The Lowspec spectrometer

[Thalestris24]

1 hour ago, Merlin66 said:

Louise,

Well done! I see you using BASS Project....

Use the earlier fluoro reference, it's more reliable. Put it down to different CFL lamps.

You could also image a solar spectrum by pointing the LowSpec to a bright sky...

The profile you show is very promising, you'll get a better result using multi-point calibration..

The scaling of your result ----not quite correct. I think you may have used a setting which would apply to a target spectrum, not a reference spectrum.

I would also change the wavelength units to Angstroms...more widely used in astronomy.

 

Hiya - thanks Yes first go with BASS Not sure what you mean by 'earlier fluoro reference'? I'm obviously not really understanding/familiar with BASS. I've only used Rspec before and with the SA100. On the 'element' menu there is only elements like neon, mercury etc (as you know ha ha). So I just assumed that first brightish peak was mercury and left it at that for now. I'll have to study it some more. 

Cheers and Merry Christmas!

Louise

[Merlin66]

Louise,

I recopied the fluoro reference I use to my last message..

Just select the line, look up the wavelength (from the reference image I gave) and enter that, continue for the lines you can see.

We're not trying to really identify the lines at this stage, just use them to provide a calibration for our system.

Some good tutorials here: https://groups.io/g/BassSpectro/wiki/Tutorials-%26-Guides

 

[Thalestris24]

10 minutes ago, Merlin66 said:

Louise,

I recopied the fluoro reference I use to my last message..

Just select the line, look up the wavelength (from the reference image I gave) and enter that, continue for the lines you can see.

We're not trying to really identify the lines at this stage, just use them to provide a calibration for our system.

Some good tutorials here: https://groups.io/g/BassSpectro/wiki/Tutorials-%26-Guides

 

I recopied the fluoro reference I use to my last message.. ok, I understand what you meant now

Just select the line, look up the wavelength (from the reference image I gave) and enter that, continue for the lines you can see. I don't understand exactly how to do that yet. As I say, I'll study some more.

I was trying to follow the slit processing tutorial (word doc on the group page). It suddenly starts talking about a 'reference spectrum' without explaining what it is, which confused me earlier. I'll have to read through the BASS documentation... Don't rush me - I'm old and dim!

Thanks

Louise

[Merlin66]

No rush, no pressure...

identify the line from the fluoro profile I posted and read the wavelength from the same fluoro profile, enter that in the calibration pop up screen.

When calibrated, this image then becomes your “reference image”

when you take a Spectral image of a star, you will also take an image of your reference lamp ( fluoro, Neon etc) then use that image to calibrate your star spectrum, the “target spectrum”

I’ll let you have a good nights sleep......

[Thalestris24]

Received the fluorescent starters so, excitedly, connected up the bulb from an S10 one. Here is raw spectrum:

The micrometer setting was the same as yesterday so goes from ~4000A to ~7200A. Exposure was 3m so takes quite a long time. I'll try and put it through BASS Project later and compare it to the spectrum data Eric kindly provided.

Louise

[andrew s]

Nice work Louise, starting to look the business. What width slit are you using for the spectra above?

Regards Andrew 

[Thalestris24]

Just now, andrew s said:

Nice work Louise, starting to look the business. What width slit are you using for the spectra above?

Regards Andrew 

Hiya

Thanks! That was with the 10um slit. I left everything as was yesterday. I need to try and get a better focus, I think. 

Louise

[Merlin66]

Louise,

Looking promising....

You could use your FCL as a "reference" image to provide the calibration for the S10 spectrum....

This would get you into the ball park..you can then refine the calibration using identified lines.

 

[Thalestris24]

1 minute ago, Merlin66 said:

Louise,

Looking promising....

You could use your FCL as a "reference" image to provide the calibration for the S10 spectrum....

This would get you into the ball park..you can then refine the calibration using identified lines.

 

Hiya

I have all the identified lines for the S10 bulb given to me by Eric. I just have to make sure my spectrum matches. Will have a go tomorrow  

Louise

[Merlin66]

I realise you have the S10 data. But calibrating say the Fuoro then using it as a BASS reference (or visversa the S10 calibrated to the data given used to re-calibrate the Fluoro) gets you closer to the final real world processing of a reference and a target start spectrum.

