spectro-aras.com

I made a few measurements with Joan's data :
https://drive.google.com/file/d/1zImx8p ... TLxcA/view

and this is what I get

First I measured the centroid of each image on an emission line of the Rs Oph spectrum as shown in the image below



there is a gradual shift of the position in X and Y of each spectrum. Here're the values ​​measured with the Mira Pro software for the same emission line (I think we can have better result using the calibration lamp instead of a star in the sky, but a centroid is a centroid unless the spectrum changes during acquisitions).


The maximum shift is dX=2,88 pixels and dY=3,818 pixels
The pixel size of the ASI 2600 MM pro is 3,76µm, so a shift dx=10,83µm and dy=14,35 µm

And an animation with the first spectrum and the last



So the spectrograph "move" comparing with some values taken with an eShel with the calibration lamp (during more 3hours of observation). The eshel was in a thermal enclosure and the temperature variation doe's not exceed 0,4°C during 3 hours.




here, the maximum shift is dx=0,03 pixel and dy=0,05 pixel
The pixel size of the ATIK 460ex in bin 2x2 is 9,08 µm so the shift is dx=0,27µm and dy=0,45µm

We have this result because the eshel is in a thermal enclosure, but in the ambiant air (without any thermal regulation), the shift can be more important depending of the temperature variation during observation due to expansion of materials.

Here's an animation with 6 raw spectra taken with an eShel and ATIK 460ex in bin 2x2 (same scale from the previous animation). Observation time around 45minutes.



and here is an article that independently compares 2 commercial echelle spectrograph (eshel from Shelyak vs Baches from Baader)
https://iopscience.iop.org/article/10.1 ... 115001/pdf

Bonjour Olivier,

Many thanks for your interest and effort where you show the deficiencies of my NOU_T echelle opposite the perfect operation and comportement of the eShell - a matter that I have never put in doubt.

But don't you think your study is really incomplete ? I don't want to say that it is partial or tendentious.

Don't you know that the first step on the ISIS traitement the binning 1x1 becomes on binning 2x2, and it is that really count ?

Don't you think in this case the shift is halved from dX=2.88 and and dY=3.818 ?
So for NOU_T + ASI2600MM pixels 3.76 binning 2x2 (7.52x7.52) , dX=1.44 and dY=1.94

Don't you believe that you could explain more clearly here that the real comparison would be both echelles results with pixels 9.08x9.08,
the NOU_T shift would have been reduced to dX=1.19 and dY=1.58 ? Do you think that I'm wrong ?

I would be really appreciated if you could help me to recover a bit the NOU_T prestige and mine, if you have time to make
a new animation with the same files with binning 2x2 (7.52x7.52), that I have left there - for all -, that would be in honor to the truth, but I wouldn't want to bother you with more job but... https://drive.google.com/file/d/1M9AM7A ... sp=sharing

Really, the NOU_T echelle has lower resolution than the eShel, and, however, the NOU_T has these - littles? -shifts, it works
very well on spectroscopic measurements, with the same % ratio-error as eShell and others, or not ? You can see it in some
graphics by an analyst below down.

In other words, the main eShel advantage over the NOU_T, is the optical fiber used and its storage in a thermal box, or not ? Am I wrong ?
So, if you have read and seen my precedent message and picture , I said in it that in a few minutes the same NOU_T can change to an optical fiber spectrograph, then in my humble opinion the difference between both echelle disappears, am I wrong ?. With the same thermal stability the results are going to be the same, or not ? Am I wrong ? But, personally, at the moment, I don't want to work with this system, I have my own reasons.

Everybody knows that you are the Shelyak voice here, and you are an important member of this forum and that your opinions
have very high credibility here, so, don't you think with your incomplete explanation someone could deduce that NOU_T is a disaster and I'm an unreliable guy ? Above all, these two people that I sent a NOU_T echelle to each one ? You don't doubt that for me you are a really honorable person and I know that you never would do it consciously.

I always had and still have great respect, sympathy and admiration to Shelyak and its team, I always speak well of them and their products everybody, especially of the eShel echelle and I'm looking forwards about the happy day that you seft will be able to show to us a graphic from 3770 to 10400 Angstroms, made with your own one, so the same range as the NOU_T echelle...

Amitiés, J.Guarro.

Hola Joan,

My intention is just to objectively compare the data produced by 2 different spectrographs. Different because they are not used in the same way, one is not fixed on a telescope but placed apart in a temperature regulated enclosure, the other is fixed on a telescope and undergoes temperature variations (and also mechanical flexion even if these are weak). One is a fiber spectrograph, the other not (less efficiency for the fiber, but more precision for the light injected in this fiber).

