The Pulsar Engineering Community Board

With this post, I Iike to provide some guidance for others like me who have very little optical experience. If I am correct, my setup is also the first with a Daheng camera (incl. Sony IMX265).

My situation : The purchased hardware corresponds with the performance edition with solid add-on (2022) with the exception of a Daheng camera (MER2-302-37GM-P). The used camera software is called GalaxyView instead of Spinakker for FLIR cameras. Automatic spectrum extraction was not possible yet but Luc’s Matlab code was already helpful in extracting the paracetamol spectrum. I did not have accessories such as a lens tube, fiber or stage frame at my disposal, so I had to be a little creative and I am very thankful to Luc for his support.

These are my experiences before I retrieved my first spectrum:

• Alignment (1) : The initial mechanical alignment of the lens group I did by elongating the bottom steel bars until they could reach the camera (and later the powder cell fixator). I used the powder-cell-bars for that. Around the camera lens I fixated a card board ‘cage plate’ and two CP33B’s for sturdiness. This way, all lenses were aligned just fine without additional accessories. In case you want to do this as in the Youtube tutorial, I think you’ll need the additional CP36 cage plate, SM1L05 lens tube and SM1A36 adapter.

• Alignment (2) : With the sample holder in place, I was able to manually align the green laser spot in the aluminum cup. However, as it turned out later, the reflected beam didn’t point to the slit. So, I purchased the long rods (ER8-P4), 2 frosted disks (DG10-1500-H1-MD) and 2 threaded cage plates (CP33/M) to align the incoming beam. By putting a simple mirror behind the last disk, I could aim the outgoing beam for the center of the slit.

• Alignment (3) : After removing the disks and installing the solid add-on, I adjusted its tilted mirror (BB1-E02) for centering the laser spot in the sample holder with a frosted disk below it. Again, when you take away the sample holder tray and flush mount a simple mirror below it, the reflected beam can be visibly directed towards the slit to confirm all your tuning as you go (without the edge filter installed of course).

• Alignment (4) : I found out that, even with the bright white paracetamol sample in place, the reflected beam spot was visible on the slit (even for 5 mW). This was a useful confirmation, as the paracetamol’s large variation in Raman intensity made me fly blindly software wise and made me doubt all my mechanical tuning so far. But, as I could visually verify that my alignment was OK, I focused successfully on the powder positioning (@250ms,24dB) and the laser power (30mW).

• Connecting (1) : When I tested the neon lamp with the GalaxyView software, I first had to switch the X-axis direction in the software. This means putting the ImageFormatcontrol\ReverseX parameter to true. In the same ‘ImageFormatControl’ tab, you might also want to change the ‘PixelFormat’ to ‘mono 12’ (i.e. ‘Bpp16’) before you press the camera’s play button. Afterwards, you read the wavelength as the cursor’s X-coordinate in the bottom line of the screen.

• Connecting (2) : The only two other tabs I needed were ‘AcquisitionControl’ (for the ‘ExposureTime’ and ‘AAROIOffset’) and ‘AnalogControl’ (for the ‘Gain’, FYI: max. 24dB).

• Calibrating (1) : Because the solid add-on already contains an achromatic doublet (AC127-019-A), it is sufficient to put the neon or fluor lamp below the powder cell (without sample holder in place). This allowed me to put a light beam on the slit (visible in a dark room). So, a sufficient alignment procedure, with the rotation and rectification of the spectra, occurs completely similar to the bread board instructions without the need for fibers or stage frames.

• Calibrating (2) : I tried an old and fresh fluor compact lamp as described on the webpage but I did not find the demo spectrum. Luc confirmed that it might be due to his lamp’s nature as also other people have found other spectra. Regardless, the pixel trim job can also be done neatly with a cheap white led.

