DIY spectroscope first light.. with CF bulb :)

[NickK]

Ok, it's probably not that pretty at the moment but considering this is a skeleton without a box protecting the light path...

The adjustable slit is ok to about 20um.. I have a fixed slit of 10um to try that will give a better separation and definition.

[Merlin66]

You're getting closer....

I'd double check the slit orientation - the emission lines, in a Classical design, should be vertical in the spectral band ( the ends will "bend" due to the off-axis aberrations).

Set the grating to the two yellow mercury lines - these are good for measuring focus/ resolution (FWHM) and clarity - they are 20A apart....

The optimum slit gap (with such a bright source) will be controlled by the pixel size of the camera.

Put your lens/ slit/ grating/ camera data into SimSpecV4 to get the theoretical performance....

[NickK]

I think I'm looking at this in reverse but:

I think it's this portion (my image is reversed from right to left):

Between the two most right lines (on my image, left on the spectra above) it's approx 384 pixels.

In theory, it's showing 0.67A, 479A range..

These are the two lines?

[Merlin66]

Yes,

That's the ones you want to look at....

[NickK]

I've researched the lines.. according to the following link the two yellow lines are 22A apart. Back of a napkin calculations using the spectra I took with the 589.5nm and 593.4nm bars are 384 pixels apart means it's 0.1015A/pixel.. which is substantially different from the designed 0.67A (would be 29.8 pixels).. now if they were 0.1A that would be rivalling the top-end daystar filters.. hence I don't quite believe it..

http://en.wikipedia.org/wiki/File:Spectra-Philips_32T8_natural_sunshine_fluorescent_light.svg

I'll set up and check the yellow lines.. at least they're a known and easier to check.

[NickK]

Hmm the peaks at 617.2 and 610.8 are 113 pixels apart, so 64A.. or 0.56A/pixel.. which seems a bit more believable.. but let me check this with the yellows..

[Merlin66]

Nick,

Never trust all you find on the web!

The yellow lines of Mercury are 5769.6 and 5789.66A, a separation of 20A.

Dispersion isn't the same as resolution.

With my 200mm fl. Littrow, a 29 micron slit, 1200 l/mm grating and an ATik 314L (6.45micron pixels) I get a dispersion of 0.25A/pixel and a resolution of R=5000.

[NickK]

Here's the yellow lines… (warning full sized 17.1MB tiff as an attachment)

I've measured (using preview box in PI so not that accurate perhaps) peak to peak - 182 pixels.

Therefore 20A / 182 = 0.109A/pixel .. errr holy moly. Can someone double check the resolution.

And a preview snippit cropped at native resolution for the lines only..

capture-2014-02-27 20:07:18 +0000-1.tiff

[NickK]

Nick,

Never trust all you find on the web!

The yellow lines of Mercury are 5769.6 and 5789.66A, a separation of 20A.

Dispersion isn't the same as resolution.

With my 200mm fl. Littrow, a 29 micron slit, 1200 l/mm grating and an ATik 314L (6.45micron pixels) I get a dispersion of 0.25A/pixel and a resolution of R=5000.

novaCen_Ha_270114.png

Looking at the spectra, some of the labels seemed a little non-linear.. so I had my suspicions (617 vs 620 for example).

I'm was hoping for about 0.5A - which would be perfect for my needs at the moment, phase two was to push this further. Here's a wider crop:

[Merlin66]

Nick,

You should always crop the full frame to give a "strip" about 60 pixel high.....

Based on the image...

The attached screen shot shows a dispersion of 0.1A and a FWHM of 0.9A giving a resolution of R=6000.

[NickK]

Nick,

You should always crop the full frame to give a "strip" about 60 pixel high.....

Based on the image...

The attached screen shot shows a dispersion of 0.1A and a FWHM of 0.9A giving a resolution of R=6000.

Hg lines_Capture.JPG

Thank you hehe go easy on the padawen

So would I be right in saying, currently I'm at 0.9A in terms of solar resolution?

I understood R to be resolving power (where above 10,000 was considered higher resolution).

Reason I ask is I'm trying to visualise what scale of image I would get from solar imaging at the current resolution. Actually - according to daystar they indicate "Half bandwidth FWHM" starting at 0.8A at their low end.. to 0.3A at the top end. So this would mean I'm at 0.9A .. not too shabby but I need to get to about 0.4-0.5A. Supposedly 40mm PSTs are "Nominal spectral bandwidth of <1.0A" and a doublestack solarmax II 60mm is <0.5 - <0.7A bandwidth.

