Alnitak XL Flat Fielder Test Results
The Flat-Man XL is our stationary observatory model electroluminescent panel that will accommodate telescopes up to 44" in diameter. This product is revolutionizing the way flats are acquired at observatories around the world. They are used by both professionals and amateurs on several continents.
Testing the Alnitak Flat-Man XL:
Evaluating the quality of a flat field is straightforward:
- Generate a series of images with the telescope pointed to the flat field illumination source.
- Rotate the camera by 90 degrees, and generate another series of images as flats with ADU values in the midrange of your camera’s linear regime.
- Dark subtract each image, then divide each image in the first set of images by a sigma-combined stack of the second flat series.
- Median (or sigma clip) the first stack of images.
If the camera has been evenly illuminated, then the resultant stacked image should show a poisson distribution with very small standard deviation.
Product Testing Results:
The product testing was carried out at the Galaxy Quest Observatory in Lincolnville, Maine with a 12.5 RCOS Ritchey-Chretien reflector and a SBIG ST-2000 camera with no filters. There was no moon, the roof of the observatory was closed, and the Flat-Man XL was positioned 22" from the secondary ring of the telescope. The telescope was focused on a star immediately prior to testing.
Figure 1: A capture of the Maxim DL screen with a stack of 20 one-second duration images of the Flat-Man XL that were dark subtracted and then co-added. Dust donuts are clearly evident in the image.
Figure 2: The Master flat produced by Maxim from 20 one-second duration images after having rotated the camera by 90 degrees on the telescope. Again the dust donuts are apparent.
Figure 3: The coadded stack of 20 individual frames that were dark subtracted and divided by the image in figure 2 as part of Maxim's calibration routine.
Note the standard deviation of the flatted stack (Figure 3) is ~ 8 ADUs. Three standard deviations comprises 99.7% of the pixel value difference, which in this case would be ~24 ADU variation in the image. Since the average pixel value in the image is 31293 ADU, 99.7% of the pixel values are within 2X 24/31293 x 100 = 0.16% of each other!
We were skeptical of this value, so we did another test, this time leaving the camera alone and rotating the entire Flat-Man XL panel. Figure 4 shows that the results were just as good.
You can also see how flat the image is by looking at the line profile shown in Figure 5. Again, there is very little variation from one end of the image to the other.
Note that the above analysis relies on combining a number of images as would be done when processing aesthetic images. For those who do time series photometry and are not stacking frames, no flat can match the performance of a stacked set of frames. For single frames, we find that the Flat-Man XL is uniform to <0.75% across the field. That's a conservative estimate. If you are careful to expose the flats to the upper range of the linear part of your chip, you may be able to get even better uniformity.
Narrowband filters are no problem for the Flat-Man XL. We did testing through an Astrodon Ha filter with excellent results. Using the same test setup as above, we found that with an ST2000 monochrome camera and the brightness turned up to 255, we got 30,000 ADU flats in 5 seconds. Figure 6 summarizes the results. The results: less than 0.4% deviation from even illumination.