Hi there,
I'm currently trying to align the Raman spectrometer so that we get see a spectrum on the grating and capture it with our camera. I'm watching the OpenRAMAN tutorial video, and I see that the alignment process is made easier with the OpenRAMAN software. However, I currently don't have a camera that pairs well with the software, so I was wondering if there was some other techniques or demo I could watch to see how I can do it without the OpenRAMAN software.
This is the image I'm getting so far. There's some red dots and the green light is from the laser.
https://freight.cargo.site/w/1164/h/880 ... .25-PM.png
Alignment Question
Re: Alignment Question
Hi,
It is possible to use the camera software directly for the alignment. At which stage of the alignment process are you stuck ?
Also, but that's more of a general remark, it is not recommended to use a color camera sensor because you are throwing away more than 75% of the signal when using a green laser and some of the peak resolution too. So when I'm using a ~10 sec exposure time on a sample, you can expect your camera to require at least 40 sec, even up to 1 minute exposure for the same results.
Not all camera are equal also in terms of noise and the one I chose was carefully selected for its low-noise capabilities. Could you tell me the exact camera model you're using? I'll check if I can find some info on its noise specification.
It is possible to use the camera software directly for the alignment. At which stage of the alignment process are you stuck ?
Also, but that's more of a general remark, it is not recommended to use a color camera sensor because you are throwing away more than 75% of the signal when using a green laser and some of the peak resolution too. So when I'm using a ~10 sec exposure time on a sample, you can expect your camera to require at least 40 sec, even up to 1 minute exposure for the same results.
Not all camera are equal also in terms of noise and the one I chose was carefully selected for its low-noise capabilities. Could you tell me the exact camera model you're using? I'll check if I can find some info on its noise specification.
-
BowenZheng
- Posts: 13
- Joined: Wed Jan 31, 2024 11:38 pm
Re: Alignment Question
Hi Luc,
I have all the components I need on the baseplate. I've focused the laser on the cuvette sample and have it reflected back. Now, I'm working on aligning the lens group, compensation window, and longpass filter so that the camera can properly detect the spectrum on the diffraction grating.
For the camera, I'm using an Allied Vision Prosilica GC camera https://www.alliedvision.com/en/product ... silica-gc/ and the VimbaX software. I'll try using only black and white. Just to clarify, are you saying that if we don't use the color camera sensor, the exposure time would only be around 10 sec?
Thank you,
Bowen
I have all the components I need on the baseplate. I've focused the laser on the cuvette sample and have it reflected back. Now, I'm working on aligning the lens group, compensation window, and longpass filter so that the camera can properly detect the spectrum on the diffraction grating.
For the camera, I'm using an Allied Vision Prosilica GC camera https://www.alliedvision.com/en/product ... silica-gc/ and the VimbaX software. I'll try using only black and white. Just to clarify, are you saying that if we don't use the color camera sensor, the exposure time would only be around 10 sec?
Thank you,
Bowen
Re: Alignment Question
Which camera model of the ProSilica are you using exactly?
Exposure time is typically around ~10 seconds on the Starter Edition for precision measurement to fully expose the sensor and get the highest dynamic range but it's still possible to recognize spectra with lower exposure time, down to 100 ms (but then it's super noisy). With a 75 mW laser I can recognize acetone at 1 ms exposure.
Exposure time is typically around ~10 seconds on the Starter Edition for precision measurement to fully expose the sensor and get the highest dynamic range but it's still possible to recognize spectra with lower exposure time, down to 100 ms (but then it's super noisy). With a 75 mW laser I can recognize acetone at 1 ms exposure.
