I have just begun setting up an interferometer to take some measurements. I am trying to reproduce one paper I found in literature for 3 DoF homodyne interferometry. The detector is merly a CMOS sensor. Here is a picture of the paper I am trying to reproduce: from the spatial interference pattern, these guys managed to extract information about relative tilts and displacements between the mirrors applying FFT algorithms to each pixel line. enter image description here

Applying FFT to their picture, I got something similar like that (for only one [pixel line), and it seems to agree fairly well apart from the scale. However, when I try to do the same to a fringe pattern I got from my setup I obtain this: enter image description here

enter image description here

The intensity is not sinusoidal adn there are plateaus where the maximum intensity is detected: is there a way to get more spacing, or smoother fringes?


1 Answer 1


Your problem is in part due to the small dynamic range, the ratio between the maximum and minimum measurable light intensities, of your detector.
This is a problem often encountered whenever one tries to "photograph" fringe patterns and you can usually tell if an image is real or simulated by looking to see if there are overexposed fringes or not.

Reducing the intensity of the fringes which are being analysed by your detector sufficiently should enable you to analyse the "expected" fringe intensity pattern.

  • $\begingroup$ Thanks! The dynamic range of the sensor is up to 69 dB and the minimum exposure time is 0.040 ms. If I understood right, to visualise fringes properly I should either decrease the exposure time (it was already set to the minimum) or increase the dynamic range? This would mean change the detector! Could a higher frame rate help in that sense? $\endgroup$
    – Gianluca
    Commented Feb 4, 2021 at 12:23
  • $\begingroup$ Could you use a neutral filter to reduce the intensity? $\endgroup$
    – Farcher
    Commented Feb 4, 2021 at 14:33
  • $\begingroup$ You are probably right. As I am pretty new to optics, I cannot even imagine the existance of some components, or how to concisely describe any issue! I have now a couple of questions: - can a 0.8 mW laser damage the CMOS sensor? Apparently I already have some dark spots...; - is there a way to understand which neutral density filter I should need for my application? $\endgroup$
    – Gianluca
    Commented Feb 4, 2021 at 14:51
  • 1
    $\begingroup$ Like Farcher said, just decrease the power hitting your detector if all the electronics are already set to their minimum sensitivity. Farcher should have mentioned that you are saturating your detector (going above its maximum capability), hence you do not obtain information. If I dump 12.367 L of water to you and you only have a 5L bucket, all you can measure is 5L and how much I exactly gave you is lost. So, decrease the signal going into your detector until you see a smooth image and not an overblown/saturated one. $\endgroup$ Commented Feb 4, 2021 at 14:51
  • $\begingroup$ I would not think that a 0.8mW beam could damage a CMOS detector (unless if it is ultrashort or very very tightly focused). Are you sure that the dark spots are not just small dust particles on the sensor/optics? $\endgroup$ Commented Feb 4, 2021 at 14:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.