New NOTCam Science Grade Array (SWIR3)

• Detector overview
• Cosmetics and bad pixels (hot, cold, zero)
• The dark level
• Readout noise
• Gain
• Non-linearity
• Detector flat field
• Memory effect (charge persistency)

For data taken before this date, check detector characteristics of the Engineering Array or the First Science Array.

NOTCam is offered with two possible readout modes: the standard reset-read-read mode and a ramp-sampling mode (multiple non-destructive reads during the integration). Read more about this in the NOTCam User's Guide.

Fig 2: Two dark images obtained with the reset-read-read mode (command dark t). Integration time (t) is 0 seconds (left) and 42 seconds (right). Note the shift register glow and the higher number of hot pixels on the 42 s long dark integration. The increase of hot pixels with exposure time has to be looked at more carefully (data is obtained), but a preliminary result is that we do not seem to have those accumulating groups of hot pixels as with the previous Science Array.

The dark images show a cosmetically clean array with only a few bad areas. The largest bad area is about 33 x 13 pixels centred on x=150, y=508, another one is about 10 x 5 pixels centred on x=390, y=586. (The lower left corner is pixel x=1, y=1.) The total number of bad pixels is about 0.8 %, of which most are located in one corner (upper left). The bad pixels are both zero value pixels, hot pixels and cold pixels. A pixel is defined as hot/cold if it deviates by more than 6 sigmas from the mean level. Both darks and high S/N ratio domeflats are used to define the bad pixels.

The four quadrants are read simultaneously. The fast readout is in the horizontal direction starting in the lower right corner of each quadrant. Because the controller can not handle the first column, we have a dead column at x=1024 as well as at x=512. (NB! Please, note that since January 2006 the images are flipped in x-direction, and the above is valid for data taken from 2006 and onwards.)

Amplifier glow starts to show up on 12 second darks at a level of 500 adu. This effect drowns in the background for broad band imaging and apparently subtracts out well.

Please, check the NOTCam User's Guide for a description of the two different readout modes available with NOTCam.

Fig 6: Processed flat field obtained from 8 differential twilight flats taken with the WF camera through the Ks band. The differential method (pair-wise subtraction of "bright" minus "faint" images) is used to eliminate the thermal contribution from the master flat.

The detector flat field looks relatively flat and has few disturbing features. The figure above shows the master flat obtained from 8 differential twilight images for the WF camera and the Ks filter. The standard deviation in small boxes of 20 x 20 pixels is less than 1%. The deviation over the whole field is ± 3%, i.e. flatter than the previous Science Array and much flatter than the Engineering Grade Array.

Thus, for the New Science Array (SWIR3) the magnitude of the persistency effect is smaller than that measured for the First Science Array and also the recovery is faster.

Note that the memory effect is negative on the first image and positive on the subsequent images. This is believed to be due to a transient change in the gain.

If you can not avoid saturation, it is recommended to clean the array with a couple of clear commands between each science exposure. Taking 3 clear commands after the saturated image gives a memory in the subsequent dark image of only 0.03 %.

The memory effect in detail - readout by readout.

Image	Comment
	An argon arc lamp spectrum with two saturated lines (peak >> 56000 adu).
	The first reset read after the arc image. There is a negative memory of the brightest lines with peak values > 40000 adu on the arc image. The reset level is at 5500 adu and the negative peak can be as deep as 5000 adu.
	The 1st dark (reset subtracted). The memory effect is negative and at a level of 0.5 %. Because this is a dark image - i.e. it has few counts all over (about 300 ADUs) - the negative traces are of even fewer counts. When the pixel values become negative, they wrap around to > 65000 ADUs (data is saved as unsigned integers), which makes the image look very ugly. If you have very low background (narrow-band images and/or very short exposures) and sources peaking > 40000 adu, you should consider wiping the array with a couple of clear commands in between exposures in order to avoid this effect.
	The 2nd reset read shows no traces of negative memory. (On the other hand the pick-up noise pattern is clearly visible.)
	The 2nd dark (reset subtracted) with a persistency of 0.03 %, meaning that only the brightest lines with peak values > 30000 adu have a positive memory at a detectable level on a dark image.
	The 3rd dark (reset subtracted) with a persistency of 0.014 %, meaning that only the saturated lines (peaks > 56000 adu) show positive memory at a detectable level on a dark image.