Forum der Vereinigung der Sternfreunde

Hi everyone,

I hope I can ask for some guidance regarding building a custom spetrograph for my telescope; a 36"f4.5?

I had found Buil's website with an article titled "Design a classical spectrograph", http://www.astrosurf.com/buil/us/spe2/hresol.htm. This design has many attributes I'd like to take advantage of;
1. R=3000
2. no entrance slit will be used, thus avoiding the difficult problem of positioning the star's image in the middle of a slit a few microns in width.
3. (F/D=4)
As for item #1, if R=3000 is not achievable, the goal, as stated to me, is the resolution to be near 0.5A/pixel. For item #2, no entrance slit is almost a requirement as I'll have alt-az tracking with field rotation. For item #3 my scope is F/D=4.5.

However, this document had an upgrade, http://www.astrosurf.com/buil/us/stage/ ... ign_us.htm, named "Design a spectrograph", in which now a slit is employed. But I'd still prefer not to have a slit....

Could you comment on the slit-less design with regard to my scope so the system resolution is near 0.5A/pixel to useful in the Pro-Am Campaign for Summer 2013, Wolf-Rayet stars WR 134 and WR 135 in Cygnus?

Using SIMSPEC V2.2 and adding my details, the R value appears to be ~1250 using an SBIG ST-8300M with its 3326 x 2504 pixels at 5.4 x 5.4 microns square. The lens details appear to be: f1=120mmf4 collimator and s' - distance between the collimator and the grating - is 158.7mm derived from using the original "Design a classical spectrograph" document. When using a 1200 Groove, 500nm Ruled Diffraction Grating, which is 30mm square, the effective diameter for the beam, d1, is 26.22mm and the 120mmf4 collimator lens seems to allow this.

In the second document, "Design a spectrograph", the collimator camera lens has been replaced with a simple achromat. Is this an acceptable replacement? Some cost savings would be nice.

Any other comments are most welcome and I hope to hear back regarding the questions....
Thanks, Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven,

0.5A/pixel is a suggested dispersion not the resolution. Achieving the required resolution of <1A is the key factor. This is equivalent to R~5000. With a slitless design, the size of the star image at the camera will dictate what dispersion you need to achive the required resolution.

Personally, I would avoid running slitless. It produces more problems than it solves:

Tracking is actually more critical, particularly for an alt az mount if resolution is not to suffer due to smearing
Resolution will vary with seeing.
Sky brightness will limit sensitivity
Intereference from the background will be difficult to remove.
Wavelength calibration is difficult without a fixed slit to a produce calibration lamp spectra and give a fixed reference.

Acquiring and tracking with a slit spectrograph is best solved by using a reflective slit arrangement. Field rotation is not a critical issue for spectroscopy provided the guiding system can keep the star in position


Robin

----- Original Message -----
From: Steven Aggas (fg-spek-convento@vds-astro.de)
To: fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)
Sent: Tuesday, December 20, 2011 8:07 PM
Subject: [fg-spek] Custom Spectrogaph

Hi Steven!

In addition to Robin’s correct remarks you need to keep your focal length in mind. It is about 4000mm. According to the seeing disk if say 1” will result in 20 micron disk on the chip if you would use the same focal length for camera and collimator in the spectrograph. And this 20 micron image represents your resolution element which needs to be covered by about three pixel. Three pixel because two pixel represent the Nyquist criterion (you need two pixel to resolve a spectral element) plus additional coverage due to diffraction effects in the spectrograph. So, I would say, working without a slit should be possible. The pixel size is about 5 microns, right? If yes, you would cover the resolution element with about 4-5 pixel. That seems to be ok but then you would see the seeing fluctuations again. Hence, at least 2x2-binning or even 3x3 is required to “undersample” and keep the disk stable with the stable pixel-size itself.

Second, I fear you will loose light due to your f/4.5 configuration. The beam will exit the slit with the same f-ratio and you would need an f/4.5 collimator. This collimator would, hence, be very big and expensive… You see, a big telescope requires big spectrographs. I have the same problem with my 80cm telescope. The perfectly adapted spectrograph with maximum efficiency is a problem.

You could accept lower efficiency, though, but need to know the result. You should play with the Excel-Sheet from Christian Buil (www.stsci.de/simspec_slit_e.xls) and see what is coming out. Then you introduce it here and we all can give some input.

It might be an idea to use a Barlow lense. It would increase your seeing disk in the focus and you could reduce the collimator size. But then you definitely need a slit.

Cheers, Thomas

An important remark: The ConVento Category is exclusively established for ProAm projects. If you discuss technical issues and other general questions which do not necessarily touch our ConVento campaign(s) then please use the respective threads in our forum. Only the ConVento threads are directly sent to our professional partners by email. We should not overload them with questions which can be answered without them.

Thanks and cheers, Thomas

Hello,

habe mir erlaubt den thread zur Technik zu verschieben.

have allowed me to move the thread to technology.