Here's some new comparison with a pixel size of 9,08 µm (bin 2x2) for the eShel and ATIK 460ex, and a pixel size of 11,28µm (bin 3x3) for the Nou-T and ASI 2600 MM Pro.
And for a beter comparison, i try the eShel with others data :
- Same target : RS Oph (but not at the same date)
- Around the same shift of temperature (difference of 2°C between the 1 exposure and the last exposure because I had done tests without the thermal regulation of the eshel last august) and during this night I have not a good weather that's why my last exposure is very weak.
- Here for the eShel spectrum, this is 23 exposures of 180s each (1h 09 minute total exposure)
- Same emission line for comparison between Nou-T and Eshel

Here's the values of the eShel. Always a centroids.



The shift in pixel is in this case (with 2°C difference from the first and last exposure) dx=0,795 pixels and dy=0,139 pixels
Remember that my first test was with a thermal regulation about 0,4°C, so these values are necessarily less good.
So a real shift of 7,218µm in X and 1,26µm in Y

Now the values of the Nou-T in bin 3x3, so a real pixel size of 11,28 µm (9,08µm for the eShel)



The shift in pixel is in this case (with 1,1°C difference from the first and last exposure) dx=2,345 pixels and dy=0,402 pixels.
So a real shift of 26,45µm in X and 4,53µm in Y

And the animation for the Nou-T (first spectrum #1 and last spectrum #10)



And for eShel (first spectrum #1 and last spectrum #23)



But I think it's normal to have differences. In general case for all spectrograph, there are the typical values :
- a difference of 1°C is an error of 300m/s
- a difference of 1 mBar (atmosphere pressure) is 90m/s error
That why all professional spectrograph in High resolution are in a thermal enclosure with a very good temperature control precision around +/- 0,01°C or less

The second difference is because eShel is not fix on a telescope but yours can have some mechanical flexion.
That's why I put this article comparing eShel and Baches (Baches is fixing on a telescope and has the same problem than Nou-T) :
https://iopscience.iop.org/article/10.1 ... 115001/pdf

All spectrograph design is a compromise between several parameters :
- A spectrograph on a telescope have a better efficiency (no optical fiber, no injection module) but have more shifting issue than optical fiber spectrograph (temperature variation and mechanical flexion).
- The wavelength range of the eShel is less than the Nou-T because eShel spectrograph have a filter on the calibration fiber to prevent any bright Argon line up to 7600 Å to disturb others weaks lines from the Th/Ar lamp. We have all orders under 7600 Å, but no calibration lamp to calibrate the orders under #30. That's the choice of Shelyak to not go up after 7600 Å and have a better precision from the range 3900 to 7600 Å.



Do the test by yourself :
- Test your spectrograph "on table" during one hour and take many calibration lamp during this time and compare the values of the same emission line (and also the spectrograph temperature)
- Test your spectrograph on your telescope during one hour with of course the mount on the sidereal velocity during the same time.

In conclusion, if you put the Nou-t in a thermal regulation enclosure and not fix it on the telescope, you can reach the values of the eShel.

Bonjour Olivier,

Many thanks for your explanations, advices and time.

Until the next one.

Amitiés, Joan.

Hello Joan,

Not wanting to hijack the thread, I remember you mentioned working on a fibre head for your wonderful Nou_T echelle and wondered if you had any success in getting the fibre feed to work?

Stu

Hi Stu,

It is possible that I haven't spoken very well, really I have experience with the optical fiber, so these last years I made and was observed for a long time with  two spectrographs, the MUSSOL and the LINX, they both work with optical fiber, you can find more or much information in this Forum about how have  they made and observation results if you find for these words here.

How well Olivier Garde has exhibited here, if you want to observe exo-planets or do some similar jobs, you need a very high precision and a spectrograph with optical fiber (OF), is the best system if you have to count its thermal stability  and put it into the convenient place. But this system works bad for the modal noise that the same OF produces.

How well exposed C.Buil the solution is to make an OF agitator device that moves the OF during the all integration time. And if you clicked there,http://www.eso.org/~cguirao/caos/ and click again above (directly doesn't work), http://spectroscopy.wordpress.com,  you can find three very interesting and precise parts into "Optical Fiber". So it is extremely necessary that an echelle that works with OF needs an OF agitator for work, if not it is incomplete and loses a lot of fine details, in other words, doesn't work well.

Following the NOU_T philosophy, with its current equipement is an accessible, "easy" and practice echelle to make very good and precise observations around a R=8500, but if someone would do more precise results,  they need a 135 mm. focal objective and use the NOU_T on the OF version and obviously with a thermal box and a good OF agitator.

You can see Stu, in this world perfection is the same as lightspeed, you can approach it but you never reach it.

Cheers, Joan.