STScI Logo

Hubble Space Telescope
Shading

Shading

Diagnostic

Shading is a time-dependent bias that changes across a quadrant as the pixels are sequentially read out. This bias is imparted by the readout amplifiers. It starts out highly nonlinear, rises to a peak and falls off exponentially, rising slightly again at the end of the readout. The amplitude of the shading depends on the camera and the time since the previous readout of the array. The direction of the shading across an array depends on the readout direction of that particular detector, and the 3 NICMOS arrays are rotated relative to each other in terms of readout direction (see figure above). The first 20 or so rows/columns (first 5000 pixels read out) are actually read twice with each readout to help reduce the non-linear effects at the start of each readout. This causes a ripple in the smooth shading curve in those pixels (seen as a "spike" near row 20 in the plots above).

The shading has been found to be variable on both short and long timescales, and it is this variability that is thought to be the source of the "pedestal effect" (see the section on "pedestal" in this document). The dark subtraction done in the calnica pipeline should remove the shading completely as a component of the NICMOS dark, but if the shading has changed since the dark reference file was made there will be a residual left in your image. This time (perhaps detector temperature) dependence is most pronounced in NIC2, as its shading amplitude is the largest of the 3 cameras. The residual will appear as a curved ramp across a quadrant in a calibrated image, and will usually be accompanied by the symptoms of pedestal - namely a negative imprint of the flatfield pattern on the sky.

Cure

The most common method for removing this effect is to process the raw data with CALNICA up to the point of the flat field correction, and then use a standalone task to compute and subtract the median value of each column or row in the image (whichever is perpendicular to the fast read direction). CALNICA can be run on the results to complete the processing.

A tool to allow users to generate synthetic darks is available. This tool allows the observer to vary the degree of dark current, amp glow, and shading to better match their data.