Steven, you can calculate your spectrograph design very simply by using an excel sheet: http://spektroskopie.fg-vds.de/simspec_slit.xls[/u]

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Perfect, Lothar, I already asked Jürgen to do that. Which is the german version of the english sheet I already mentioned.

Thomas

Just for the record...
The current English version V3.3 contains some corrections to the original Excel macros ( the effects of the slit gap on final resolution)


On 21 December 2011 12:16, Thomas Eversberg <fg-spek-tech@vdsastro.de (fg-spek-tech@vdsastro.de)> wrote:
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
http://tech.groups.yahoo.com/group/astr ... =322612425
"Astronomical Spectroscopy for Amateurs" - Springer
"Grating Spectroscopes - How to use them" - Springer

_________________
" Astronomical Spectroscopy for Amateurs" - Springer

Hi Ken, it seems that you missed the correct thread. Our admins can move it. Very well! I have two questions:

1) Where is that sheet?
2) What in detail has been adjusted?

I ask because the documentation of simspec has always been marginal and the calculations haven't been easy to understand. Klaus Vollmann did that and the result was satisfying. On the other hand, the S/N output calculation was never reliable. It might be that the Planck function in the sheet was not correct. This will not be adjusted by taking effects of the slit gap on final resolution into account. Please understand: Giving information on a new tool is nice. But additional background information is essential to properly use the tool (this is valid for ALL software tools, of course). So, as long as version 3.3 is not reliably tested I will stick on version 2.2 because it works.

Cheers, Thomas

Thomas,
The revision history is shown....
The latest correction is quite complex in that the basic formula was correct but the terms were not correctly bracketed within the macro and this led to errors.....
Hope this helps
ken


 
On 22 December 2011 09:37, Thomas Eversberg <fg-spek-tech@vdsastro.de (fg-spek-tech@vdsastro.de)> wrote:
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
http://tech.groups.yahoo.com/group/astr ... =322612425
"Astronomical Spectroscopy for Amateurs" - Springer
"Grating Spectroscopes - How to use them" - Springer

_________________
" Astronomical Spectroscopy for Amateurs" - Springer

Thanks Robin and Thomas,

I had played with SimSpec before posting the first time but I'm glad there is an english version.... My German is really bad, my French is worse.

I had started to customize my copy by including an optical layout diagram to help me better understand. I could also then put real values there since optics rarely come in the calaculated values.

The V2.2 copy attached is titled: simspec_slit_e for 36 inch.
Please look it over and comment? There's bound to be errors that you may find them long before I would.... The third and fourth tabs contain the optical layout relative to the scope, with and without a slit. My preference for slitless is due to the potential of using the Zero Order for finding/positioning with the use of a degree-scribed optical base I have.

The V3.3 copy attached is titled: simspecV3.3 for 36 inch.
I transferred all the pertinent information to it to see the differences, if any, compared to V2.2 with regard to SNR, Resolving power, Dispersion. Is my understanding correct, when using a cylinder slit, in needing yet another camera for finding/positioning? The slit function does improve R. What value could change to improve R to the point of usability for the Pro-Am Campaign for Summer 2013, Wolf-Rayet stars WR 134 an,d WR 135 in Cygnus, without resorting to a slit?

I understand Robin's list of drawbacks and why they would occur in a slitless version.
*To include a reflective cylinder requires some other method then, of finding the right star and positioning it correctly. Some comments on how would help. The scope should have good pointing to within a few arc-minutes of center, more or less, during the GoTo, then there is a need for fine positioning
*Tracking would have field rotation and then only keep that star on the cylinder it it happen to be the one chosen for tracking purposes. The ST-i is envisioned as the tracker using the 100mm FL f/2.8 lens that can come with it for a stand-alone system. The 35.7mm diameter optics may not always reach my spectroscopic target star's magnitude.

So how far off am I in this spectroscopic endevor, the spectrograph design I've shown, and systems integration in the method of finding/positioning/tracking?

Thanks, Steven
Sorry for posting it in the wrong forum but my German is bad....

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven,

Not a reflective cylinder virtual slit (very inefficient) but a slit in a mirror. See here for an example of a spectrograph using a mirror slit system.
http://astrosurf.com/thizy/lhires3/index-en.html
(The LHIRES III littrow spectrograph designed by Christian Buil). The slit is conventional but with reflective jaws (or in the latest design in a mirror) so you can guide on the surrounding field or on the overspill on the slit. The slit be built into the spectrograph or as a hole for a fibre optic guide head. Similar systems are used on all the Shelyak spectrographs (Shelyak will sell the reflectve slit on its own to self builders) and in the Baader DADOS and SBig SGS. It works very well as the guide camera sees exactly what the spectrograph sees without the flexure or alignment issues of a separate guidescope for example.

