B&W-tech Spectrometer & 473 module: Setup+Mods+Info

[Mosc007]

Re: B&Wtech 473nm unit / Spectrometer Mods & Info

I have seen a couple of proper drivers from OS. Most are 110v input which means I would have to use a step down tranny from 240v. I have seen the 12v one from the US. I can get a local one with holder but $330 AUD seems a bit much.

I will give mine a run off a CCFL driver i have here. But maybe not enough output voltage. I have a HeNe supply with about the correct voltage but the current is only about 6ma. It may be ok. If not i might order the 12v one from the US.

Will leave it till the weekend to test.

 

[macona]

Re: B&Wtech 473nm unit / Spectrometer Mods & Info

I picked up one of these units as well as did a professor friend of mine. I pulled the flags and removed the bandpass filter yesterday and calibrated it with an ILT 5900 multi-line argon and a HeNe. Took me a bit to figure out which lines my argon was actually putting out. But once I did I got a nice tight pixel fit for the linear regression.

After cal I found that the spectrum range on my unit is 335nm to around 690, so a little shifted but gives me access to a good chunk of the UVA range. 632.8nm landed on pixel 1758 and 454.6nm landed on 611. We use a lot of 355nm lasers at work so i might bring the spec to work and check it.

I am using the science-surplus software which does work pretty well, only issue has been that it does not store the coefficients. Anyone else have this problem?

 

[paul1598419]

Re: B&Wtech 473nm unit / Spectrometer Mods & Info

Nope. I calibrated mine using several different software packages available here. The Science Surplus software stored my coefficients fine. I have concerns that you are not realizing that the calibration equation does not extrapolate at all, so anything outside of your two outer points are not even close to what you think they are. Example: 335nm.

 

[chloderic]

Re: B&Wtech 473nm unit / Spectrometer Mods & Info

I picked up one of these units as well as did a professor friend of mine. I pulled the flags and removed the bandpass filter yesterday and calibrated it with an ILT 5900 multi-line argon and a HeNe. Took me a bit to figure out which lines my argon was actually putting out. But once I did I got a nice tight pixel fit for the linear regression.

After cal I found that the spectrum range on my unit is 335nm to around 690, so a little shifted but gives me access to a good chunk of the UVA range. 632.8nm landed on pixel 1758 and 454.6nm landed on 611. We use a lot of 355nm lasers at work so i might bring the spec to work and check it.

I am using the science-surplus software which does work pretty well, only issue has been that it does not store the coefficients. Anyone else have this problem?

Hi, be welcome, have at look at the posting #359 and before and after


Here an other interesting picture of the ilumination of an HE-NE-tube without powerconnection, only holding it at an plasma-ball ...

 

[macona]

Hmm, I am going to have to disagree with the calibration, none of the spectrometers I have used (Ocean Optics and EG&G) required more than about 6 samples to get a good linear regression to scale and offset for calibration. Here are the instructions to calculate the coefficients: https://oceanoptics.com/wp-content/uploads/Spectrometer-Wavelength-Calibration-Instructions.pdf

To verify this i coupled the spectrometer to my double grating scanning UV-Vis-NIR monochromator and the spectrum was dead on till 380nm where it cut out. The signal drops off big time below 400nm, guessing the blaze angle is wrong for anything below that. I am shifting the grating to see if I can get more useful rad range, I shifted it a but last night to make sure the 400nm cutoff was not because of the sensor.

 

[diber]

The potential problem is that we are not doing linear regression (best fit of a linear function to a number of points) but higher order regression (due to the optics/CCD configuration, a linear function is not very accurate to describe the relation between pixel number and lambda... both Ocean Optics and Spectrum Studio use 3rd order polynomial). The higher the order of the polynomial, the more points you need to determine the coefficients, but also the higher the risk that the fitted function diverges from the real function (x^n changes very rapidly over a small range of x for large n, and hence the curve becomes very sensitive to small errors in the coefficient Cn) , especially outside the range of the measurement points, but possibly even inside the range). It can go well with low number of calibration points and even outside the range covered by the calibration lines, but it is much better and safer (i.e. less sensitive to calibration measurement errors) to use more points, covering a range as wide as possible. This is valid for the BW spectrometer, but also for the Ocean Optics spectrometer!

 

[chloderic]

I would like to describe another mod of another part of this unit.
First, i removed the collimator from the beam splitter,
to do this you just need to loosen the small grub screw and then unscrew the part (see picture 1).
Then I dissected up a leftover microscope eyepiece (see picture 2 + 3).
In addition I have a few seals, a pipe-screw-connection and an ultra langpassfilter (see picture 4 + 5).
Then I installed these parts instead of the lenses in the eyepiece body.
Finished is a SMA-fiber-filterholder for my DIY laser raman microscope (see picture 6 + 7 + 8).

As you see, you could use other part of this unit as well .....

