Forum der Vereinigung der Sternfreunde

Hi, following an email exchange with Thomas Eversberg he encouraged me to join this forum as UK based echelle users seem very rare!

I have built an echelle with a similar optical layout as the Shelyak Instruments eShel but it has an achromatic 100mm collimator and Cannon 100mm camera lens, This provided a better fit to optics of my F5.4 Newtonian (via a 50 micron fiber from a Shelyak Guide Head) and FLI ML8300 CCD. By the standards of some of the instruments on this forum it is very crude in its construction but it works well and its results match reasonably well the predictions from "simechelle i.e. R ~7000 to 8000".

However, I am getting a new 400mm F6.8 corrected DK (from Orion Optics) and they have optimised it to have a 20mm corrected field and minimised longitudinal chromatic aberration on axis. Because of this I am planning a new echelle to best match its capabilities. The aim is have a similar performance to a Lhires III but across the whole visual spectrum. The design is in its infancy but I have had the guide head modified to have a 75 micron hole reduced by 1.5 to match the 50 micron fiber and will use a 200mm collimator. I will be using "simechelle" and Thomas's and Klaus's excellent book to guide me and it is this that led to the email exchange.

I hope to use the f/4 = f/4.3 exit beam from the fiber without slowing it down thus avoiding the need for image slicing. I realise this will require a larger grating (50 x 120mm) and high quality optics but I think it can be done.

All advice and ideas welcome. I will update this thread as I make progress.

Regards Andrew

PS I am sorry I have no German.

Welcome Andrew,

hmm... I'm wondering why you use a corrected DK? A standard DK is perfect on-axis without any aberrations! The chromatic aberrations come from the refractive corrector for the field. However, using a fibre, there is no need for a corrected field of view!

When slowing down the beam exiting the fibre, there is no need to use an image slicer first of all. You can use a collimator with a longer focal length to compensate the larger object size. However, you end up with the same beam size but with lower efficiency (transformation optics will eat up some photons).

I would recommend to think about an R4 echelle grating instead of an R2 grating. This will allow to reduce the beam size by same resolution. However, a larger grating would be required. Look at Richardson, they have some overstock from time to time.

Furthermore, what seeing conditions do you have? It may be advantageous to increase the speed of you 400 mm CDK. But focal-ratio-degradation measurements of the fibre you are using would be necessary. They are from Shelyak?

A direct linked echelle with R20k is possible by using very simple imaging optics (beam size 12 mm). The chromatic aberrations and field-curvature are the most limiting factors here. With a Petzval lens of two achromatic lenses, I got some very good results with beam sizes of 40 mm and R35k over the visible using a 50 micron fibre.
However, for higher resolution at beam sizes of 50 mm, it is hard to correct over a wide wavelength range with off-the-shelf optics.

kind regards,
Daniel

_________________
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 for your input Daniel,

I did look at a DC (uncorrected) but wanted a reasonable corrected field to allow me to automate the the centering of the target. With my low resolution work with a 300mm F5.4 Newtonian on my Paramount ME I used it's T-Point facility and a custom VB script to acquire the target on to the slit and take the spectra automatically. I hope to do the same with the new scope but with it's longer focal length I need a 15mm field to ensure enough stars to ensue a plate solve when away form the plane of the galaxy.

I will look at your design ideas I has missed increasing the collimator length to compensate for the increased beam size . Obvious once you mention it!
I will also look at your other points. As to the fiber you made it for me! So I am sure it is very good.

Thanks again Andrew

Daniel, I have looked at your point on increasing the focal ratio before the spectrograph collimator and I think I understand the trade-offs. What is the preferred way to increase the focal ratio? If I use two achromatic doublets then the longer focal length lens will amplify the chromatic aberration before it.

Unfortunately, I can't get a response from the Richardson web site so still need to look at R4 v R2 gratings.

All thought welcome.

Regards Andrew

Hello Andew,
did you have considered to use a small refractor for rougly centering your target and then to perform the final centering with the metallized slit?
Regards Christian

I have done that in the past but I want to run as fully automated as possible. I can, just with T-point, get the star close to the field center but the slit it'self was not quite centered so I needed to plate solve to calculate how much to jog the mount to get the star on the slit. This way I could get spectra of 10's of stars a session without manual intervention while looking for candidate Be stars. Theirry Lemoult and I manage to look at some 100's of star in a few weeks. His system was more fully automated then mine.
Our results are here http://www.astrosurf.com/aras/be_candid ... idate.html.

Thanks for the idea.