Cheers
Robin

Here is an example of the guide camera view of the slit and surrounding field (on LISA spectrograph)
http://astrosurf.com/buil/supernovae/2011/_sn2011dh.png
from this page on Christian buil's website
http://astrosurf.com/buil/supernovae/2011/obs.htm

Cheers
Robin

Hi Steven, I took a quick look on the sheets. They both give more or less the same output. Only minor negligible differences occur.

You have two basic choices:

1. Without slit: Then you could use a collimator of 300mm focal length and a camera of 200mm. An f/4.5 300mm collimator is big and, hence, expensive. And you need a 83mm grating - expensive. The camera would be bigger, as well. So, for about 1 Angstrom 2 pixel resolution you would spend some (a lot of?) money.

2. With slit: You could use a 180mm collimator and could then use a standard photo lens size and a 50mm grating. The collimator and camera could also both have 200mm (Littrow configuration) and you would still lose little light at the grating. With a slit of 20-25 microns you would match the 1 Angstrom resolution requirements for the campaign.

2x2 Binning is essential for both choices. And the calculation is for 2" seeing, which means a 40 micron disk. You told me that you can count on 1" at you site. So, if you have 1" then you almost get full light throughput. Your site is fantastic! Please check that out in detail, especially the loss potential of light and check if I made a mistake (Robin and others?).

Cheers and all the best for Christmas to all, Thomas

Thanks again Robin and Thomas!

The links to a Shelyak reflective slit, and views through the guider, are most helpful to know what I can expect in my eventual system. The Orion Starshooter with its 1280x1024, with 5.2um pixels, would give me a 4.5'x5.6' fov, more than enough to find, position, and then track (though I don't know if tracking is possible with it on the halo remnant of the target star surrounding the slit).

The option of the slit really is a requirement.... The additional autoguider and slit increase the costs but the R-value is much better. The lens selection increased as well, with reduced lens costs by going to 180mmf4.5.

I had originally discounted the 180mmf4.5 collimator camera lenses because they violated the f1 rule causing a larger exit pupil than the 50mm grating could accomodate, but after plugging in that value, using a 20um slit value in V3.3, I noticed the R value increased by 38% by exceeding the grating by only 0.1mm. Not a bad trade off.... In essence, I think this would only affect the outer edges of the spectrum swath as seen by the ST8300. The pivotable grating means any portion of the spectrum is viewable, just not at that time. With the 20um slit, the R is ~6554 at 6563A with an angle of 26.3975 and an R of ~5252 at 5500A at an angle of 24.6359. I see the Neon lamp is used for the Ha line calibration, but for the 5500 range is a sodium source better for dialing in the range containing the HeII 5411A line in WR134 and the CIII 5696A line in WR135 by using the D lines?

I have fiberoptic cable. If pointed at the primary would the Neon or other lamp light be usable or is there some other way? How would the calibration light be introduced?

A question regarding V3.3, should the 'Spectrum size/spread' values be tied to the binning amount; 1, 2, or 3 (1, 2x2, 3x3)? A note on the 'Width of Spectrum in Pixels (n)' state '9' should be used, is this a 3x3 configuration where the others look at only one axis of the binning?

Between the two versions I uploaded, the scope details were nearly identical and outputs close, but to get the SNR similar (both not using a slit) the exposure times are vastly different.... Do those exposure times make sense? I've found computers are very good at making humans look really bad at math order of operations. The noise equation has some addition along with multiplying and deviding but no parenthesis, the signal has no addition so no worries there.

Happy holidays. I'm off of work over the holidays so I have time to spend on what parts are needed and how these parts would fit together.

Thanks again for the help.
Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Steven,
the width of the spectrum is the binned height of the spectrum....it will depend on your final set up, the size of the target star and any aberrations in the system.
The SNR is based on the algorithm laid out by Buil in the web link given on the spreadsheet.
Hope this helps.


On 23 December 2011 23:41, Steven Aggas <fg-spek-tech@vdsastro.de (fg-spek-tech@vdsastro.de)> wrote:
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
http://tech.groups.yahoo.com/group/astr ... =322612425
"Astronomical Spectroscopy for Amateurs" - Springer
"Grating Spectroscopes - How to use them" - Springer

_________________
" Astronomical Spectroscopy for Amateurs" - Springer

Thanks Ken. I found one SNR formula in Buil's pages but it's not as complex as what's in the .xls spreadsheet. I'll just leave that alone for now.... I did change the atmosheric transmission value from 0.8 to a scaled value. Looking online at some graphs it's only 0.8 at ~1um, but closer to 0.66 at 6563A and lower at shorter wavelengths. I did notice that the spreadsheet starts to give negative collimator values as you get more blue. I will be able to rotate the grating to view the blue end and it's only an artifact of the spreadsheet, yes?

Regarding grating mechanical mounting, I have a micrometer rotational stage. I'd like to attach the thick grating to it, much like Buil did with an optic clamp, but can it be put on its edge and still stay perpendicular to the incoming beam or does that require a tip/tilt behind the grating on top of the stage? Like what's seen on 'Grating Holder' jpg.

