Sa100

[bojan]

While waiting for BW CCD sensor for my Logitech 4000 camera (I will also attempt the long exposure mod then), I was playing with SA100 grism, mounted on C11.

I took spectra of Betelgeuse and Sirius on Wednesday evening... it was quick&dirty job, exposure (couple of seconds) was done with moded Canon450D @ ISO100. To avoid saturation, I drifted the scope along RA, thus obtaining wide spectra like there was a slit in the light path...
Of course, this was possible only because both stars are sufficiently bright.. With fainter objects, I will have to use higher ISO settings and longer exposures with accurate tracking.

It was relatively easy to align the grism in my mechanical solution - grism is rotatable with respect to camera (~+/- 10degrees), and of cource the whole thing can be rotated so the flrame could be aligned perpendicluar/paralel to celestial coordinate grid.


EDIT:
In the current setup, grism distance from sensor is ~100mm (limitated by prism holder diameter and inner diameter of the C11 baffle tube).

There is a bit of overlapping of second order spectrum at IR side (right side, not easily visible on attached images), I will sort it out with 680nm IR filter (I have one left over from Canon 400mm FP lens setup, now I will use it here).

[Merlin66]

Bojan,
Very nice result!
Terminology: the SA100 is just a transmission grating; a Grism is a combination of a prism plus grating.
You first image shows sufficient drift, just needs better focus on the spectrum.
BASS Project will allow you select and sum the spectral columns and present a profile for wavelength calibration...
Looking at the image profiles the dispersion changes between them, 3.66 A/pix for the Betelgeuse and 5.4 A/pix for Sirius. For a grating this is very high, normally 8-10 A/pix target.
(cropped/ changed images for IIS??)

Well done!

[bojan]

Ken,
Thank you for your analisys.
The images I attached earlier are full size (from 450D), no cropping.
I did some wl calibration in BASS, it was not easy to identify some features (especially with Betelgeuse)... I (wrongly?) assumed Halpha was marked with red circle...
I also assume the feature marked by Green circle is camera response. Have to learn how to calibrate this.
Will get there :-)

[Merlin66]

Bojan,
Interesting, if both spectra were taken with the same set-up - grating spacing/ camera etc. I'd expect the dispersion to be the same .....

The green dip is the "Canon" dip caused by the overlap of the Bayer filters in the DSLR camera. This can easily be compensated when doing the instrument response processing.
See the example "Astronomical Spectroscopy for Amateurs", p100.

Betelgeuse being a red giant M2/M3 has many Tio bands in the red.
Remember the Saha graph.
(http://www.astronomicalspectroscopy.com/How_to.html)
This shows there is little possibility of Hydrogen in the spectrum.

The red dip looks to me like the pronounced Telluric band at 7600A, a good and easy marker....
EDIT: Looking at the MILES M2 reference...I had another go at re-calibrating (at 4.5 A/pix)

[bojan]

Yep, the spectra were taken with the same setup, same evening, one after another.

Difference is in position of the "0" order, I did not take sufficient care of this.. Next time I will use the lap screen for more accurate aim at target.

[Merlin66]

Bojan,
Hmmmm
The position of the zero order doesn't impact on the dispersion...
It's best to position the zero order towards the left of the frame and centre the spectrum across the chip. Focus on the spectrum.

[bojan]

For positioning/ficussing I was using magnifier lens (http://www.iceinspace.com.au/forum/showpost.php?p=1300302&postcount=4), placed at the centre of the camera screen, this way it is easy to focus on spectrum... but not easy to place "0" order on the edge, as it's not visible with the lens...


BTW, my setup is similar to this (http://www.astronomicalspectroscopy.com/images/grating_02.jpg) one, except I have prism in front of grating.

[bojan]

My setup...
It screws on C11 2" thread, grism is centered in the barrel.

[Merlin66]

Ahhh, OK.
Which prism wedge are you using?
The prism obviously impacts on the linearity of the dispersion.

[Merlin66]

Tim Stone used his SA100 to get this result.
Not 100% sure about the line identification.....

[bojan]

I tried to recalibrate the spectra (starting with Sirius, using Balmer series as reference).
Came up with 0.432nm/pxl, almost the same as your 2nd result.
Applied the same dispersion to Betelgeuse (corrected for hor shift), both spectra now look consistent (green dip and some other less prominent features).

[bojan]

I ordered the prism from Bxoptic (http://www.bxoptic.com/), they had it in stock and it was pretty cheap (I am trying to find the quote...) and the delivery was very quick, I highly recommend them (I used them for other things in the past when I worked for MCE Lasers).
I think it was ~3° or so, almost ideal match for SA100.

