract and hide the hot pixel (and any others?)

    I don’t like all this talk of spectrometers, it’s making my wallet nervous…

    1. Oh I see what you mean – when I take a spectrum of a bright object the signal at 460 pixels virtually disappears – swamped – so it is only a low level signal and your explanation is best I have come across so far so thanks for that.

  2. This probably the equivalent to a “Hot” pixel on a DSLR screen, except in this case its a hot pixel on the spectrometers sensor.

    Pete H

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.