Regards Andrew

Hi Andrew

There are several ways to increase the focal ratio. An achromatic doublet is for many cases a good choice. But you can also try to use two triplets, e.g.:

https://www.thorlabs.de/newgrouppage9.c ... up_id=5370

Or you use two off-axis parabolas, but take in mind that they are very sensitive to misalignment and there is only suitable imaging quality on-axis.

http://www.edmundoptics.de/optics/optic ... _s|*C5137I

There are also aspherized and real aspheric achromatic lenses on the market as well as UV-NIR corrected triplets.

http://www.edmundoptics.de/optics/optic ... nses/2953/

http://www.edmundoptics.de/optics/optic ... nses/3405/

http://www.edmundoptics.de/optics/optic ... nses/2479/


In any case, it is necessary to ray-trace the performance. Do you have a tool to do this?

Thorlabs offers an R4, but only with 79 l/mm and with 25x50mm²

https://www.thorlabs.com/thorproduct.cf ... E2550-0875

kind regards,
Daniel

_________________
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 for the comprehensive reply and the links Daniel.

I have access to the Educational Version of OSLO which allows me to model up to 10 surfaces. That should be enough even if I can't model the whole set up at once.

I have eventually got on the Richardson Grating site. It seems my IP address was in a block they blocked due to improper use! I finally got it fixed via the Newport site support.

Unfortunately they don't seem to have any echelles in the overstock list at present. I have inquired about the cost of some echelles via their UK agent (we are not allowed to go direct from the UK) but it seems they take a long time to respond.

While working through the options I have been thinking of improving my existing echelle instead by improving the S/N by getting a better camera lens (less chromatic defocus at the extremes) and using a camera with a 90% QE rather than the existing 50% QE.

At least playing with ideas does not cost much other than time.

Thanks for your help and advice.

Regards Andrew

Hi Andrew,

there are overstock echelles, e.g.:

http://www.gratinglab.com/Products/Over ... uffix=414E

Cick on the little red cross at the catalog number in column one:

http://www.gratinglab.com/Products/Prod ... es/T3.aspx

kind regards,
Daniel

_________________
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 must be going blind!

Regards Andrew

After much delay I have completed the rebuild of the R~8000 spectroscope. It seems to be performing as expected (see attached). Unfortunately the weather has been poor so I have not yet been able to test it on the stars. It's one innervation (if it is one at all) is that i have built it into a temperature controlled housing. The heat generated from the CCD is extracted from the enclosure using a commercial air to air bidirectional peltier cooler controlled by a commercial H-bridge controller.

Using this I can control the temperature to better than +/- 0.05 degree.

Regards Andrew

Hello to all:

With the aim of helping, Here is the tuning phase of the Echelle Mussol and your characteristics.
http://www.spectro-aras.com/forum/viewt ... f=8&t=1270

Also worth noting Module Guide, where there are the lights of tungsten and calibration. For each star observed, take the tungsten and calibration. It is a way to reduce the effects of temperature changes.

Also attach a scheme of the Echelle, the Linx, ( Luchs, Lynx ), and some photos of their state of construction.

The module guide is built by Antoni Jové (actually a small serie of five units, for those who wish).

Regards, Joan.

Hello Andrew :

A question. What is the diameter, or the ratio of the focal objective CANNON 100 mm. camera lens ?

Cheers, Joan.

Joan, it is a canon 100mm f2.8 camera lens as collimator and a Leica 90mm F2 camera lens.

Regards Andrew

Hi Andrew:

Looking at photos of the conference of Yahoo, it gives me the impression that the prism of your Echelle not have the right size and clogging, much of the light that reaches the network Echelle. The rest of the light leaves the prism, may again be obstructed if it uses CANNON 100 mm. 2.8 in (36 mm. in diameter). This is only my opinion.

I might put on-notice your image "Observatory", where the Orders are degraded enough in the end. Maybe I see it wrong.

At the conference of Yahoo, to https://groups.yahoo.com/neo/groups/ast ... op=eupdate, there is a spreadsheet, the Richard's Echelle Spectrograph, which may help to find out for sure.

I wish wrong.

Regards, Joan.

Thanks for your comments at the focal ratio the spectrograph works at there is no significant obstruction. While the prism could be a bit larger it is reasonably sized.

Regards Andrew

Just managed to calibrate Order 51 using Bass software. Measuring a Th line gives R>9000 with a calibration error of 0.02 A/px at an average dispersion of 0.16 A/px.

