I tried CCDinpector for the first time and I am having the same problems as a recent poster when trying the defocused star collimation. I keep getting messages that either the star is too dim and not detected, or the star is distorted and try to decrease the exposure. I did get it to work on an occasional image, maybe one in 50, so I was afraid to try to collimate. I am using a modified Canon 350D through a C9.25, using ImagesPlus to drive the camera. I used atars between mag 0 and +1, varying exposures (1/8 second to 8 seconds), varying amounts of defocus, with and without dark subtractions, and kept the star centered. I was wondering if maybe DSLRs are too noisy; is the background interfering with star detection? Also, what should the ADU saturation level be set at for a DSLR? Also, I am working in heavy light pollution; could this be a problem?

Hi,

What format are the images saved in from ImagesPlus? Are they color or monochrome? 8 or 16 bits? How's CCDInspector configured: Bayer matrix or Monochrome?

If you have a single image that gives such an error, please send it to me (or post it to a website and send me the link):

paul [at] pk.darkhorizons.org

Regards,

-Paul

Firstly, thank you for your prompt reply.
ImagesPlus is saving the image in the native raw format from the canon 350D (.CR2 files); CCD inspector is then automatically opening them in the viewer. They are color images. (I put a filter in which mimics the orginal filter on the non-modified DSLRs, which seemed to helpa little; the star became white instead of red). I'm not sure about the bit depth of these images, but I think it is 12. I have CCDInspector configured to Bayer matrix.
Unfortunately, I had CCD inspector set to auto-delete the images, so I don't have any. But I will give it another try my next available clear night and get you one.
Thanks again.
Ed

Hi Ed,

I'll be interested in seeing one of those images. Use WinZip or other compression so it's easier to transport.

When processing DSLR raw images in CCDInspector, you should set the sensor type to Monochrome, as the debayering is done by the raw image processor before CCDI actually gets the image.

I don't believe that's the main reason for the problem you are seeing, but next time try to set sensor type to Monochrome in CCDI.

Regards,

-Paul

This message has been edited. Last edited by: Paul Kanevsky,

Hi Paul

I took more images tonight, same problem. The raw image size is over 4.5 megs, even after using WinZip; it won't let me upload it. So I'm attaching a jpeg screenshot of the image as it appeared in the CCDInspector viewer. If you need to see the raw image, any ideas on how I can get it to you?
Thanks.
Ed

1_second.jpg (417 KB, 14 downloads) Defocused jpeg

Hi Ed,

The easiest way would be to email it as an attachment. If this doesn't work and you have a website, you can upload the image there and send me the link to download it. If that's not possible, email me privately and I'll provide you with an ftp access where you can upload the image:

paul [at] pk.darkhorizons.org

It's hard to judge from the JPEG, but it would appear the noise level is very high relative to the star brightness. I'll be able to tell for sure once I see the raw image.

Regards,

-Paul

OK. I emailed you the raw image. Also, I converted the raw image to TIFF using ImagesPlus, both with White Balance option on and with it off (Mono Bayer Data).

Thanks.

H Ed,

Sorry it took so long: I had to deal with two computers failing within 24 hours :-(

Now that I have one back in working shape, I was able to take a look at your shot. The raw image has, unfortunately, very low S/N -- the brightest parts of the star are within 400-500 adu, while the background is around 200 adu. This is not enough for the defocused star algorithm to figure out the correct flux distribution in the star. It will normally look for at least 10x the difference between the background and star adu (yours is only around 2x).

A few things to try:

1. Find a brighter star
2. Expose longer
3. Try increasing ISO to 800

Let me know how it goes.

Regards,

-paul