4 nights out with a Star analyser

[Jack Martin]

Most of the 17 spectra were obtained on 13-14-15-16 Jan 2012.

Any comments?

Jack

Star analyser results .pdf

[andrew s]

Nice set of images Jack. Have you tried extracting the profiles and callibrating them? If not take a look at this free software ISIS software. It even has a special set of processes sfor teh star analyser.

Andrew

[Helen]

Thye look good Jack.

Thanks for the link Andrew.

I got a star analyser for Christmas but haven't used it yet.

Helen

[Merlin66]

Jack,

Some good results there. Well done.

The Star Analyser gives about 30A resolution.

Did you apply a back ground sky removal?

What about an instrument correction?

You really need to process and calibrate them through RSpec or VSpec to generate the profiles, this will clear show the detail.

Onwards and Upwards.

[chiltonstar]

Jack,

Some good results there. Well done.

The Star Analyser gives about 30A resolution.

.

Is this really correct? I quote from the SA100 manual:-

"Because of these limitations, the resolution of this type of spectrograph is restricted to typically 1/50 to

1/100 of the wavelength (eg 5-10nm at 500nm), independent of the diffraction grating design"

ie about 50-100A, unless you are using it in 2nd order for example....?

Chris

[Merlin66]

With the right combination of camera pixel size and distance to grating, you can achieve a resolution of 30A.

It's being done all the time.

(The limitation is the optical aberrations inherent using the grating "in the converging beam" - add a collimator and imaging lens and you'll do MUCH better.)

[chiltonstar]

With the right combination of camera pixel size and distance to grating, you can achieve a resolution of 30A.

It's being done all the time.

(The limitation is the optical aberrations inherent using the grating "in the converging beam" - add a collimator and imaging lens and you'll do MUCH better.)

Ah....but was it done here?

I'm a pessimist (forensic scientist/chemical spectroscopist, what do you expect) so when I see apparent "line like" features in some of the SA100 data reported (not in this post) I get worried, because the implied resolution of the spectrum shown is vastly better than the actual resolution of a simple system like this, in other words, some of the features are likely to be processing artefacts.

Chris

[Merlin66]

I always get a bit sus when I hear of resolutions <30A for a grating....

If the target spectrum is not recorded horizontal across the CCD then yes, there is a high risk of introducing artifacts.

You just have to look at some of Torsten Hansen's work with a grating and DMK31 to see some very nice, good resolution results.

I have a TransSpec spreadsheet which analyses the grating in a converging beam and takes into account the various aberrations.

TransSpecV2_1.zip

[tom--rspec]

Hi, Jack! Nice spectra. I made a quick YouTube video that shows how it can be calibrated: Calibrating a Star Spectrum - YouTube

On Youtube, select the buttons at the bottom of the playback so that it plays it in HD, full screen so you can see what's going on!

As I mention briefly in the video, calibration of unknown spectra is much easier if you can see the star (called the "zero-order" along with the spectrum. Move your grating a bit further from your sensor to lower the dispersion. (Or, unzoom your DSLR if you're using an objective grating on the front of the lens.)

[Jack Martin]

Tom,

Glad to have you onboard. Just watched your video on UTube very good. I have noted your comments and will experiment accordingly lets hope for clear nights next week as the star analyser is on loan I have to return it next Saturday but I can use my Rainbow Optics grating instead.

Regards,

Jack

[Jack Martin]

Nice set of images Jack. Have you tried extracting the profiles and callibrating them? If not take a look at this free software ISIS software. It even has a special set of processes sfor teh star analyser.

Andrew

Andrew,

I have not extracted the profiles or calibrated them I will look at the link you sent thank you for you imput.

Regards,

Jack

[Jack Martin]

Thye look good Jack.

Thanks for the link Andrew.

I got a star analyser for Christmas but haven't used it yet.

Helen

Helen,

Thanks for your comments although many hours of work are involved its worth it so get out there use your star analyser and post the results.

Regards,

Jack

[Jack Martin]

Jack,

Some good results there. Well done.

The Star Analyser gives about 30A resolution.

Did you apply a back ground sky removal?

What about an instrument correction?

