Introduction

The NIRC detector has a number of adjustable voltages. Normally the default parameters set up by the facility scripts are sufficient. However, observers may benefit from understanding the issues involved even if they do not tweak the parameters away from their defaults.

Linearity at different detector biases.

One adjustable parameter is the detector bias, set by the detbias command. The default is 1 and is set upon startup of the software.

The detector bias affects two characteristics of the images: dark current and dynamic range. A higher detbias value reduces the dark current, and setting the value too high will reduce the dark current to a negative value. Unfortunately, it also decreases the dynamic range to an unusable value.

Dynamic range and linearity are intimately entwined. In fact the dynamic range is normally defined by a certain linearity level. For example one could define the point at which the linearity curve deviates from linear at the 1% level as the highest useable signal in the dynamic range calculation.

Examples

Below are examples of linearity measurements for three detbias values: 0, 1, and 2. These data were obtained by taking images with the tertiary mirror cover closed and the K filter in the light path. Exposures of different integration times were taken. Exposures with the f77k dark plug in were taken at each tint value as well, to estimate both the dark current and provide a good bias frame for the K-band images.

For the three detbias values we plot the mean DN (with the dark frames subtracted) vs. tint:

Linearity at a given tint value was calculated by fitting a line to all data points of lesser exposure time, extrapolating that to the tint of concern, and measuring the deviation of the data from the extrapolation.

Dark Current

Note that the lower the detbias setting the better the linearity. The drawback is that the dark current is higher at the lower detbias settings. Dark current rises rapidly in a relatively small number of pixels, so the median of a dark frame may not change much. The mean dark current is:

detbias	Dark Rate (DN/sec)	1% linearity
0	27	40,000
1	4	33,000
2	0	23,000