The JWST NIRCam detector subarrays reduce data volumes and readout times, enabling rapid observations of bright objects without saturation.

A new subarray readout mode has been implemented for Cycle 4: multistripe subarrays are available to support the new NIRCam short wavelength grism time-series capability within the APT NIRCam Grism Time Series template. NIRCam detector multistripe subarrays expand the normal subarray capabilities by allowing multiple substripes (rows of pixels) to be read out, sampling different locations on the detector or re-sampling a single substripe multiple times, within the subarray frame time.

NIRCam users may either observe the full field of view (for a given observing mode) or read out smaller portions of the detectors, called subarrays.

Subarrays are read out more quickly than the full detector, allowing for shorter integration times. Shorter integration times can allow brighter objects to be observed without saturating the detector. See NIRCam Bright Source Limits for more information.

Each pixel is read out in 10.00 µsec. For most subarrays, pixels are read out one at a time (N_outputs = 1). For subarrays spanning the full width of the detector (2,048 pixels), 4 parallel output channels can be (and generally are) utilized for faster read out times (N_outputs = 4).

Thus, the total time to read out each frame of N_rows × N_columns pixels through N_outputs output channels (including small overhead delays due to telemetry) is:

for N_outputs = 4, 

t_frame = ((N_columns / N_outputs + 12) × (N_rows + 1) + 1) × 10.00 µsec

for N_outputs = 1,

t_frame = (N_columns / N_outputs + 12) × (N_rows + 2) × 10.00 µsec 

The smallest subarray available for science (64 × 64 pixels) can be read out in t_frame = 49.4 ms (the shortest integration time, though multiple groups of readouts are recommended: N_groups > 1). The full 2048 × 2048 array is read out through 4 output channels in 10.737 s.

When observations are obtained simultaneously at short and long wavelengths, subarrays with identical numbers of pixels are used in both wavelength channels. Each long wavelength pixel covers 4× more area on the sky than each short wavelength pixel. So in some cases (e.g., SUB400P ), the long wavelength subarray covers 4× more area on the sky than the short wavelength subarray. In other cases (e.g., FULL or SUB640 ), 4 short wavelength detectors combine to roughly cover the same area observed by one long wavelength detector. Subarrays ending with a "P" are designed for point source imaging (or for compact object imaging), and therefore have uneven coverage between the short wavelength and the long wavelength channel. 

Each detector has 2048 × 2048 pixels consisting of a central block of 2040 × 2040 pixels sensitive to light and a 4-pixel wide border of reference pixels along all edges used for calibration. All subarrays positioned along edges include reference pixels, subtracting slightly from their total area available for science.

After each subarray integration, the subarray pixels are reset pixel-by-pixel, then the remaining detector rows outside the subarray region are quickly reset, one row at a time. While less effective at clearing out charge than standard individual pixel resets, this scheme should effectively mitigate latent images that might otherwise build up and leave persistence on portions of the detector outside the subarray region. The row resets contribute slightly to the overheads, generally less than a percent of the total integration time. When using the FULL and FULLP subarrays, all detector pixels are reset individually, thus no additional row resets are required.

The exposure times reported by APT and ETC include time for resets. Subarrays require both pre-exposure pixel resets (for all pixels in the subarray) and post-integration row resets (for additional numbers of rows specified below). The full detector requires neither reset for the first integration of an exposure; the pixels are reset between integrations.

Figure 1 and the tables below summarize the supported subarray sizes, frame times, and fields of view for the NIRCam observing modes. Subarrays are defined on one NIRCam module or the other (A or B), chosen for optimal performance for each observing mode.

---

Imaging

All imaging subarrays are on module B.

Table 1. Imaging subarrays

---

Coronagraphic imaging

All coronagraphic imaging subarrays are on module A. The first 2 rows are for target acquisition, rows 3 and 4 are for target acquisition astrometric confirmation images, and the last 3 rows in the table are for occulted images at short and long wavelengths.

When observing with the FULL subarray, short wavelength coronagraphy returns data from all 4 detectors. Smaller subarrays return data in just 1 detector per channel.

Table 2. Subarrays for coronagraphy and target acquisition

---

Time-series imaging

Time-series imaging subarrays are on module B. Each observation obtains data on one short wavelength detector and one long wavelength detector.

Table 3. Time-series imaging subarrays

---

Grism time series 

Grism time-series subarrays are on module A. In this mode, the user is offered a choice between one and 4 detector outputs. Each observation obtains data on one long wavelength detector and 2 short wavelength detectors.

Table 4. Subarrays for grism time series

---

Wide field slitless spectroscopy

Subarrays are not offered for grism wide field slitless spectroscopy (WFSS). Note grism spectroscopy is obtained in the long wavelength channel only, while imaging is obtained in the short wavelength channel. Additional supporting images are also obtained at long wavelengths.

Table 5. WFSS subarrays

---

References

Norwood, J., Hammel, H., Milam, S. et al. 2016, PASP, 128, 025004
Solar System Observations with the James Webb Space Telescope

---

---