Physics 408: Optics Laboratory
Michelson Lab survival guide
- Once you have found the white light fringes, immediately take the Fourier Transform spectroscopy data, check that it is okay and
then move on to the calibration proceedure.
- Due to a software problem in the APT-USER program that controls the micrometer stage, loading the "move sequence for FT-Spectroscopy" does NOT work! In fact, it bugs in a bad way the APT-USER program, so don't use it. You will therefore need to code the move sequence for the FT-Spectroscopy section yourself. Once you have the white light fringes code the following move sequence. If you have already loaded the
"move sequence for FT-Spectroscopy", kill the APT-USER program and restart it since this code somehow disables the move sequencer. Don't worry, when you relaunch APT-USER it remembers the position of the stage.
Now that you have coded this move sequence, check that when it is run, the full move sequence executes. This code should move the stage back by 30 um, then forward by 60 um, then back again by 30 um quickly.
Move No. Dist/Pos MinVel Acc MaxVel DwellTime Return Absolute Relative
1 -0.03 0 0.005 0.005 1 False False True
1 0.06 0 0.005 0.005 1 False False True
1 -0.03 0 0.005 0.005 1 False False True
Once you try this code and it all *works*, now you are ready to
set the MaxVel for the 60 um forward moving step to a slow value (choose 0.0003 which will give 300 nm/s and with a frame rate of 30 frames/sec gives 10 nm between frames) so you can take the data with MatLab.
Also, you will need to set the DwellTimes to a few seconds. This code looks like:
Move No. Dist/Pos MinVel Acc MaxVel DwellTime Return Absolute Relative
1 -0.03 0 0.005 0.005 2000 False False True
1 0.06 0 0.005 0.0003 1000 False False True
1 -0.03 0 0.005 0.005 1000 False False True
At this point you are ready to take all the FT data. Be aware that filter "C" is a deep blue filter and doesn't have the best transmission since our white light source doesn't have a great output below 450 nm.
Check carefully the "FT-Spectroscopy.m" file since it may need to be corrected
The old version of this code (still floating around) is off
by two bins, which biases the spectrum low by more than 5 nm at 640 nm
(so the laser calibration step will be off by around 7nm).
The problem is that the line
fftdata_oneside =fftdata(NFFT/2:NFFT)
includes a negative frequency bin (N/2) and the 0 frequency bin (N/2 +
1). But when the wavelengths are calculated, the index vector 1:N/2+1 is
used, which starts at the first non-zero k (which should correspond to
fftdata(N/2+2).
So the quick fix is to change that line to
fftdata_oneside = fftdata(NFFT/2+2:NFFT)
and the n = 1:1:NFFT/2+1 line to
n = 1:NFFT/2-1
- correction by Matthew Hasselfield: Oct. 30, 2008