As it's FHWM, basically it's reducing the width at the 1/2 point in the peak..

[Merlin66]

Measuring the FWHM is the safest way of confirming the resolution.

Yes,

the PST has a bandwidth FWHM around 0.7A to 1A etc

R is the measure of resolution for spectroscopes and equals the wavelength/ FWHM i.e. 5769.6/0.9= 6410

To improve the R value, you'd need to increase the focal lengths of the collimator/ imaging lens and use narrow slit gaps in conjunction with small pixel mono cameras.....

[NickK]

Measuring the FWHM is the safest way of confirming the resolution.

Yes,

the PST has a bandwidth FWHM around 0.7A to 1A etc

R is the measure of resolution for spectroscopes and equals the wavelength/ FWHM i.e. 5769.6/0.9= 6410

To improve the R value, you'd need to increase the focal lengths of the collimator/ imaging lens and use narrow slit gaps in conjunction with small pixel mono cameras.....

Ok, currently the collimator is 300mm f5.6, the final phase one design will boost that to 600mm f11.2 to match the vixen's 910mm f11.4.

[Merlin66]

Sounds good!

The imaging lens would also have to change...but this gives other issues - the plate scale (i.e. the size of the solar disk) will change and you'll probably need to build mosaics to cover the surface features.....

[NickK]

Sounds good!

The imaging lens would also have to change...but this gives other issues - the plate scale (i.e. the size of the solar disk) will change and you'll probably need to build mosaics to cover the surface features.....

Agreed - I'm not too fussed about full disks, instead I prefer to focus on targets. One reason for keeping this light is to co-mount with the pentax. Phase two means I can build in a viewer but that's going to be next year..

I'm going to summarise my "learnings" in the build thread..

[NickK]

I can see some additions to the capture application I use quickly appearing lol..

[Merlin66]

Nick,

I've been using AstroArtV5 for imaging with all my spectroscope cameras....DSLR/ DMK/ ATik/ Lodestar/ QHY5 etc. etc.

It allows you to do the necessary (batch) pre-processing (flat/ darks/ stacking/ cropping), then import the reference lamp and target spectrum into BASS Project and fix the tilt/ slant/ smile, then finally the wavelength calibration, normalising and presentation of the profile.

[NickK]

I can do dark, flats, stacking and cropping in realtime.

I should also be able to automate a spectra scan - with an audino controlling the grating, then perhaps using FFT based alignment

However I take your point - walk before running  post processing is something I need to get to grips with.

[NickK]

Here's a first attempt at 600mm collimator (f11.2) using the PM 2x to double up with the 10um fixed slit, cropped at native resolution - there's a lot of alignment and adjustments to be made:

To say this is hard is an understatement .. especially when the imaging lens is f4.. it's so hard I've had to modify the imaging lens to get the back focus..

However, how bad it is.. just look at them lines!!

Here's the order 0 reflection.. I now really pity those with f4 scopes..

[Merlin66]

Nick,

The spectrum image seems to show the brighter 1st order and then a faint 2nd order at the rhs.....but I don't recognise any of the lines..

Maybe re-setting the grating to centre on the yellow doublet?

Changing the collimator without changing the camera lens will, as I said, will not change the dispersion and not necessarily give a great increase in resolution.

Onwards and Upwards...

[NickK]

Nick,

The spectrum image seems to show the brighter 1st order and then a faint 2nd order at the rhs.....but I don't recognise any of the lines..

Maybe re-setting the grating to centre on the yellow doublet?

Changing the collimator without changing the camera lens will, as I said, will not change the dispersion and not necessarily give a great increase in resolution.

Onwards and Upwards...

I've marked on this spectrum where I think it is. It's definitely not adding more pixels between the lines, or really making them more defined (or tighter), but what it appears to be doing is to use the grating more efficiently - the net result is an image at 10um far better than last night (using 1 second exposure here)..

I think the left and right are actually reflections.. but I would expect the two 20A lines to be on the right of the image..

edit: but then thinking about the first images.. they were reversed.. so actually I think my guess is wrong - I think this is the aquamarine (green-cyan boundary) to the left but reversed.. *mental note* reverse camera for next attempt.. it may make it easier!