Alpy600 Calibration Problem

[Keith Graham]

Hi,

I just got my ALPY calibration module and put it to the test. I tried calibrating a series that I took a few nights ago using the Ar-Ne lamp. There are some very strange results coming up, and I am really in need of some help.


First, I calibrated this series without using the calibration module. I processed this series of 4 images using the Predefined Mode (ALPY600 Balmer lines) and used the first image in the series as the calibration image (per instructions in calibration without the calibration module). I also entered the flat, dark, bias, and cosme files and the appropriate Tilt and Smile data on the Calibration page. With this method and clicking the Go button, my final profile has an RMS of .4 and appears reasonable (even though there are a few small glitches I need to work out),

Next, I took a calibration image using the AR-Ne lamp. Per the instructions when using the calibration module, I entered this image into the Calibration box in the General tab. I also used the Predefined Dispersion Equation per the instructions. As above, I included the flat, dark, bias, and cosme files. I click next to get to the Calibration tab, and obtain the Tilt and Smile. I then go to the Calibration Assistant. Here is where the confusion begins and some weird things happen:

1. I notice that sometimes the Calibration neon spectrum image name does not carry over from the General tab. It will have the image I used for calibration from a previous setup. I will need to shut down ISIS completely and start over for the new calibration image to appear in the Calibration Assistant.
2. The X Coordinate of line at wavelength on the Calibration tab sometimes is shown as 5852.49 and occasionally as 6562.83. Sometimes I can change it and sometimes not. I am assuming from the instructions that it should always read 5842.49 and cannot be changed.
3. I see in the instructions that when the mean image is computed in the Calibration Assistant , this name apparently carries onto the Image to Process box on the Calibration tab. This does not happen for me, and I enter the name of that mean image manually. I then click on the H alpha line to get the X and Y coordinates.
4. Now here is the really confusing part and the part that really concerns me. When I click the Go button in the Calibration Assistant, it gets as far as Save the @.dat file, OK, and I get the message:

Error: Index was outside the bounds of the array.

If I click OK, the procedure continues to the RMS line and that is always way too large. One example is RMS 27.82 I have gotten RMS ranges from 25 to 32. Never have I gotten close to the .4 in the Balmer line process.

The message then appears:

The calibration error is too high. Verify parameter and data.

So as you can see, I am not having much luck using the Ar-Ne lamp calibration method with the Calibration Assistant. I get much better results with the Predefined Balmer line method. Can someone please help me to see where I am going wrong? I have been going through the process for about 4 hours trying to find my error and am having no luck.

Thanks,

Keith Graham

[Francois Cochard]

Hello Keith,

I've tried to reproduce your problems... and I do (with recent images taken in conditions similar as yours).

I've no solution so far. I just remember that the ISIS calibration algorithm is very sensitive to the pixel size parameter (it uses it to explore the spectrum to find Hbeta and so on lines).

I'm not able to recover the way to tune this parameter - or find any work around. Let's put Christian in the loop !

The temporary solution is to :
- Process your data without wavelength calibration,
- Use the 'dispersion" tool in the profile tab, to define the calibration polynom,
- re-process the data with the "predefined equation" (without changing parameters at this step).
It works... but it's not that fast. This is why Christian developped the assistants to make the calibration easier and faster... which should work !

Stay tuned !

Friendly,

François

[Christian Buil]

Hello Keith, hello François, hello all,

I have resumed major calibration modes et pipeline in a single page for Alpy 600 users
(but the characters of the page is general, LISA, Lhires users can applied similar techniques):

http://www.astrosurf.com/buil/isis/guid ... onnage.htm

My opinion (I agree with François opinion) is a bad value for pixel size (it is a scale factor
for ISIS for recover spectral lines is an automatic process). Try 9.02, 9.04, ... 9.10, .... microns
and check the RMS value.

MOST IMPORTANT : read with attention the new resume calibration page and test in sequence
the method 1, method 2 and method 3 proposed. Not only the method 3
It is the better technique for understand ISIS !

Sorry, the page is only in french, I hope help english for traduction !?

Note also a new version of ISIS available: V5.2.2. I recommand download because the
dual calibration tool (star + relco lamp) is now improved (better precision).

========================================================================================

Une page (non encore référencé sur mon site) qui décrit les modes principaux
d'étalonnage avec Alpy 600, que vous ayez le module d'étalonnage ou pas.

Tout est résumé pas à pas. Il est impossible de rater un etalonage de cette manière !

http://www.astrosurf.com/buil/isis/guid ... onnage.htm

Je recommande très vivement de tester les 3 méthodes sucessivement. C'est la meilleure
façon d'apprendre les principes de fonctionnement de ISIS.

Cette page va interesser aussi les utilisateurs de LISA, Lhires, ... car les techniques
sont voisines.

Une traduction en anglais de cette page assez synthétique serait bien venu je pense.
Je ne me sent pas capable (par contre je peux faire les copies d'écran en anglais).