Looking at the angle Alpha when plugging in different reference wavelengths, the angle decreases when decreasing the wavelength. However, the two diagrams are different on Buil's site. The picture Dispersion1 is correct, yes, (and not Dispersion2)? I ask because my micrometer stage increases its value as it rotates clockwise (apparently decreasing Alpha) when looking down on it but I'd prefer the value to increase as I dial in a longer wavelength segment. It seems I would only need to mount it and the optics on the 'roof-plate' that isn't shown on Buil's pic where he's attached them to the 'floor-plate'. On my scope it'd all be on the left plate closer to the primary not on the right plate.

Would anyone have a good picture of the Shelyak Instruments 4-Position Slit (15/19/23/35µm) support? This device is at a 45 degree angle to reflect all but the slit to the guider and slit passes through to the collimator? Does it then move along the plane of the 45 for each slit width to be in place as needed? With a good picture I could ask far better questions. With ~1.5" of back-focus available and the need for a 1.25" focucser for the guider, the possible need for a micrometer linear bearing to move the slit along the 45-plane, I've got to make it all fit....

I also have a couple of matched optics holders that may work well for the collimator, plus I have that rotational stage, but the plan for the 200mmf3.5 camera lens is a couple of custom lens holder rings so the bayonnett mount protrudes out of the enclosure to be able to attach the ST-8300. This way I can detach the main camera and do other things with it. Besides the obvious caviat that the lens can't move during the twisting action, is there something I'm missing that I should be aware of? As long as it's set to infinity and locked in place at the prescribed distance of 255.42mm from front lens to the grating, I'm good, yes?

Hopefully the descriptions are understandable. If I've understood and the descriptions are correct I'll make fewer mistakes resulting in many unused holes....
Thanks,
Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven!
Don’t worry, standard freeware execute that without problem.

I presume that the slit has no alternative. But the cost effect is not as dramatic as expected. There are some low cost solutions as, e.g., as slit out of the blades of a pencil sharpener (successfully done in our group) or even a non-adjustable slit (as the Shelyak device). Lenses can be found at Ebay for really small prices.
Yes, now you learn how to deal with your instrumentation. The calibration issue is indeed important to deal with. There are a number of different calibration lamps possible. This is an extra point to count for. The calibration light should pass the spectrograph as the stellar light to avoid glitches. Best would be a source right in front of the slit on the entrance side.

The S/N issue in simspec is somewhat weird and I do not count on it but on my personal experience. I am absolutely sure that you easily will acquire an S/N of 100 for a 8mag star within less than 1 hour.

Cheers, Thomas

Hello Steven,

The slit has a angle of about 10 to 15° compared with the focus plane of the telescope, not more ! When the angle is too big then one of the edges of the slit ist not more in the foci of the collimator, telescope and guiding camera. Then you see the star on the slit unsharply.
You can see in the picture the slit in the first version of Lhires III. The slit has to be oriented perpendicular to the gradient of its inclination.

You use 2 foto lenses. I think, that makes the spectrograph havy. Better it would be to install an achromat for collimator.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Thanks Lothar, that is most helpful!

I've not been able to identify a simple achromat to replace the 180mmf4.5. F4.5 is not common.... Do you know of one?

To all;
I started to change V3.3's static values in some of the boxes to formulas I found in Buil's site, even adding a box in the telescope system section for the distance between the secondary and the focal point which affects:. "Distance from Camera lens to grating(T) :", which in V3.3 was a 30mm static number but according to Buil it's this formula s'=(f1*s)/(s-f1), where he states "s is the distance between the telescope secondary and the collimator, if f1 is the focal length of the collimator lens, then the distance s' between the collimator and the grating must be approximately" so I changed it. With 609mm as 's' and the value spit out for 'f1' I'm looking at ~255mm..., where T=s’ (and the optical diagram shows this as ‘X’ because they didn’t always use the same names between Word and Excel (and pictures). So, too, for the camera values (Camera lens-FWHMc) and others, those are static too. I think there’s a need for V3.4…, no static values except for like I did surrounding the optical diagram with real-world values because a lens might not come in a 172.935mm but I have to choose between a 120mm or 180mm in a f4.5 cone.

So, I made a V3.4 and attached it. It has some conditional formatting in the “Applied Results” where one has to enter the real-world collimator lens value, specifically with regard to the grating size, because of how the collimator lens affects the grating size required (you’ll see it red because 50.1mm grating is required whereas I don’t want to buy anything larger than 50mm so it’s highlighted in red as a point of reference to accept or correct. I think that overflow of light past the grating creates a slight loss of primary mirror but the R value increases by having that lens type. The other formatting added, is the minimum required camera lens diameter. The 200mmf3.5 works whereas a 120mmf1.8 does not. User Inputs are also clearly marked as well as a different color coding for the provided input data and calculated outputs. One should only have to put proper values in the yellow boxes.

