Campaign Features and Events
Not In Fine Point Data
The second half of the C0 data is more indicative of the quality of data users should expect from K2. The Kepler spacecraft was not in fine point for the first part of C0, causing large photometric scatter. The data quality is much improved in the second half of the campaign, beginning on cadence 89347 after the safe mode, when compared to the first half of the campaign. See the QUALITY flag (bit 16) to determine when the spacecraft was in fine point.
The Kepler Spacecraft was in safe mode between cadences 88198 and 89346. Data is flagged in the QUALITY column with bit 2. The project used this time to fix large attitude errors that were occuring during resaturation events.
Photometric Jitter Caused by Lower Bandwidth
During the development of the K2 Mission, the number of guide stars per fine-guidance sensor (FGS) was changed from ten (for Kepler) to one (for K2). This change was demanded by the need for increased aperture sizes given the uncertainties in the star-tracker to boresight alignment and the need to acquire an entirely new field-of-view every 80 to 90 days. To compensate for the increased sensor noise and assure that fine-point lock could be achieved, the attitude control bandwidth was decreased from 0.1 Hz (for Kepler) to 0.02 Hz (for K2). This change in bandwidth means that the cross-boresight attitude (i.e., RA and Dec) has a time constant of 50 seconds, comparable to the short-cadence duration. Engineering studies have shown that the photometric precision of the long-cadence data is also compromised by the larger pointing jitter associated with the lower bandwidth. The project has approved a change in bandwidth to 0.05 Hz (20 seconds) starting with Campaign 3.
Large Pixel Masks
When planning C0 observations, the pointing performance of K2 was not accurately known. The worst case scenario was that a star at the edge of the focal plane could move as much as 40" from its nominal position. Therefore each star was assigned a large pixel mask by first computing a Kepler-style optimal aperture and then adding 10 rings of pixels to account for a potential 40" pointing offset. During the second half of C0, the pointing performance was excellent and the pointing drifts were no more than 6" for any target star. Care will be needed when performing photometry on C0 data. Simply including all collected pixels for a given target will not create a high signal-to-noise light curve. For tools to help choose your photometric aperture, see for example, PyKE contributed software.
Because K2 points along the ecliptic, its field of view will occasionally contain bright solar system objects. Jupiter was in the K2 field of view during C0 from 2014-03-14 through 2014-05-12, but fell on dead module 3. It creates a bright antipodal ghost on module 23, channel 79, and impacts all the targets observed in this region. See the FFI ktwo2014074233223-c00, extension 79, for an image of the reflection.
While Jupiter was on the focal plane, the background level was increased over its nominal value for nearly half the channels, with the largest impact seen in channels 53 -- 84. In addition, as Jupiter moved on and off the focal plane, a specular reflection lasting approximately 6 hours was seen in these channels. The relative background levels as measured in the smear signal from channel 83 are shown below as Jupiter enters the focal plane (near cadence no. 87525) and leaves the focal plane (cadence no. 90375). The specular bump resulted in an increase in background level of 25-30% for the affected channels, while the quasi-static background increase for the time Jupiter was on the focal plane was 3-5%.
FFI showing the reflection of Jupiter as seen on channel 79.
The background level on channel 83 as Jupiter enters (left) and leaves (right) the focal plane.
Observations of M35 and NGC 2158
The open clusters M35 (NGC 2168) and NGC 2158 were observed during this campaign by placing 154 separate 50x50 pixel masks over the densest portion of these two adjacent clusters. Each mask was given a custom aperture number to act as the unique identifier found in the file name. The target pixel files for these clusters have custom aperture numbers ranging from 200,000,811 to 200,000,964.
Individual masks when tiled together cover the field of view containing M35 and NGC 2158.
Channel 10 Black Correction
The black correction on Channel 10, mod.out 4.2, has problems because a bright star bleeds into the black region. For data release 1, the affected regions of the black are excluded from the fit of the black, so the fit is poorly constrained and we see chatter in the residuals. This mostly affects stars in the last 200 rows of the channel.
Image Artifacts in K2
The thermal environment is changing more rapidly in K2 than it did for the Kepler Mission. As a result the number of channels with significant rolling band (changes in the black level that are both time and spatial dependent) is larger for K2. The channels observed to be most impacted by rolling band in C0 are 1, 2, 10, 11, 14, 25-28, 36, 44, 58, 62, 74 and 79. Other known image artifacts, such as Moiré patterns and undershoot from bright stars, are also likely to be enhanced in K2 data as compared to Kepler. See the Kepler Instrument Handbook.
Module 7 Failure
Prior to the start of C0, on January 21, 2014, the photometer was autonomously powered off by an under voltage fault in the Local Detector Electronics Power Supply. Since that time, module 7 (i.e., channels 17 to 20) has yielded no star data or charge injection signal. The subsequent behavior of this module is very similar to that of module 3 after it failed on January 19, 2010. K2 continues to operate and collect simultaneous data from sources falling upon the remaining 19 detector modules over 105 square degrees. There is no indication of any accelerated degradation on these other modules.
Cropped Calibrated Pixels in Data Release 1
The problem described below was corrected in Data Release 2 of the Campaign 0 data. See the DATAREL keyword in the primary header of the target pixel file to determine which data release you have. We recommend that everyone use the latest release of the data.
In Data Release 1, for cadences after the Safe Mode (cadence no. 89346), the calibrated
data (see the FLUX column in the exported target pixel files) is
incorrectly reporting a value of zero for pixels collected near the edge of the aperture. The raw pixels (see the RAW_CNTS column) do not have this issue, so the collected data is properly reported in the target pixel files. An example of the issue is shown in the figure below. The left image is the mask (white indicates a collected pixel), the central image shows the calibrated pixels and the right image contains the raw pixels
for one cadence. Those columns and rows that are black in the central
image incorrectly have a value of zero, but there is clearly data in the original raw data on the right.
This issue occurs for both the long and short cadence data. It does not appear in the custom apertures, so the calibrated pixels are being exported as expected for these files.
Image of mask (left), calibrated pixels (center) and raw pixels (right) for one cadence of one target.