self-mixing-equations.ipynb

self-mixing-equations.ipynb

Hi Aleksas,

Long time. I just wanted to inform you that I reworked my interferometer project in my free time as I needed to measure Piezo displacement between 40kHz and 300kHz and at low amplitudes less than 6 micrometers. I noticed there were quite some errors in my past project. Please see changes to the self mixing interferometer:

1. The updated laser diode is the ADL-65052TL from http://laserlands.net
2. The controller was switched to the Analog discovery 2 (AD2).
3. I moved to use a simpler modulation scheme using an Opamp and N-channel MOSFET to create a Voltage Controlled Current Source (VCCS). The scheme is briefly illustrated by Norgia here on IEEE Xplore: https://ieeexplore.ieee.org/document/10155245 - Have a look at section V.
4. The photodiode section is quite standard needing only a trans-impedance amplifier.

Some learning points:

1. The self-mixing setup is limited by the TIA Gain Bandwidth Product (GBW), you typically need 4MHz GBW Opamp to visualize fringes for a 500Hz motion due to the higher frequency components in the interferometric fringes. My rule of thumb us that you need a TIA bandwidth of 10-20 x the sinusoidal motion frequency of the object.
2. You require a retroreflector to enable backscattering of as much power into the laser cavity - However, some papers show that self mixing is possible on diffuse targets but in my opinion the signal is below the noise floor. However, I notice general noise from the breadboard, speckle noise from laser and fading effects from the retroreflector show up in my signal. These effects limits the frequency of motion I can detect by raising the noise floor, and therefore not suitable to a KHz piezo.

Currently I am using the S-curve of a Sanyo HD-65 DVD optical pickup. This is based off this work: https://www.youtube.com/watch?v=5bqujaldaCQ that shows that a DVD optical pickup can serve as a MHz level vibrometer - See minute 2:48 in the video. However, the vibrometer FFT off of the pickup is in volts instead of absolute distance.

As FYI, other links showing hacking Optical Pickups is here: http://www.diyouware.com/node/161 and here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6066758/ .

To calibrate the S-curve, I noticed that the optical pickup also can act as a self-mixing interferometer, and therefore I can calibrate the intensity measured by the pickup photodiode using the monitor diode signal at a low frequency of motion. This would then allow usage till MHz frequencies. I believe this is possible because this research mentions that noise is caused due to back-reflection of power back into the laser cavity (See figure 15(OPUStructure) and figure 18 (Backreflection noise) here https://research-information.bris.ac.uk/files/319227179/Final_PhD_Thesis_FRP.pdf) - Therefore it could work as a self-mixing interferometer also - In principle,

Happy to collaborate more on this research.

I have stopped working on the project 3 years ago due to time shortage. I have managed to reproduce interference steps both with retroreflective tape and from just a white colored wall from ~30-40cm distance. A weaker interference pattern but still very clear visible. If I remember correctly distance estimation was somewhat consistent with theory, but there was a big error margin when calculating using peak count. I have a repo with pcb design and software if you're interested.

Thank you for the feedback and insight on your results. Do share