atik vs h9 comparison

Atik 314L - 6 min dark frame viewed full range 0 to 65535: While relatively 'clean' compared to Kodak chips, it has a reasonable number of fully saturated pixels.

SXVF-H9 - 6 minute dark frame viewed full range 0 to 65535: In comparison to the Atik dark frame, there are only 2 pixels in the entire frame that reach full saturation!

Statistical data from the dark frames:

Here are some local measurements on image (using Maxim), but representative of the entire frame.

Atik 314L:

SXVF-H9:

Several additional comments regarding the data and cameras (some from my friend):

1 - The Atik darks have higher noise than the H9 darks (both small region and full frame as shown above)

2 - There are some unusal parings of the hot pixels that indicates an unusual chip behavior (not likely a readout issue), possibly some cosmic ray damage to the isolation between pixels. This behavior is not seen on the H9 dark frame. These could be the result of shipment by air, excessive storage time, or a supply from a gray market (seconds) source. Sony ships by sea to avoid damage like this.

3 - Some quick cooling tests on the chip indicates that the Peltier is undersized and does not achieve the -25C delta as specified. After 5 minutes, the cooler dropped the ccd chip temperature by 15C from ambient (to +8C), but actually warmed up to +12C with addition operation. The cooler is not controlled to a setpoint by software (i.e. Maxim) nor monitored, despite comments by some.

4 - The average background level of the Atik is quite low. Some of the individual pixels are near zero (i.e. less than 10), so some care should be taken with the light frames (could go negative with calibration).

You can download the full FITS files from the 2 cameras below if you would like to take a look for yourself.

Raw Atik 314L 6 minute Dark File (FITS)

Raw SXVF-H9 6 minute Dark File (FITS)

Addendum: I received an email from Steve Chambers, who is the developer of the Atik cameras. Regarding comment number 2, there is a firmware update that is available for first cameras that show the paired bright/dark pixel. All new cameras have the new firmware update, so this artifact should no longer be a problem. Check with an Atik dealer if your camera shows this or if you have questions about it.


H9 vs. 314L		Updated: 8.17.08
	There have been several comments regarding the published performance of the new Atik 314L vs. the SXVF-H9 camera. A friend who has both cameras has provided some data, some comments and raw dark frames to me for evaluation and posting. Here are some of the results and combined comments.
Return to new images
Atik 314L - 6 min dark frame viewed full range 0 to 65535: While relatively 'clean' compared to Kodak chips, it has a reasonable number of fully saturated pixels. SXVF-H9 - 6 minute dark frame viewed full range 0 to 65535: In comparison to the Atik dark frame, there are only 2 pixels in the entire frame that reach full saturation! Statistical data from the dark frames: Here are some local measurements on image (using Maxim), but representative of the entire frame. Atik 314L: SXVF-H9: Several additional comments regarding the data and cameras (some from my friend): 1 - The Atik darks have higher noise than the H9 darks (both small region and full frame as shown above) 2 - There are some unusal parings of the hot pixels that indicates an unusual chip behavior (not likely a readout issue), possibly some cosmic ray damage to the isolation between pixels. This behavior is not seen on the H9 dark frame. These could be the result of shipment by air, excessive storage time, or a supply from a gray market (seconds) source. Sony ships by sea to avoid damage like this. 3 - Some quick cooling tests on the chip indicates that the Peltier is undersized and does not achieve the -25C delta as specified. After 5 minutes, the cooler dropped the ccd chip temperature by 15C from ambient (to +8C), but actually warmed up to +12C with addition operation. The cooler is not controlled to a setpoint by software (i.e. Maxim) nor monitored, despite comments by some. 4 - The average background level of the Atik is quite low. Some of the individual pixels are near zero (i.e. less than 10), so some care should be taken with the light frames (could go negative with calibration). You can download the full FITS files from the 2 cameras below if you would like to take a look for yourself. Raw Atik 314L 6 minute Dark File (FITS) Raw SXVF-H9 6 minute Dark File (FITS) Addendum: I received an email from Steve Chambers, who is the developer of the Atik cameras. Regarding comment number 2, there is a firmware update that is available for first cameras that show the paired bright/dark pixel. All new cameras have the new firmware update, so this artifact should no longer be a problem. Check with an Atik dealer if your camera shows this or if you have questions about it.