-
BowenZheng
- Posts: 13
- Joined: Wed Jan 31, 2024 11:38 pm
Re: Alignment Question
Hi Luc,
The camera model is GC1380CH
Best,
Bowen
The camera model is GC1380CH
Best,
Bowen
Re: Alignment Question
I looked up on Internet to gather some data on this camera model and I found this page : https://vitalvisiontechnology.com/allie ... -145b-poe/
It's an old CCD sensor (ICX285) with a not-so-good light sensitivity. The website mentions an absolute detectivity threshold of 11 e- with a quantum efficiency of 54% which translates (at 529 nm) to a photon detection threshold of 20.37 photons -- to be compared with the 3.9 photons of the 31S4M sensor. In the monochrome (black/white) version, you can therefore expect a factor 5.2× less sensitive for this camera + the additional losses due to the Bayer filter (~4×). I would therefore expect a total loss of 21× or slightly higher with this camera.
This is going to be problematic because, during alignment, I typically use exposure time on the order of 100-200 ms to adjust the laser steering. You would therefore have to use exposure times in the range of 2 sec ~ 4 sec for the alignment which makes it impractical to get real-time feedback to tune the system angles.
I would recommend that you look for another camera to avoid spending too much time struggling with the alignment process. If you're on a budget, I can recommend Daheng cameras which have very nice performance/cost ratios. I have personal experience with the MER2-160-75GM-P for OpenRaman and one user have used a MER2-302-37GM-P. Note that the MER2-160-75GM-P will require a 35 mm camera lens so if you already bought a 50 mm lens, it's better to use the MER2-302-37GM-P.
It's an old CCD sensor (ICX285) with a not-so-good light sensitivity. The website mentions an absolute detectivity threshold of 11 e- with a quantum efficiency of 54% which translates (at 529 nm) to a photon detection threshold of 20.37 photons -- to be compared with the 3.9 photons of the 31S4M sensor. In the monochrome (black/white) version, you can therefore expect a factor 5.2× less sensitive for this camera + the additional losses due to the Bayer filter (~4×). I would therefore expect a total loss of 21× or slightly higher with this camera.
This is going to be problematic because, during alignment, I typically use exposure time on the order of 100-200 ms to adjust the laser steering. You would therefore have to use exposure times in the range of 2 sec ~ 4 sec for the alignment which makes it impractical to get real-time feedback to tune the system angles.
I would recommend that you look for another camera to avoid spending too much time struggling with the alignment process. If you're on a budget, I can recommend Daheng cameras which have very nice performance/cost ratios. I have personal experience with the MER2-160-75GM-P for OpenRaman and one user have used a MER2-302-37GM-P. Note that the MER2-160-75GM-P will require a 35 mm camera lens so if you already bought a 50 mm lens, it's better to use the MER2-302-37GM-P.
-
BowenZheng
- Posts: 13
- Joined: Wed Jan 31, 2024 11:38 pm
Re: Alignment Question
Hi Luc,
I don't have that one specifically. I've attached the link to the datasheet for the camera I have, so I don't know if what you mentioned for the Allied Vision Manta G-145B PoE still applies to this camera.
https://cdn.alliedvision.com/fileadmin/ ... eet_en.pdf
Best,
Bowen
I don't have that one specifically. I've attached the link to the datasheet for the camera I have, so I don't know if what you mentioned for the Allied Vision Manta G-145B PoE still applies to this camera.
https://cdn.alliedvision.com/fileadmin/ ... eet_en.pdf
Best,
Bowen
Re: Alignment Question
Sorry, my bad -- for some reason when I googled for the GC1380H it proposed a datasheet for another model and I didn't pay attention to the model number 
Here is the correct one : https://www.alliedvision.com/en/camera- ... -gc/1380h/
The absolute sensitivity threshold is 18e- with a QC of 54% @ 529 nm which translates to 33.4 photons -- in monochromatic conditions.
Also, I see that the sensor (ICX285) has 6.45 µm pixels which means they can collect 3.5× as much photons as the pixels from the camera I'm using. That means a light flux sensitivity ~9.5 photons if we want to compare with the 3.9; still a factor 3. Since these properties are given in monochromatic conditions, we need to scale by a factor ~6.2 for the red (based on the Quantum Efficiency curves and the bayer pattern). That adds up to a sensitivity loss of 15× so the conclusions drawn previously still hold.