 

[Agastar]

I would like to describe another mod of another part of this unit.
First, i removed the collimator from the beam splitter,
to do this you just need to loosen the small grub screw and then unscrew the part (see picture 1).
Then I dissected up a leftover microscope eyepiece (see picture 2 + 3).
In addition I have a few seals, a pipe-screw-connection and an ultra langpassfilter (see picture 4 + 5).
Then I installed these parts instead of the lenses in the eyepiece body.
Finished is a SMA-fiber-filterholder for my DIY laser raman microscope (see picture 6 + 7 + 8).

As you see, you could use other part of this unit as well .....

Man, that looks really cool!

 

[chloderic]

Hmm, I am going to have to disagree with the calibration, none of the spectrometers I have used (Ocean Optics and EG&G) required more than about 6 samples to get a good linear regression to scale and offset for calibration. Here are the instructions to calculate the coefficients: https://oceanoptics.com/wp-content/uploads/Spectrometer-Wavelength-Calibration-Instructions.pdf

To verify this i coupled the spectrometer to my double grating scanning UV-Vis-NIR monochromator and the spectrum was dead on till 380nm where it cut out. The signal drops off big time below 400nm, guessing the blaze angle is wrong for anything below that. I am shifting the grating to see if I can get more useful rad range, I shifted it a but last night to make sure the 400nm cutoff was not because of the sensor.

In the datasheet of the sensor you find the spectral sensivity, it is not defined below 400 nm but it seems to go down ...

 

[chloderic]

Hi, because it is a little bit silent here, i would like to post my last MOD of parts of this device to discuss ... ?
This is the prototype of my DIY raman-spectroskope with an 532nm DPSS-laser (about 930mW), an russian microscope, premium-longpassfiler, XY-Stage, LED illumination and an DC-powersupply to regulate the laser output ....... , here it is about 300 mw .... therefore the filter has no chance ... .
You will see in spectrum-studio the raman-shift of a diamond. This assembly is to identify unknown minerals in this case.

 

[paul1598419]

Very nice how you have repurposed parts from other parts of the laser assembly to get your RAMAN spectrometer up and running, chloderic. I have been following just to know what you have been doing as RAMAN spectroscopy is not what I am particularly interested in. Still, it is great information. :thanks:

 

[chloderic]

Very nice how you have repurposed parts from other parts of the laser assembly to get your RAMAN spectrometer up and running, chloderic. I have been following just to know what you have been doing as RAMAN spectroscopy is not what I am particularly interested in. Still, it is great information. :thanks:

Hello Paul, yes, you are absolutely right and expressed it politely and elegantly. The theme RAMAN is certainly not interesting for everyone here and also not quite suitable for the topic. Maybe i should open a new topic on this. You're actually everywhere here , so you know yourself pretty well an in every case better than me ... is there such a topic? I have not found anyone yet, but I do not want to open a double theme ...

 

[paul1598419]

Actually, I can see myself setting one of these up to do RAMAN spectroscopy. I used to use all sorts of spectroscopy in college mostly in chem courses. Have used NMR, GC-mass spec., IR spectroscopy and others. Your project is fascinating to me and since I have been interested in mineralogy since I was a kid I could find a use for this. So, that is why I've been following your success here. Since this is a laser forum, there are fewer people here interested in what you are doing specifically. Most just want an accurate spectrometer to measure wavelengths of laser diodes. But, there is so much more than that which can be done with these.

 

[Civitus]

Hi, because it is a little bit silent here, i would like to post my last MOD of parts of this device to discuss ... ?
This is the prototype of my DIY raman-spectroskope with an 532nm DPSS-laser (about 930mW), an russian microscope, premium-longpassfiler, XY-Stage, LED illumination and an DC-powersupply to regulate the laser output ....... , here it is about 300 mw .... therefore the filter has no chance ... .
You will see in spectrum-studio the raman-shift of a diamond. This assembly is to identify unknown minerals in this case.

Hi Chloedric!
Very nice job! great!
I am also working on my own Raman system. I am getting some parts from a supplier on eBay.
But, for the laser, as we are in a forum of lasers, it has to be a narrow linewidth, and they are quite expensive ...

I would like to know more how you hav connected the optical fiber to the microscope eyepiece. Can you tell me more or we can exchange on private messages.

Again .... great job

C

 

[paul1598419]

If you are using one of these B&W TEK 473nm lasers with a spectrometer, you already have a laser to use. It is the 473nm DPSS laser that came with the spectrometer. There are other ways to get a narrow linewidth with a direct diode laser. You can diffraction grating tune the output of most any single mode laser in order to get a linewidth less than 1 MHz. There are YouTube videos on how to do this if you are not sure.

 

[kecked]

I’m doing the same but hooking the Raman up to a confocal microscope. Maybe we should start a separate thread for this.