Could another pair of eyes look it over? My goal was to not miss some important input that needs to be relative to the input scope design. There is still an outstanding question of the grating angle and its affect on making the grating size negative when calculated for 4500A, and the collimator lens value goes to -180mm too but stays positive for 5500A and 6563A.
Wavelength: Grating Angle, Grating Size, R Value, Collimator FL
6563A: 26.3975d, 50.15mm, 7027, 180mm
5500A: 24.6359d, 47.2mm, 5754, 180mm
4500A: 23.2591d, -45.15mm, 4615, -180mm

I adjusted the grating angle to maximize the 180mmf4.5 collimator and therefore R-factor. Is there a way to know that I’m actually at the max R-factor grating angle (like 6563A is centered in the ccd?)?

Thanks, Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hello Steven,

http://www.edmundoptics.com/products/di ... c#products
there is a 160 mm D=40mm achromatic lens

http://www.qioptiq-shop.com/Praezisions ... esser.html
here exists also a 180 mm version: Achr. VIS ARB2; D=40; F=180 € 110,00

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Thanks again Lothar. While weight is one consideration, does a camera optic have better spot size at various wavelenths, or an achromat? What redeeming qualify does a camera lens have over an achromat to accept the weight it adds? Are there pros and cons for each?

In comparing V2.2 and V3.3, I found static values had replaced some formulas. Since I don't know which was right I did favor the formulas as it then shed light on the parts involved, and a static value may have been inserted to a specific value for which the primary and focal lenght were entered, much like my primary mirror diameter, FL, and camera data was entered. I'm still learning what parameters are rules and which are mere suggestions.... I'd rather fight with the formulas than have someone design the spectrograph for me and not know how they arrived at that configuration. Are there some formulas, like the one to determine the grating-to-camera lens distance, that are better served with static values over prevoiusly used formulas?

Thanks Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven

A camera lens in general has less aberrations than an achromat. If the camera is not too fast (not faster than F/4 or F/3.5) an achromat will do a reasonable job, to reach the necessary resolution. The most important aberrations are chromatic, spherical and surface curvature. Spot diagramms are given by the manufacturer, and musst be smaller than the image of the slit on the detector. The aberrations could also be calculated with an ray tracing software (WinLens 3D Basic as a freeware). Camera lenses for fotography are only corrected for the smal visual, so the cromatic aberrations is greate to the ends of this wavelength. In fotography they use an IR-UV filter to block these wavelength which would result in poor image quality.

Because it is (mostly) not possible to cover the whole spectra with the detector the chromatic aberration affects only a smal region of the spectra. Therfore it is easier to bild a high dispersive spectrometer than a spectrometer with covers the wohle (and short) spectra.

The distance between grating and camera is set by the designer, because it is mostly given by geometrical facts and dimensions of the used components. The necessary diameter of the camera must be calculated and this value must be used to avoid vignetting.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Hi Steven,

but you use the foto lens not in the direction for wich it is constructed (and corrigated). You introduce the light from the backside. The foto lens cannot play its capabilities.
And the spectrum sector is small, only <100 nm, therefore chromatic aberrations don't play a role.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Hi Lothar! I don't understand what you mean here. Do you mix up collimator and camera? About what lens do we talk here?

Cheers, Thomas

Yes, I talk about the collimator.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Thanks Daniel!
How important is it to have an achromat that matches the input focal ratio? My scope is f4.5, must the achromat also be actually f4.5, or just close to it like an f5, or f4? I would ask about the use of a coma corrector for improving aberations but that chews up half of my focus in-travel needed for the guiding optics and I'll need every millimeter for those parts prior to the focal point/slit....
With regard to the grating camera lens distance, what geometrical facts and dimensions of the used components should I account for? The camera lens should be as close to the grating as possible without blocking collimator to grating light as a minimum, and, Buil has a formula that makes that distance rather large (which is right?), but are there other facts to consider?

Thanks, Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven

The collimator should be as fast as the telescope but not slower. In your case I would consider an achromat some faster than F/4.5, so all the light is collected by the collimator. This f-matching is very important in terms of efficiency.
A coma corrector should not be necessary becaus the spectrometer works on the optical axis of the telescope (the slit is centered on this axis).

Ruled gratings are blazed, which means, that there is an angle at maximum efficiency, see the "Diffraction Grating Handbook" from Newport for reference:
http://gratings.newport.com/handbook/handbook.asp

The grating should work near that angle (this isn´t possible exactly) and the angle between the incident and diffracted beam should not be greater than say 20°.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

If I use a 160mm focal length f4 achromat collimator like Lothar suggests (actually f4.1 because the manufacturer states its 39mm clear aperture), Edmond Optics Edge Blackened pn: NT47-740-INK, I have an exit pupil of 35.55mm which is no problem given the 39mm clear aperture. At a total angle of 28.5d, like Buil used, (Alpha of 38.2d and Beta of 9.7d), at 6563A it says I'd need a grating of at least 45.3mm. So I'd pick a 50mm grating (which the manufacturer states has 90% clear aperture, so ~45mm clear aperture but probably has more usable area than that). The minimum d2, imaging lens, clear aperture required, according to Simspec is 44.6mm. If I use a 200mm focal length f3.5 imaging lens, with its clear aperture of 57.1mm, it should work, yes?

