KAF3200 : The best sensor

Discussion in 'Legacy Models - Community Support' started by Gert, Nov 18, 2017.

  1. Gert

    Gert Standard User

    Joined:
    Oct 21, 2017
    Messages:
    10
    Hello Forum,

    I want to follow up on a discussion that evolved from my other thread on the STT 32o0 camera no longer being available.

    http://forum.diffractionlimited.com/threads/stt-discontinued.3905/unread

    An observing friend of mine recently had the opportunity to compare his own imaging with KAF3200 sensor using a 0.45m telescope with imaging that was obtained during rented observing time at the Mt Lemmon 0.8m Schulman telescope.

    1. Schulman 32-inch Mt. Lemmon SBIG STX 16803,
      Exp. 3600 sec. (9 x 400 sec.).Aperture 813 mm, FL. 5690 mm, Pixel: 4096x4096 9µ, scale 0.33"/Pixel.
      FWHM 1.74" (Gaia)-limiting magnitude +22.6mag at S/N=5.

    2. Newton 17.4-inch Emberger Alm Austria SBIG ST10-XME,
      Exp. 3600 sec. (12 x 300 sec.).Aperture 443mm, FL. 2048mm, Pixel 2184x1472 6.8µ scale 0.68"/Pixel.
      FWHM 2.45" (Gaia)-limiting magnitude +22.6mag. at S/N=5.
    Note that with the 3200 sensor a scope 1/3 the light gathering area equals a 0.8m observatory instrument with 16803. Both scopes using 3elem correctors and 'clear' filter as documented in FITS header.

    [​IMG]

    Clear Skies,
    Gert
     
    Last edited: Nov 18, 2017
  2. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    9,932
    I’ve always said the 3200 is the best front-illuminated sensor ever made. It’s too bad the architecture apparently hasn’t been used on other sensors.

    Unfortunately it is not as popular as it used to be because it is “only” 3 megapixel.

    We still sell the STF-3200, and will in due course release an Aluma version.
     
  3. Jason

    Jason Cyanogen Customer

    Joined:
    Oct 11, 2016
    Messages:
    30
    Location:
    Long Beach, CA
    @Gert,

    That is a fanatastic comparison and immediately gets to the heart of why dedicated amateurs equipped with high QE, high dynamic range, low noise CCDs rival professional-sized observatories (not to mention how far imaging has come since Burnham's Celestial Handbook was published with images from the great mid-century observatories).

    In the simplest approximation, signal rate is proportional to pixel etendue * QE, where:
    • pixel etendue is collecting area * pixel solid angle ("A Omega")
    • collecting area, A = pi * (1 - f**2) * d**2 / 4, where f is central obstruction fraction and d is telescope aperture
    • pixel solid angle, IFOV**2 = (pixel size / focal length)**2
    Just comparing pixel solid angle * QE, all else except aperture assumed equal (optical transmission, bandpass, integration time, and object radiance) and ignoring factors of 1e-6,
    • Schulman = (9.0/5.690)**2 * 0.41 = 1.0 [sr] x 1e-12
    • Schulman 2x2 = (18.0/5.690)**2 * 0.41 = 4.1 [sr] x 1e-12
    • Schulman 3x3 = (27.0/5.690)**2 * 0.41 = 9.2 [sr] x 1e-12
    • Newton = (6.8/2.048)**2 * 0.72 = 7.9 [sr] x 1e-12
    Unbinned, Schulman needs 8x the collecting area or nearly 3x the diameter to overcome the deficit for signal electrons, but it's only a factor of two larger in diameter, so it has already lost. It does have the edge, of course, in resolution when seeing permits.

    Signal-to-noise ratios will be similar, as one can see in the image. Newton will record a higher background signal for the same reason it collects more object signal, pixel etendue * QE; however, the 3200 has lower spatial and temporal noise than a 16803, so it will tend to come out ahead. That said, the 16803 is a great large-format CCD that has 9x the physical area and is hard to beat for wide field imaging.

    Attached is a single raw frame comparison of NGC 891 taken with a STT-3200 and an -8300 on the same telescope. I have matched apparent field and false color histograms, changing nothing else about the raw images. The better spatial and temporal noise characteristics of the 3200 are evident as is the higher SNR due to the pixel solid angle * QE; however, the 8300 image is more oversampled leading to rounder stars and perhaps exhibits a little more structure in the dust lane, though it is hard to tell with just a single frame. The original data was taken by another gentleman who originally posted the comparison on Astrobin (see link for 12x300-sec coadded final images). I requested and he provided me with a single raw image from each camera.

    Despite thin cirrus and a Windows 10 update that is still causing no end of trouble, I had first light this weekend. I've attached the first light image, 60-sec of the Horsehead in luminance, taken with a 4" refractor at f/7 and guided with the AO-8T at 12 Hz. Lots to play with, even in a single uncalibrated image; looking forward to a clear night (and a stable operating system...).

    Jason
     

    Attached Files:

    Last edited: Nov 20, 2017
  4. Gert

    Gert Standard User

    Joined:
    Oct 21, 2017
    Messages:
    10
    Hello Doug & All,

    Great to see that the Aluma 3200 is now featured on the camera website!
    I wonder if there will be package deals for camera+filterwheel/guider+ filter glass+guide cam+AO device.

    The Maxim DL software page has a feature comparison of the the different levels offered. But the 'Maxim Lite' mentioned in conjunction with the cameras is missing. Would it be possible to extend the features comparison table accordingly?

    Thanks & Clear Skies,
    Gert
     
  5. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    9,932

Share This Page