You really need to process and calibrate them through RSpec or VSpec to generate the profiles, this will clear show the detail.

Onwards and Upwards.

Merlin66,

I have not done any of the abovementioned. I don't know how to apply background sky removal.

Thanks for your input.

Jack

[Merlin66]

Jack,

if you save your original images as .fits files, you can open them under Vspec V4.0.5.

Under the "Extraction" tab - "Sky background subtraction"

This will automatically remove the background from your spectral image. You can then go on to prepare a profile for calibration.

[tom--rspec]

Tom,

Glad to have you onboard. Just watched your video on UTube very good. I have noted your comments and will experiment accordingly lets hope for clear nights next week as the star analyser is on loan I have to return it next Saturday but I can use my Rainbow Optics grating instead.

Regards,

Jack

Th RO grating is 200 lines/mm. The SA is 100 lines/mm. So, the RO grating will spread the spectrum out twice as much as the SA. The images you posted already had so much dispersion that the star wasn't visible. (The calibration in my video showed dispersion of about 3 Angstroms/pixel in the label just above the graph. That much dispersion also makes imaging dim objects more difficult because the light is more spread out.)

Doubling the dispersion is the opposite direction of what you need! If you're mounting the grating on the front of a DSLR, just reduce your lens' focal length (zoom).

If you're using the grating between your camera and your telescope, the you'll need to decrease the distance between the sensor and the grating.

If you want to explore a bit of the math around grating spacing, check out this write-up: link. There's a link there to a simple Excel workbook (SimpleCalc.xls) that lets you estimate where the grating should be placed.

The good news regarding grating spacing is that, unlike focusing for example, it's not critical to get it "exact." There, in fact, is no perfect grating spacing. Just get it approximately right, and then get out and then have fun doing science!

Also, for an example of how background subtraction is done, check out video #10 here: link.

[Jack Martin]

Jack,

if you save your original images as .fits files, you can open them under Vspec V4.0.5.

Under the "Extraction" tab - "Sky background subtraction"

This will automatically remove the background from your spectral image. You can then go on to prepare a profile for calibration.

Merlin 66,

Thanks for the information as the images are in colour will it work in V Specs?

Regards,

Jack

[Jack Martin]

Th RO grating is 200 lines/mm. The SA is 100 lines/mm. So, the RO grating will spread the spectrum out twice as much as the SA. The images you posted already had so much dispersion that the star wasn't visible. (The calibration in my video showed dispersion of about 3 Angstroms/pixel in the label just above the graph. That much dispersion also makes imaging dim objects more difficult because the light is more spread out.)

Doubling the dispersion is the opposite direction of what you need! If you're mounting the grating on the front of a DSLR, just reduce your lens' focal length (zoom).

If you're using the grating between your camera and your telescope, the you'll need to decrease the distance between the sensor and the grating.

If you want to explore a bit of the math around grating spacing, check out this write-up: link. There's a link there to a simple Excel workbook (SimpleCalc.xls) that lets you estimate where the grating should be placed.

The good news regarding grating spacing is that, unlike focusing for example, it's not critical to get it "exact." There, in fact, is no perfect grating spacing. Just get it approximately right, and then get out and then have fun doing science!

Also, for an example of how background subtraction is done, check out video #10 here: link.

Tom,

Thanks for the links the original images are raw and include the star for calibration. There is no lens between the DSLR and grating. I have some more of Cassiopeia to process.

Regards,

Jack

[Merlin66]

Jack,

I use Buil's freeware IRIS to convert the Canon RAW files to a fits format. IRIS also does the background removal and rotation of the spectrum to the horizontal.

The resulting fits file can then be read directly into RSpec or Vspec.

(RSpec can do all this automatically - RAW file in=Spectrum out!)

[Jack Martin]

Jack,

I use Buil's freeware IRIS to convert the Canon RAW files to a fits format. IRIS also does the background removal and rotation of the spectrum to the horizontal.

The resulting fits file can then be read directly into RSpec or Vspec.

(RSpec can do all this automatically - RAW file in=Spectrum out!)

Merlin 66,

I have IRIS and will give it a go.

Regards,

Jack