The R factor for the above configuration is ~7400 at a spectral resolution of 0.89A and dispersion of 0.44A/pixel, using a 20um slit, and binned 2x2 using 5.4um pixels with 2 arcsecond seeing. Then the question is (besides the question of being able to squeeze a slit and guiding optics in the ~1.5" in-focus travel) what distance the 200mm lens should be away from the grating (besides the minimum needed not to block light exiting the collimator lens headed towards the grating)?

How does this system look overall so far?



The Edmond Optics 1200lpm, 500nm grating, pn NT43-006, has an absolute efficiency above 50% from ~3750A to ~7000A, and, because of the 2x2-binned pixels of the ST-8300 I'm only imaging ~500A wide at a time. The grating only needs rotating to some lesser angle to bring other 500A ranges besides the 6563A data mentioned above into view, yes? This system is good for other wavelengths besides 6563A, I hope, but at a lower R-factor (by changing the total angle from 28.5d to 29d for 5500A with an R~6140, and, 30d for 4500A with an R~5000 (all staying below a spectral resolution of 1A and 0.5A/pixel))?

Thanks, Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Steven,
The total angle is alpha-beta (beta being negative) = 38 degrees - this is still OK.
The distance from the grating to the imaging lens is constrained by the physical dimensions of the lenses...if you layout the components for "minimum" total angle you'll quickly see the results....


On 30 December 2011 23:05, Steven Aggas <fg-spek-tech@vdsastro.de (fg-spek-tech@vdsastro.de)> wrote:
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
http://tech.groups.yahoo.com/group/astr ... =322612425
"Astronomical Spectroscopy for Amateurs" - Springer
"Grating Spectroscopes - How to use them" - Springer

_________________
" Astronomical Spectroscopy for Amateurs" - Springer

Hi Steven! Ken is basically right. Just to clear up the details for you:

The total angle is always . The reason for the difference is the definition of the coordinate system. In fact we have .

The direction angle depends on the angle of incidence and the diraction angle changes by (see image attached)

If we change by , then changes by . The alteration is proportional to . The proportional factor is called the anamorphous factor A. Even would work but on the costs of lower resolution. To keep the anamorphous factor possibly independend from the grating position, should be kept small.
This is a matter for the spectrograph designer. In practical sense I would say it vice versa. The physical dimension of the camera lens is constrained by the distance from the grating to camera.

From my point of view these consideration differences are important especially for beginners. It is easy to mix design problems up and an accurate description helps avoiding those problems.

BTW Ken, you ended up in the wrong thread.

Cheers, Thomas

Hi Steven,

a drawing of the system is attached.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Thanks guys,
So my components will be laid out on a 28 degree angle like Daniel's attachment or 28.5d like Build showed? But neither optic should block the others' light, and that's what might affect the imaging lens chosen and therefore affect the R-factor if the 200mmf3.5 can't be used. I've ordered that lens since it wasn't pricey to see what can be done. Is there anything else in the parts I described that is concerning?

One thing I've considered is placing the slit, guiding optics and collimator on a motorized micrometer linear bearing (since the collimator to grating distance is apparently not too critical) in order to have the spectrograph focus independent of the telescopes' focuser. This would alleviate having all that weight hanging on the focuser alone at various elevations/angles. To prevent flexure I'll have three mounting attachment points connecting the spectrograph to the telescope; the upper and lower cage rings and a third on the lower 2" OD aluminum cage tube (plus the focuser for four). The focuser would be racked in all the way and only provide the initial alignment point for loading the device on the scope, then the other points are attached and secured.

Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Guys,

Here's a first attempt at an optical diagram with proportionately sized parts, dimensions, and an angle..., relative to the 36" scope. It's done in Excel because I've drawn so many things using the shapes.... The shape of the 200mmf3.5 is approximate, slightly larger than the clear aperture value. I think I've identified all components.... I used Daniel's and Buil's picture for reference.

Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven
The collimator to grating distance isn´t critical, you could make it as big as you want.
The distance betweet collimator and slit is very critcal, because the resolution will be affected.
Do you know how to adjust a collimating system (collimator and slit)?

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

If the collimating lens is 160mm focal length....then this is the distance between the slit and the lens not the 200mm+ you show....