Notez que j'en ai profité pour poster une nouvelle version de ISIS (V5.2.2). Il y a des
modifications subtiles, mais surtout, le mode de calibration mixte pour Alpy et LISA
donne maintenant des résultats significativement meilleurs. Je recommande
le téléchargement.

Christian

[Keith Graham]

Hi Christian,

I had already performed some of your suggestions before you posted them here, and I had already printed out the new documentation you have prepared. I now have v 5.2.2 installed. I now went back through the methods as you suggested, and here are my results:

1. My camera is the Atik 314L with pixel size 6.45u I tried different sizes in the general tab and found 6.42u to be optimum with RMS= .77 using method 1 and RMS =.076 and resolving power =544 using method 2 . For calibration, I used the first image in the sequence.
2. When using both Method 1 and Method 2, the resulting profiles were very reasonable save for some irregularities between the H beta and H delta lines (this is another issue currently being investigated).
3. Method 3 is what is giving me the problems. When using the Calibration Wizard. I first reset the Spectrograph Model in the Settings tab to Alpy600 with calibration module. I set the Calibration image in the General tab to the calibration module ArNe image I took. I set Spectral Calibration to Predefined dispersion equation. I did have to manually put in the 5852.59 line as that did not automatically change from 6862.82 as used in Methods 1 and 2. I then clicked on the Calibration Assistant. I computed the mean image by pressing the GO button in that section. I then checked to see that the correct ArNe image was in the calibration neon spectrum box, checked that the X and Y Ha coordinates were correct, and made certain the UV Calibration box was checked. I then pressed GO.
4. After the line that says “Save the profile
OK”
I get the error message that says “Index was outside the bounds of the array”

5. I clicked OK and the process continued down through RMS=33.179377 when I get the message “The calibration error is too high, verify parameters and data”.

At this point I tried numerous pixel values between 5 and 6.7. The best RMS I could get was 3.4 and that was at about 5.5u. Most RMS readings were between 15 and 25. When using Methods 1 and 2, I was getting the best RMS (.076) at 6.42u.

So this is where I am at now. Changing pixel values does not seem to work on Method 3, so I am at a loss as to what to try next.


Cheers,

Keith

[Christian Buil]

Please Keith,

Send to me (on a zip file for example) ONE actual stellar image and ONE ArNe line image. Try also to post on the forum (or in the zip file) a screen copy of General tab, of caliration tab and setiings tab. It is the only solution for debug your problem.

Christian

[Christian Buil]

Thanks Keith for the images.

The problem is now solved (easy). It is related to moderate size of your CCD chip, not fully taken
into account by ISIS.

All is OK now. Download just updated V5.2.3 ISIS version.
Adopt 6.43 microns size pixel. Choose also good coordinates
for Halpha position (X=1019, Y=471).... and now calibration is perfect.

One question Keith, you use your LX200 10" telescope with focal
reducer or without focal reducer ? Your spectrum is really excellent
in the UV region. Very perfet! I suspect Alpy 600 fixed at the f/10 focus.
Exact ?

Christian

[Keith Graham]

Hi Christian,

The new 5.2.3 version took care of my issues. The profiles look great even with the IR block filter still in place. The continuum has been evened out, and all of those bumps and dips are gone. The calibration and response wizards do a spectacular job. Thanks for the continued help.

I was under the impression that the ALPY was designed for f/5-f/6. so I am using a f/6.3 focal reducer on my LX200 10". Is this still a good idea, or do you suggest I try f/10?

Cheers,

Keith

[Christian Buil]

Hi Keith,

First, yours questions help to progress!

Operate Alpy 600 on a f/6.3 optical beam is a very good choice
(remember, Alpy can also accept f/5 to f/4 beam - for Newton
telescopes users).

Your spectra are so good that I have some interrogration.
Many focal reduced produce large chromatic aberration,
clearly visible in the bleu and UV part (the spectrum trace
is enlarged). The chromatism seem absent from your data.
It is exceptionnal (you observe down to 375 nm wavelength....! ).

Keith, can you precise the focal reducer model (reference) ?
It is a Meade model ? (auxiliary question: what about distance
between focal reducer lens and the spectrograph slit in millimeters).
You have a photo of the setup ?

Christian

[Keith Graham]

Hi Christian,

Thanks for the kind words. I am now confident that I can begin working on the RR Lyr campaign and some others.

In answer to your questions, the f/6.3 focal reducer is a Meade Series 4000. The distance from the FR lens to the slit is about 85mm. As you can see in the photo, I did not use the Alpy T mount for connection to the scope. I first attached the FR to the scope and attached the ALPY to the FR using the SBIG T thread visual back.

Cheers,

Keith

[Robin Leadbeater]

This focal reducer also seems to work ok for me but only provided the reduction is limited to the design level. I typically operate mine with ~100-110mm spacing but if I try to push it even a bit further then the results (chromatism and various other defects) are terrible.

Robin