A second problem with the larger pixel size is that the setup was made for 3.45 µm pixels, which means that you will not respect the guidance rule of a factor ~3 between dispersion and resolution -- this can be problematic when trying to resolve peaks.
You can try with that camera but it's going to be very complicated; I wouldn't risk it.
Here is the correct one : https://www.alliedvision.com/en/camera- ... -gc/1380h/
The absolute sensitivity threshold is 18e- with a QC of 54% @ 529 nm which translates to 33.4 photons -- in monochromatic conditions.
Also, I see that the sensor (ICX285) has 6.45 µm pixels which means they can collect 3.5× as much photons as the pixels from the camera I'm using. That means a light flux sensitivity ~9.5 photons if we want to compare with the 3.9; still a factor 3. Since these properties are given in monochromatic conditions, we need to scale by a factor ~6.2 for the red (based on the Quantum Efficiency curves and the bayer pattern). That adds up to a sensitivity loss of 15× so the conclusions drawn previously still hold.
A second problem with the larger pixel size is that the setup was made for 3.45 µm pixels, which means that you will not respect the guidance rule of a factor ~3 between dispersion and resolution -- this can be problematic when trying to resolve peaks.
You can try with that camera but it's going to be very complicated; I wouldn't risk it.
-
BowenZheng
- Posts: 13
- Joined: Wed Jan 31, 2024 11:38 pm
Re: Alignment Question
Hi Luc,
Thank you for the informative answer. Could you describe what specs I should be looking for in a camera? I know you've already mentioned some like pixel sizes smaller than 3.45 µm pixels, but I was wondering if there are any others I should be aware of.
Best,
Bowen
Thank you for the informative answer. Could you describe what specs I should be looking for in a camera? I know you've already mentioned some like pixel sizes smaller than 3.45 µm pixels, but I was wondering if there are any others I should be aware of.
Best,
Bowen
Re: Alignment Question
3.45 µm pixels, monochrome, very low absolute sensitivity threshold and at least 2048 px horizontally if you're using a 50 mm lens.
The FLIR camera I'm using has a IMX265 sensor so you can look at substitutes that relies on the same sensor (in monochrome). Apart from the FLIR and Daheng cameras already mentioned, the Basler acA2040-35gm might be a match or the Baumer VCXG.2-32M. I'm not a big fan of Baumer btw but have some consideration for Basler.
With a 35 mm lens, I got satisfactory results with IMX273 cameras which are less expensive but also have fewer pixels. This sensors works great on the OpenRAMAN so I can recommend it.
I don't recommend going for larger sensors (IMX264, IMX304 etc.) because they are more expensive and won't offer you a lot of value compared to the price.
I also don't recommend sensors that are not from Sony; either because I don't have experience with them and can't say if they are any good in term of noise performances, or because I already have experience with them and can tell you they aren't suitable for Raman spectroscopy (Aptina, Cmosis etc.).
The FLIR camera I'm using has a IMX265 sensor so you can look at substitutes that relies on the same sensor (in monochrome). Apart from the FLIR and Daheng cameras already mentioned, the Basler acA2040-35gm might be a match or the Baumer VCXG.2-32M. I'm not a big fan of Baumer btw but have some consideration for Basler.
With a 35 mm lens, I got satisfactory results with IMX273 cameras which are less expensive but also have fewer pixels. This sensors works great on the OpenRAMAN so I can recommend it.
I don't recommend going for larger sensors (IMX264, IMX304 etc.) because they are more expensive and won't offer you a lot of value compared to the price.
I also don't recommend sensors that are not from Sony; either because I don't have experience with them and can't say if they are any good in term of noise performances, or because I already have experience with them and can tell you they aren't suitable for Raman spectroscopy (Aptina, Cmosis etc.).