Dark Current and Total Noise Measured by Spatial Averaging

This technique, reported by Hall, et al.[1], utilizes a set of identical ramps, each consisting of resets followed by a fixed number of frames. The standard UH data set consists of five ramps, each made up of 2 pixel by pixel resets followed by 145 frames resulting in a 6.1 Gbyte data set. The DC dark current for each ramp is determined by subtracting the average of the upper and lower reference pixels for each stripe and fitting each pixel in the corrected frames up the ramp. The total noise for each stripe is determined by subtracting corresponding frames in adjacent ramps, masking out anomalous pixels, determining the variance of the difference frame and normalizing by two.

2.3 Dark Current and Total Noise Measured by Temporal Averaging

The temporal averaging technique requires a data set of enough identical ramps to significantly determine the properties of individual pixels. The standard UH data set consists of thirty six ramps, each made up of 2 resets followed by 145 frames. The raw, four dimensional (36 rampsx145 framesx 2048x2048) data approaches 44 Gbytes and takes 18 hours to accumulate. After flagging pixels affected by cosmic ray hits in individual ramps, the data set is used to generate a ramp by ramp data cube of dimensions 2048x 2048x36 for the required parameter. This typically accumulates charge and its sigma for each pixel, over a specified time interval using either CDS samples (which may be averaged) or slope fitting. These are then converted into dark current and total noise.

The dark current in the NIRCam arrays is so low (<0.002 e-/sec) that over V hour ramps its determination for individual pixels is dominated by read noise. For Fowler CDS averaging, we divide frames 9-144 of each ramp into four blocks of 32 frames each and then derive the dark current and total noise for CDS averages of 1-1, 2-2, 4-4, 8-8, 16-16 and 32-32 frames centered in the blocks separated by 384, 768 and 1152 seconds. Typical dark current and total noise histograms are presented in Fig. 1 and total read noise is summarized in Figures 2 and 3. Note that the 768 second values closely match the optimum signal to noise for CDS averaging over the nominal JWST 1,000 second exposure limit and that for all three time intervals, the 8-8 total noise is below the NIRSpec requirement of 6 rms e-.

Figure 1. Dark current and total noise histograms for 8-8 CDS at 768 seconds.
Figure 2. Standard deviation for JWST-002 at 3B4 sec, 76B sec, and 1152 sec.


Figure 3. Variance noise power for JWST-002 at 384 sec, 768 sec, and 1152 sec.


Figure 3. Variance noise power for JWST-002 at 384 sec, 768 sec, and 1152 sec.

2.4 Spatial versus Temporal Averaging - Consistency of Results and Relative Merits

Measured CDS 1-1 and 8-8 total noise values for JWST-002 are listed in Table 1 for the two averaging techniques (note that the uncertainties are for a stripe average in the spatial case and for individual pixels in the temporal case). The consistency between the two techniques validates the use of the simpler spatial averaging for screening purposes.

Table 1. Measured CDS noise (in ADU - 1 ADU is 4.33 ^V or ~ 1e-) for JWST-002.

Stripe 1

Stripe 2

Stripe 3

Stripe 4

CDS 1-1


12.50 ± 0.15

12.50 ± 0.05

12.60 ± 0.09

12.70 ± 0.10


11.74 ± 1.69

11.77 ± 1.71

11.99 ± 1.75

11.91 ± 1.73

CDS 8-8


5.51 ± 0.03

5.65 ± 0.09

5.70 ± 0.07

5.96 ± 0.07


5.03 ± 0.73

5.04 ± 0.73

5.37 ± 0.77

5.30 ± 0.78

The advantage of the spatial averaging technique is that it requires only a modest data set and computationally simple reduction to provide reliable estimates of stripe (or sub-array) total noise along with accurate estimates of individual pixel dark current. It is thus valuable for initial evaluation and screening of arrays. The primary disadvantage is that because dark current is so low it is a negligible factor in total noise, it provides no information on total noise in individual pixels. At the expense of increased time for acquisition and reduction of test data along with greatly expanded data storage requirements, the temporal averaging technique has been demonstrated to yield reliable estimates of total noise in each pixel.

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook

Post a comment