I will process the calibration frames taken at the end of the experiment to determine any shift while the enclosure temperature was maintained to within +/- 0.05 deg C

The forecast is for a clear night tomorrow so I may get a real first light!

Well I am still waiting for some clear skies as I wasted the last semi-clear night.

I have however been trying to understand if the echelle is stable or not in its wavelength calibration. I ran a test where I started the camera and the echelle temperature control and took calibration spectra at 15 minute intervals. At each point I collected 6 spectra and median combined them to give a good S/N ratio.

I then looked at the line at 6531A. Initially I used Bass to locate its position in Px and looked at the variation from sample to sample. Following that I did a similar thing using SPLAT-VO to fit a Gaussian distribution to the lines. Both methods gave similar variation in the samples although there was an offset of about 0.5 Px between them.

In both cases the variation (once the initial settling see later) was about +/-0.1 Px. The average dispersion was 0.16 A/Px (R~9300). Thus equivalent to 0.74 Km/s for 0.1 Px.

I then tried using the CCF function in ISIS and got the results in the attached diagram. Sample 7 was used for all the correlations. Clearly I need to do the trial again over a longer period as it seem it take up to 2 hrs for the echelle to settle down once switch on even though it's internal temperate (measure in the air) quickly comes under control.

There was no obvious relation to the shift in wavelength with the external temperature or pressure during the test.

As always advice and comment welcome.

Regards Andrew

Well, that seems to be a reasonable output for a fiber fed instrument, Andrew, although more tests are required, indeed. Good work! Cheers, Thomas

Hi Andrew,

the screenshot looks really good! Seems, that this combination of caemra lenses is suitable for that task! What is the weight of the system? I gues even with modifications it would be to heavy for direct linking?

Have you tried to get higher inter-order separation in the red. Looks like that you are getting cross-talk!

ciao,
Daniel

_________________
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, it was designed to be fiber mounted and is much to heavy to be telescope mounted. I can only just lift it!

It has a massive ceramic tile as a base and I deliberately over sized the enclosure to provide some thermal mass. It is covered with 50mm of insulation apart from where the peltier heater/cooler sits in the case.

I will look at cross talk in the red but I don't intend to below the H alpha order. At a glance I can't see any inter-order leakage between the ThAr calibration spectra but I will make a thorough search.

I have rerun the stability test I did before but for longer. I got a very similar standard deviation of 0.37km/sec. To better this I think I will need to look at the stability of the power supply to the Th Ar lamp and "shake" the fiber. I also suspect the CCD cooler control may well contribute to small shifts in the CCD but I have yet to think of a way to test this.

I certainly can't find any correlation between the enclosure temperature, ambient temperature or pressure and the results I got.

Regards Andrew

That seems quite convincing, Andrew. You obviously know what you are doing! I believe your ad hoc standard deviation is a very good result and shows that you are on the right track. Very nice!
Cheers, Thomas

I have the echelle up and running well now. I attach an image of the 51 order which covers H Alpha with R ~ 10000.

I am intending to study some bright stars (binaries, hypergiants and RV standards) over the summer to learn how to make quality RV measurements and generally home my skills ready to start some real programmes in the autumn.

The spectra is of HD89758 and compares my 4 x 5min exposure with a filtered spectra from the Elodie archive.

Still need to get the reduction automated via IRAF so lots to do on cloudy nights.

Regards Andrew

Cool stuff, Andrew! Can you show a picture of your machine and a design sheet? What about the main lessens you learned during the design and construction phase? I guess we all would appreciate an article for our section journal. Cheers, Thomas

Hi Thomas,

I am sorry but I forgot to take photos of the echelle before boxing in its insulation. I will see if I can gather all the data and write it up for the magazine later in the summer when the nights are short.

I have the bits to build my Hi Res echelle target R ~ 20000 so that is tempting me! (As is the list of jobs from my wife.)

I have now managed to obtain spectra full automated. I have a Python program that controls two instances of The Sky X on two computers by sending them JavaScripts. The first acquires and centers the targets from a list in a csv file. It then starts auto-guiding the target on the 75 micron hole in the guide head with a second camera. Unfortunately the field of view of the guide head is too small to reliably plate solve especially with a bright target so I use an On Axis Guider which passes the IR below 750 nm to obtain the right field and center the target.

The second copy runs the echelle camera as the Sky X only allows two cameras per instance. I also have automated the ThAr reference capture by adding a USB relay to switch in the reference mirror . This is also controlled by the Python program.

I would like to hear form anyone else who has done this kind of automation.

Regards Andrew