• Calibrating (3) : In absence of the fluor spectrum, the calibration procedure would become something like this :
  • Rotate the grating with the bottom knob so that the neon spectrum fills the width of your screen as much as possible (i.e. aim for 2048 pixels).
  • In my case, the peak @703nm crossed the right border of the screen and the peak @585nm almost touched the most left border, let’s say its center is at pixel 11.
  • Compute the dispersion. In my case (703-585)/(2048-11) = 118/2037= 0.05793 nm/px.
  • Determine at which pixel position you need to put the first peak so that pixel ‘0’ is at 540 nm. In my case, 585-540 = 45nm. So, at 45/0.05793 = 776pixels I should put my most left selected peak (i.e. @585nm). This is close to the 779px on the breadboard web page.
    (FYI, my @703 peak was far less pronounced than my @692 peak, which is a little odd but especially confusing.)
  • You can then use a white LED to adjust the FEL550 edge filter such that you can see it clipping the first few pixels.

Thank you very much for your detailed feedback Karel! I liked the idea of using the solid cuvette to re-image the light on the slit; that's less tooling required

I think you also mentioned me that your Daheng camera is limited in 1 sec exposure time. It's the same with mine (different sensor, it's a MER2-160-75GM-P). This is a bit annoying because I believe it's going to limit the dynamic range of the spectrometer by a factor ~6 (sqrt(32 sec/1 sec)) compared to the FLIR camera and assuming they have the same readout noise. I'll experiment around that but it's worth to be taken into account for other people if they'd like to switch to the Daheng camera as well.

(Spoiler alert: an update will come for all of those who can't put their hands on a FLIR camera -- i'm pretty busy atm so I'd expect to start working on that this summer during the holidays)

If this thread is still open, may I ask a few questions about the alingment?
I have noticed that there was no light signal from the cuvette reaching my camera. I used spinview and set the gain to the highest value while shielding off stray light with a blanket in the hopes of finding some signal comming from the sample. Untill now I got nothing. I have tried one of Karels tricks of inserting a mirror behind frosted glass disc 2 and guiding the laserdot back to the slit. After removing the edge filter and reajusting the gratting I saw the spectrum of the laser on the camera. I was now certain, that the assembly was well aligned. Then I reassembled and recalibrated the spectroscope and prepared an ethanol sample to be put in the cuvette. Yet now I still see no signal in the camera even with the gain set to its maximum value(I lowered it for the calibration). In the assembly video, Luc used the OpenRaman software to align the dichroic mirror, however the software does not work with me. I viewed the camera image on spinview instead but I start to wonder if maybe this programm is not appropriate for this job. It might also be of interest that I am only using the starter edition not the performance edition of the OpenRaman build. Has somebody another methode of aligning the mirror without using the spectrometer software?

If I understand correctly, you can get spectra of stuff like neon lamp but not of a Raman sample.

If that's correct, don't worry - you are experiencing the same difficulties as most people (including me) had during their first few alignments. Raman is pretty weak and it requires really good alignment to get a signal. Once you have done it a couple of times, it also becomes much easier. I also had less trouble with the liquid samples than with the solid ones in the past even if today I can align both without problem.

Using Galaxy is perfectly fine; just crank the gain to maximum and set an exposure time of 100 - 200 ms to have some real time feedback of your alignment.

If you're working with the Starter Edition, I have also found that laser alignment using the frosted disks is more difficult because the laser spot is about ~5 mm on a ~1 mm hole. Having good laser steering is really mandatory and I'd say about 90% of the time when it's not working it's coming from there. Try repeating the process with really the best centering you can do and only then fine tune the dichroic mirror using Galaxy software.

When using the rods w/ the frosted disks, check that the rods are firmly attached. If the set screws of the CP02B are not properly tighten (do not force - just a gentle push) the whole cage system things bends a bit due to gravity and it misalign your laser.

Well first let me say that you did understand me correctly and if there was any ambigouity in my post I am sorry. Second your point about the rods bending due to gravity makes sense. I noticed, when I used the mirror instead of only the frosted disct that the laser light hit the second disc more at the top when the laser was reflected back onto the slit. This could be a result of loosely tied up rods and that bend down due to gravity. I will follow your advice then and see if I can get a good laser alingment. Thanks for your help.