On 3 January 2012 18:40, Daniel Sablowski <fg-spek-tech@vdsastro.de (fg-spek-tech@vdsastro.de)> wrote:
--
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
http://tech.groups.yahoo.com/group/astr ... =322612425
"Astronomical Spectroscopy for Amateurs" - Springer
"Grating Spectroscopes - How to use them" - Springer

_________________
" Astronomical Spectroscopy for Amateurs" - Springer

Hi Daniel, maybe not.... Other than that I placed the collimator at the distance from the slit to fill its aperture (~39mm clear aperture according to the manufacturer) with angled light of an f4.5 cone that passed through the slit at the scopes' FP. This distance is ~215mm to fill the 39mm optic whereas the collimator's FL is 160mm, but maybe it should be set at 160mm distance. Maybe the use of an f4.1 v. a matching f4.5 collimator then spreads the light to ~45.3mm at the grating, according to Simspec, instead of the 39mm if the light were parallel from having used an actual f4.5 collimator that matches the scopes f4.5?

160mm not 215mm?

Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven,

just two points. As Daniel already mentioned, the distance between slit and collimator is critical. The slit has to be at the back focal plane of the collimatot, so that light exiting from the collimator is exactly parallel. Otherwise you will have astigmatism at the focus of the camera and loose resolution. I would provide some means for focusing the collimator to compensate for temperature changes and chromatic aberration of the lens.
The slit has to run parallel to the grooves of the grating, perpendicular to the drawing plane. The guide image will therefore be projected out in the drawing plane.
The design has a nice feature. The small angle beta means that dispersion, which varies with cosine beta is almost constant over the whole wavelength range, which simplifies analysis of the spectra.

Regards, Martin

Thanks Martin,
That my scope is f4.5 and the collimator f4 will this cause the light exiting from the collimator not to be exactly parallel? An f4.5 collimator should be perfect but is f4 ok with regard to astigmatism?

With regard to the drawing, you're saying if I rotate the slit pictogram by 90 degrees ccw it would be correct relative to the other parts drawn? The guide camera is behind the plane and the slit reflection light is pointing into the guide camera?

I do have another linear bearing that the collimator could be put upon. This would allow for independent focus of the slit/collimator.

Thanks for your feedback!
Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Thanks Ken, at this point I've misplaced or misorientated every part at least once....

In this drawing the slit is rotated on the axis of the slit opening to 15 degrees. Is this one correct? I hope so, there is no interference with the guide camera and a focuser knob.

I've added the relay lens for guiding....


Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven,

now the slit and guide camera look ok in the last drawing. That the collimator is a bit faster than the telescope does not matter, much better than the other way round. A spectrograph is always a compromise betweeen resolution and efficiency, with your design you seem to be on a good track. Just make sure that you focus the collimator correctly. If you want to use a different grating some day (with larger dispersion) you may be happy to have chosen a slightly smaller beam hitting the more inclined grating and not losing light (I did not check your angles). If you are a perfectionist you could still extend the collimator focal length with a weak Barlow lens to get a match between telescope and collimator f-ratio, but I think your choice so far is very good.

Regards, Martin

Hi Steven

To reduce bending and to improve mechanical stability, it would be better to use a 90°-mirror after the slit (see my drawing).
The light after the collimator musst be parallel. To adjust the collimator: Use a smal telescope (finder scope) and focus it on a star. Then look from the direction of the grating in direction to the slit through the collimator an chance the distance slit - collimator till the image of the slit in the finder scope is perfectly focused. This distance musst be fixed for all times...

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Hi Daniel, I'm not sure I can fold it, given that I would want the device to focus independent of the telescopes focuser. The telescope focuser would not be used except for insertion of the 2" nose piece of the spectrograph. The orientation of the spectrograph would then allow for a left, right, and bottom support beam from the spectrograph to both upper and lower cage rings and one of the 2" OD cage tubes. This way nothing is hanging on the focuser and the three brackets go with the spectrograph when not in use.

Thanks for the tip on adjusting the collimator! It was mentioned that the collimator should be adjustable for temperature compensation and the chromatic aberation of the lens. How critical are these to compensate for? Must it be adjusted to properly use the device on the red v. green portion of the spectra?

On a different note, one of the slit holders, the LHIRES III, is designed to hold the slit at 10 degree angle while the other, LISA, is at 15 degrees. The LISA has two 19um slits on either side of the 19um hole and the LHIRES has a 19um slit, so, 10 or 15 degrees does not seem to be ultra-critical. What is the maximum angle that a 19um slit can be placed at; 20, 25, even 30 degrees, or more, before bad effects are noticed?

The reason I ask is that I have limited space to place the mirror that would divert the slit-reflected light towards the relay lens and on to the guiding camera. My drawings are currently using the LHIRES slit with its 15 and 19um options at a 15 degree angle but a 20, 25 or 30 degree angle would surely help things out.

Thanks for any input you have.
Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven
Aluminium has a thermal expansion coefficient of a = 2.38*10^-5/K at 20 °C, by a collimator focal length of 180 mm and a temperature difference of 40 K you will get a difference of:

2.38*10^-5 * 40 * 180 mm = 0.169 mm

This is in the range of the chromatic shift of a achromatic lens and it could be possible that the collimator has to be readjusted every summer/winter.
To adjust the collimator it is enough to have a elongated hole in the collimator holder for linear adjustment.