 

[Thalestris24]

Just now, Merlin66 said:

I realise you have the S10 data. But calibrating say the Fuoro then using it as a BASS reference (or visversa the S10 calibrated to the data given used to re-calibrate the Fluoro) gets you closer to the final real world processing of a reference and a target start spectrum.

 

Oh ok - hopefully it will all become clearer....  

Thanks

Louise

[Thalestris24]

Hi Ken - if you're about! 

I'm struggling a bit to achieve a calibrated spectrum... Trying to match more than a few lines isn't easy... A couple of questions: can I zoom into the chart to get closer to selected lines and can I change the colour of the chart spectral plots? At the moment I have one blue and one green and it's not very contrasty. Black and orange, say, would be much better!

Cheers

Louise

[Merlin66]

Louise,

Right click on the image (active image has the yellow boarder), Profile Properties/ Line/ Line colour and select the colour you want....repeat for the other profile.

Yes, you can crop the X axis - Chart/ Crop X axis range.... selected the wavelength coverage which suits, then revert by Chart/ Uncrop X axis.

To get a reasonable calibration you only need to identify say 5-6 (max) lines across the spectral image - set Calibration fit to Quadratic.........

Ken

 

Edited December 28, 2019 by Merlin66

[Thalestris24]

1 minute ago, Merlin66 said:

Louise,

Right click on the image (active image has the yellow boarder), Profile Properties/ Line/ Line colour and select the colour you want....repeat for the other profile.

Yes, you can crop the X axis - Chart/ Crop X axis range.... selected the wavelength coverage which suits, then revert by Chart/ Uncrop X axis.

Ken

 

Ok, thanks!  

What's the best easily available and reliable source of defined spectral lines? I'm probably overthinking/over-pedantic about this but different sources (ie different people) quote different values so I'm having trouble with both the cfl spectrum and the s10 starter bulb spectrum. It's driving me nuts! Although Eric kindly provided me with a detailed S10 spectrum, it's on a different scale to mine and I can't see how to reliably line them up, so to speak. How many matches do I need for an accurate calibration?

Cheers

Louise 

[Merlin66]

Louise,

Softly softly....

1. Start with the CFL as a "reference image" (place up front as #1 on the list). Us the reference image I posted with the lines identified and mark/ calibrate  all the obvious lines you can see.

This will give you a pretty good calibration of your instrument set-up. Once you have done this then.....

2. Import the S10 image (as #2 image on the list). Right click on the image, Properties/ Calibration/ select "use calibration from the first profile" / copy. This will then calibrate the S10 image form the "reference image" of the FCL.

3. If you now want to positively identify any/ all the lines in the S10 spectrum - use the mouse to move over the lines (Crop the X axis if needed) to match lines with the reference that Eric gave. If you want, you can then lable some of the lines - for future reference - Click on the "label" icon, in the pop down screen enter the type of lable you want to use (check the BASS manula for details) When done, save the chart as a .png (Chart/ save chart to file)

Re-reading your message..the S10 data is just that...data which you manually have to select and enter, you can't just superimpose the image.....

.Ken

 

Edited December 28, 2019 by Merlin66

[Thalestris24]

2 minutes ago, Merlin66 said:

Louise,

Softly softly....

1. Start with the CFL as a "reference image" (place up front as #1 on the list). Us the reference image I posted with the lines identified and mark/ calibrate  all the obvious lines you can see.

This will give you a pretty good calibration of your instrument set-up. Once you have done this then.....

2. Import the S10 image (as #2 image on the list). Right click on the image, Properties/ Calibration/ select "use calibration from the first profile" / copy. This will then calibrate the S10 image form the "reference image" of the FCL.

3. If you now want to positively identify any/ all the lines in the S10 spectrum - use the mouse to move over the lines (Crop the X axis if needed) to match lines with the reference that Eric gave. If you want, you can then lable some of the lines - for future reference - Click on the "label" icon, in the pop down screen enter the type of lable you want to use (check the BASS manula for details) When done, save the chart as a .png (Chart/ save chart to file)

Ken

 

Ok, thanks - I'll give that a go tomorrow.