[bojan]

Wedge prism details as follows:

material :BK7,
Dia :38.1mm
thickness (thick edge ) :4.66mm
wedge angle :4degree +/-1 arcmin
AR coating , 450-650nm

1 pcs, the unit price was 35.00 USD (for one) + US$20 shipping.

Better than Edmund....

[Merlin66]

Bojan,
Thanks for that - they have a very impressive set-up.

[bojan]

In anticipation of T CrB outburst, I re-redesigned my spectrometer a bit..

The main reason was to make it more compatible with my existing astro-photo gear.

Because the alignment mechanism in old version was adding ~30mm to backfocus distance (shown on first image), a lot refocusing was needed to switch from one configuration to other.

Also, the old version was quite long (grism was inserted into C11 flange opening when attached), so I couldn't use it with FR.

This grism version is shorter in overall, so now there is no clash with FR glass, and it can be screwed in 1" barrel as any other filter. To maximize the scale, I inverted the optical train order (previously, the prism was in front of SA100, now it is at the back - hopefully this will not offset the benefit of smaller star images with FR too much).

[Merlin66]

Look forward to seeing your first light results.
Ken

[bojan]

Last night I spend some time capturing eta Carinae spectrum.

The first results are not good, compared to from 06/02/2020.

1) Scale is as expected (4.4 A/px), that is reduction by 0.83 from previous design, due to smaller grating-sensor distance.

2) Resolution is significantly worse compared to "old" (FR.63 + wedge + SA100 + 450D) configuration. It seems lines are doubled (??!! especially visible at Hb.. I am not sure what is causing this, maybe camera mirror vibration? I must try with mirror lock next time to confirm).

3) It seems that modded Canon 60d (only second filter removed, I left the one with "shaker" because this way it is easier to manage dust) has still UVIR cutoff-like spectral response characteristic.
450D seems to be quite wider in both directions, especially in IR.


Back to the working bench...

[bojan]

Tonight the visibility was much better, and atmosphere was a bit calmer
After some re-adjusting of wedge against grating (rotating)and focusing (on spectrum), the result was better, but I am still not happy.

It seems the re-ordering of optical elements (prism after grating) significantly worsened the coma.

So I will have to go back to previous arrangement.

[bojan]

Problem solved..
I made a mistake when assembling the "new" arrangement.. I mounted the prism with flat side towards the sensor..
Luckily it was easy to flip it and the coma problem is no more. The third image shows the correct setup.

Spectrum is of eta Carinae.

[bojan]

A further improvement..
I added tilter to the assembly (I bought it a long time ago from aliexpress I think, for couple of $), so now the focus is better along the spectrum.

[Merlin66]

Bojan,
On the Christian Buil webpage, he shows his prism mounted to the FRONT of the SA100 (held in place with a couple of dabs of hot melt. (I'll find the link an post it)

[Merlin66]

Here's the original page:
http://www.astrosurf.com/aras/staranalyser/userguide.htm
Also an interesting read on Cloudy Nights
https://www.cloudynights.com/topic/760222-star-analyser-200-grism-configuration/

[bojan]

Hi Ken,

Yes, I know about the order prism-grating.
However, I was wandering how to increase dispersion on the setup (earlier w/o Tilter) by increasing the distance of grating from sensor, and found a discussion somewhere (when I find it again, I will post it here) that the order can be reversed (by mirroring the whole assembly), and indeed I tested this by making mistake (just reversing the order), see here: https://www.iceinspace.com.au/forum/showpost.php?p=1611656&postcount=19


My first results with this setup were on Sat night, Nova Sco (https://www.iceinspace.com.au/forum/showthread.php?p=1617986#post161798 6).


I am also thinking about slightly different setup, like on the attached image.
Basically, I want to convert f/10 beam to parallel, to avoid coma.
This will be my next project..

[Merlin66]

Providing a parallel beam to the grating is a very common solution.
As you show, a Barlow can be used, positioned at its focal length from the telescope focus. Fitting the grating to the front of a camera lens and using a DSLR for imaging has been very successful.
(If you use a positive collimating lens you can then think about adding an entrance slit for even better results.)

[bojan]

Yep.. I am considering that for the future

I have an old Panagor slide duplicator. the mechanic is ideal for that sort of thing... (coronagraph as well), see attached sketch attached).
It has standard M42x0.75 mm threads, so it is easy to change configuration.

Also, I 3D-printed inserts to hold various lenses..

The only disadvantage is length (~300 mm).. but I can live with that :-)

[Merlin66]

I also have one of those old slide copiers! Never thought of using it.
I did use a Chinese eyepiece projection tube with an eyepiece etc.
In you drawing, I can't see the imaging lens behind the grating. Am I missing something?
(See: "Astronomical Spectroscopy for Amateurs". p184-193 for grating/ slit ideas)

[bojan]

No, you are not :-)
This is my older drawing, imaging lens can be added after prism.