Ken,
the DSI3 pixels are 6.5 x 6.5 micron, and it's a b+w chip so that avoids another issue.
I noticed that you and Robin had some interesting chats on the yahoo site about the use of the aperture mask on the SA.
A filter slide or wheel that could move the SA in and out of the optical train would allow you to take an image of the star field, then the spectra without removing the camera from the scope. It would make it easier to flat field away the background stars... what do you think!!
Ken,
the DSI3 pixels are 6.5 x 6.5 micron, and it's a b+w chip so that avoids another issue.
I noticed that you and Robin had some interesting chats on the yahoo site about the use of the aperture mask on the SA.
A filter slide or wheel that could move the SA in and out of the optical train would allow you to take an image of the star field, then the spectra without removing the camera from the scope. It would make it easier to flat field away the background stars... what do you think!!
Mark
I have tried this as I have my SA on a filter wheel. It will make it easy to position a star on a knife edge etc as I can frame it with a clear filter first. I tried subtracting a clear image but it didn't work probably becuse of different focus with the SA and the clear filter.
I have tried this as I have my SA on a filter wheel. It will make it easy to position a star on a knife edge etc as I can frame it with a clear filter first. I tried subtracting a clear image but it didn't work probably becuse of different focus with the SA and the clear filter.
Yes, I had to refocus and then readjust the frame size to get the stars to match. Also you can see I didn't place the camera back in exactly the same position after removing the SA.
It was a fiddly job.
How does the knife edge/SA distance affect the image?
Ah next investigation.
Yes, I had to refocus and then readjust the frame size to get the stars to match. Also you can see I didn't place the camera back in exactly the same position after removing the SA.
It was a fiddly job.
How does the knife edge/SA distance affect the image?
Ah next investigation.
Not sure yet as it has been cloudy. In theory it should be as close to the CCD as possible. The image is not focussed when it passes through the SA so the knife edge should have a blurry side. I read somewhere that the resolution of the spectrum is determined by (among other things) the number of grating lines the star image crosses on the grating. This means the more blurry at the SA the better. This will blurr the edge also.
I will have to experiment.
The resolution of the SA is dependent on the distance from focus.
The more out of focus the star image is at the plane of the grating, the more lines are illuminated, the better the resolution.
The SA is 100lpm grating , so a 4mm OOF star would illuminate 400 lines, a 6mm image, 600 lines ( a 50% increase in resolution!) etc etc
You still loose efficency with large star images...
When the beam is collimated ( and a slit added) you have the potential of the full SA aperture 28mm, 2800 lines.... x7 the resolution.
Hi all,
I have used the spacer to extend the SA from the CCD and rotated the aperture diaphram to the correct position. Included is a .zip of the image of alptau as a .fit. This image has the dark removed.
It is 5sec through the 4".
Feel free to play and post what you find. I would be very interested in any calibrations for wavelength sensitivity for my ST7.
Jeff,
Still working on your spectra...
Would a dispersion of 14.7133A/pixel make sense to you? Compared to the Vspec library K5 III.....
I'm seeing out to about 9000A at the moment - I think, with atmos absorption at around 7000 and 7200???
Did you/ do you have any IR filter on the CCD?
I'll post later the curves I've got.
I had a go at Jeffs image of alp Tau.
I measure 15.136 A/pixel.
This is similar to my genesis CCD using the same KAF0401 chip. I get ~12.5 A/pixel.
see attached.
Found this quoted on the Spectro-L forum:
Based on Alain Lopez - for the Star Analyser - 100lpm
Dispersion ( A/pixel)= 10000 x Pixel size (micron)/100 x Distance ( in mm, distance from the grating to the CCD)
I think Terry has done a creditable job....
I'm having problems getting all the obvious "lumps and bumps" to align just using a single A/pix calibration...
The spectra seems to vary from 14.88 to 15.13 as you go towards the IR.
I'll dump a graphic to show the differences...
Hmmmm, don't understand yet?? different focus???
You can see the registration of the lines at 4973/5124 and 8484/8621 but this starts to "drift" across the middle graph, ie 7077 misses and slight drift in 5851???
Last edited by Merlin66; 13-03-2009 at 02:53 AM.
Reason: added info
BTW Jack Martin gives a good (albeit small scale!) comparison spectra on p89 of the PAS book. Covers 3100 to 6600A
Some interesting comparisons with Jeff's one.
Jeff,
The attached image is my comparison between your alptau0012 and the K5 III library, after correction for your camera responce.
I also include a copy of the responce curve I generated and used; maybe Terry can compare this with his one??
Thanks Terry and Merlin, I really appreciate your time and efforts.
You seem to have identified the issue.
When I calibrated using the O2 line, I got one l/pix reading, but as I refined it after croping the l/pix changed.
I don't understand how the focus could change, unless the spectral image is not flat?
Would the multi-line calibration produce a better result?
I will post an image of etacar taken with the same setup, on the same night.
jeff,
If you use the camera responce curve to "correct" the raw spectra; remember each new configuration ie scope/ grating distance/ camera will require a new responce curve, a three line calibration will improve the result. For the alp tau you could try the 4973 or 5124, 5851 and the 8484..
still don't know where the distortion is coming from??
Interesting results. I don't think we can take too much notice of the more distant IR part of the graph. The blue end of the 2nd order spectrum starts to interfere.
Take a look at the images towards the bottom of this page. http://www.astrosurf.com/buil/us/loris/loris.htm