The inclination of the slit should not be greater than 15°. I always try to get 12 °, and this is almost possible.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Note:

In most cases it is possible to focus the camera (temerature and chromatic effects). And this is also more easier, because the camera needs to have a fine focuser unit.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Thanks Daniel,

The thought right now is to use a USB servo motor board capable of driving 4 motors; http://www.active-robots.com/controller ... motor.html. The items currently scheduled for pullies/gears are; 1: (slit/guide optics/collimator) as a group, then 2: the (collimator) alone, 3: (grating angle),and, 4: the (imaging lens focus), most of the things you've mentioned.

The master plan involves a wireless network to control the scope movements, imaging camera, spectrograph/guider from a control room with a view of the observing area, which includes a 7.5' dish antenna, and with automation I don't have to climb a 12 foot ladder for minute adjustments.

Here's the latest drawing updated with, I believe, more accurate portrayal of components including angles. If there is a wrong angle, or some other detail not correct, could you let me know?

Thanks, Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven

Do you know the holding torque of the servos? The mechanical stability of the system ist very important, therfore I would prefer a complete mechanical solution instead of an electro-mechanical one.
The adjustmend of collimator and camera is needed only two or three times, so it could be done by hand.
It is of course an interessting craft project, which I do not want to dissuade you. But this project seems very complex and difficult.

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/

Thanks Daniel, I'll look more into the motors but they will be turning the micrometer knobs of the linear bearings instead of me. Everything should be as mechanically stable as if I had been the one to turn the knobs, though the imaging lens will be having its barrel rotated. (I found this done on the web a few years ago and they thought it worked quite well for remote focus of their camera lens.)

As for the complexity and difficulty, I'm only adding servo's to the mix of complexity. Do not others have to design their spectrograph like this too, with this much complexity? Is this one more complex than others?

Could someone provide feedback on angles within the drawing? If the model is reasonably accurate I can start procuring parts and saving money for the ST-8300. I already have enough micrometer linear bearings, optical clamps and some sheet aluminum. My goal is to know how to use it by summer of 2013.... What thickness would you recommend for the base flooring to which all components/clamps will be attached?

Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Hi Steven,

you have very fine adjustment tools...but they are heavy and oversophisticated. As Daniel said a more simpler adjustment is better to save weight: Adjusting and then fixing the collimator by screws or such simple things. But mechanically stable.

As base plate you can use a Aluminium plate (5 mm) oder high density plywood (marine, 10 mm). Very fine are low density plastic plates with low thermal expansion coefficient:
http://www.obo-werke.de/index.php?optio ... &Itemid=67
They are also very easy to work.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

I calculated the weight for having used 6mm aluminum plate for the base plate, the optics, ST-8300 and the guide camera, etc, and the total is approximately 8Kg, ~18 pounds. The fraction of the total which is linear bearings is 5%, ~7% if I add servos to the linear bearings. This is actually better than my first estimate of 20 pounds, or 9 Kg, for the spectrograph.

One requirement of this project is that it be remotely adjustable. I did not want to be on a 12-foot ladder with a laptop to look at the results of adjusting micrometers. So motorized micrometers are needed....

So it was mentioned that with a seasonal temperature change, Summer v. Winter, the collimator could be off as much as 0.169mm with a 40K temperature difference. It was also mentioned that this might be accomodated with the focussing of the imaging lens.

To prioritize the importance of pieces needing adjustment, is this the proper order?

1. Focussing; {Collimator/Slit/Fold Mirror/Guide Camera} micrometer linear bearing. (the whole assembly moves together to focus the spectrograph and the scopes' focusser is not used. At 8Kg you can see why I wouldn't want to focus using the scope's focuser....)

2. Wavelength Change; Grating micrometer rotational stage.

3. Imaging lens focus.

4. Collimator-Slit distance. (does this really need to move to adjust for the 40K change of 0.169mm? Can it be set at the mid-temperature point like in Spring or Fall so at most it's 0.085mm off and then the remainder is compensated with the imaging lens focus? In the grand scheme of things is this an ultra-critical distance to compensate for?)

At 8Kg, the spectrograph is 1% of the mass of my telescope. Adding motorized linear bearings means I can stay safely on the ground to adjust/use the device 13 feet in the air.

Thanks for the feedback on the adjustment priorities....
Steven

_________________
Apache-Sitgreaves Center for Astrophysics
http://darkskyobserving.com

Ok Ok Ok Yes No

_________________
Daniel P. Sablowski

https://www.kunstmann.de/buch/axel_hack ... 42000/t-2/
https://dpsablowski.wordpress.com
https://github.com/DPSablowski
https://www.aip.de/mitglieder/dsablowski/