Cheers

Louise

[Thalestris24]

8 minutes ago, Merlin66 said:

Re-reading your message..the S10 data is just that...data which you manually have to select and enter, you can't just superimpose the image.....

.Ken

 

Just to say, I wasn't trying to superimpose as such, just aiming to match significant lines. But it's not so easy...

Louise

[Thalestris24]

Here is my S10:

CFL:

Both:

The S10 is a bit of a jungle at the red end! Presumably that's mostly neon plus other stuff.

Anyway, I'll give it another go starting from the cfl tomorrow though different people quote slightly different wavelengths for some of the lines...

Louise

[Merlin66]

28 minutes ago, Thalestris24 said:

Here is my S10:

CFL:

Both:

The S10 is a bit of a jungle at the red end! Presumably that's mostly neon plus other stuff.

Anyway, I'll give it another go starting from the cfl tomorrow though different people quote slightly different wavelengths for some of the lines...

Louise

Louise,

Use the Buil reference image I supplied for the CFL - it has been used successfully by many.

Need to check back in the messages...but I assume you have the line/ wavelength data for the S10?

ken

 

[Thalestris24]

40 minutes ago, Merlin66 said:

Louise,

Use the Buil reference image I supplied for the CFL - it has been used successfully by many.

Need to check back in the messages...but I assume you have the line/ wavelength data for the S10?

ken

 

I have that but I don't like that it has ? against most of the lines! But if you say it's ok... I have managed to find some of Richard Walkers published plates and data for various starter and fluoro bulbs so I'll peruse that as well.

I should be asleep now, so off to bed!

Thanks

Louise

[ejp1684]

Louise, I did a quick check of your result and found the following from the S10 calibration data. I don't quite understand your comment about there being '?'s against most of the lines, I read these values straight from the charts.

I find the lower end needs more exposure to get accurate results. Other bulbs may work better for you; I used the S10 because it was there on the shelf at B&Q!

Happy calibrating!

Eric.

[Thalestris24]

1 hour ago, ejp1684 said:

Louise, I did a quick check of your result and found the following from the S10 calibration data. I don't quite understand your comment about there being '?'s against most of the lines, I read these values straight from the charts.

I find the lower end needs more exposure to get accurate results. Other bulbs may work better for you; I used the S10 because it was there on the shelf at B&Q!

Happy calibrating!

Eric.

Thankyou very much, Eric! The above is a 120s exposure. I have a 180s exposure also but thought the red end was a bit overcooked, though I've not actually tried viewing it in BASS. I obtained Richard Walker's published lines for the Relco 480 and the Osram ST 111 starters last night. Unsurprisingly, I suppose, they are both different to the S10... I've no doubt there is commonality between them and the S10 but it needs teasing out. I'm a bit ocd when it comes to lining things up... I guess a few iterations between your calibration of my S10 and my CFL should finalise things, as Ken instructed .

Happy New Year!

Louise 

PS I was referring to the ?s on the calibration of the CFL that Ken gave me

Edited December 28, 2019 by Thalestris24

[Thalestris24]

Oh well, ended up with these - S10 (via Eric ) and CFL calibrated from the S10:

Unfortunately using the S10 to calibrate the CFL results in values that don't quite line up - there must be an error somewhere...

The purple vertical lines are my attempt to add labels. Only two of them came out Upon saving the data in the label editor, the label seems to disappear. Nevertheless, the purple lines should line up with the corresponding peaks but they are a bit off. Maybe I'll have to redo the S10 calibration... I hope I don't have to do this procedure every time I need to calibrate a target even when using the same settings? That would get tedious!

Louise

[Thalestris24]

Oh, I was following the slit processing tutorial which says 'now that the target has its own calibration data, the reference spectra can be removed'. I assumed I'd be able to get it back? Looks like it's gone so I'll have to do the S10 all over again 😢.

Louise

[andrew s]

58 minutes ago, Thalestris24 said:

Oh, I was following the slit processing tutorial which says 'now that the target has its own calibration data, the reference spectra can be removed'. I assumed I'd be able to get it back? Looks like it's gone so I'll have to do the S10 all over again 😢.

Louise

If your using bass and you save the calibrated image the calibration will come back if you load the saved image. If you make it No 1 in the sequence then you can apply it to new spectra. 

Regards Andrew