Emily Lakdawalla posted an article in February describing challenges facing the Mars Reconnaissance Orbiter and the High Resolution Imaging Science Experiment (HiRISE), based largely on a NASA press release. There were two especially worrisome issues for the HiRISE team and user community: (1) blurred images; and (2) MRO battery issues requiring a future move of MRO’s orbit to a later time of day, leading to lower-quality images. There is now good news about both issues.
Blurred HiRISE images
HiRISE images taken in 2017 and early 2018 showed blurring not seen earlier in the mission. The percentage of full-resolution images with blurring peaked at 70% in October 2017 at about the time when Mars was furthest from the Sun. That percentage declined to near zero as Mars moved closer to the Sun, indicating a thermal effect. The camera has an active thermal control system (TCS) that keeps the telescope optics and structure near 20° C. To preclude possible transient noise, the original control software disabled the TCS during science image capture, which lasts for only a few seconds, not enough time to affect the optical performance.
NASA / JPL / UA
Unexpected blurring in HiRISE orbiter images
Beginning in early 2017, the HiRISE team noticed unexpected blur in some images. This comparison of two images of part of Gusev crater taken March 16, 2016 and January 9, 2018 shows that the later image is blurred. Blurring of the second image reduces its value, because it was taken as part of a campaign to map future potential landing sites at high enough resolution to help mission planning.
HiRISE has had another issue: bit flips in the analog-to-digital converters for some of the 26 image channels when the focal-plane electronics (FPE) are relatively cold. These bit flips create artifacts in the images, which have worsened over time and can lead to useless images. We mitigate the bit-flip problem by warming up the FPE before imaging Mars, which requires operating the FPE in imaging mode. Thus, we acquire long-duration warmup images that are not returned to Earth. Initially we did these warmup images in the middle of the night side (MRO in eclipse), when MRO operates on battery power. But HiRISE imaging uses a lot of power, and when the lifetime of the MRO batteries became a concern we moved the warmup images out of eclipse to just before the Mars images. That meant that the TCS was off for significant time periods before most Mars images, without enough time for all telescope elements to recover to 20° C. The engineers at Ball Aerospace in Boulder, Colorado, who designed and built the HiRISE camera, suspect that there are temperature gradients that lead to out-of-focus images.
The temperature-gradient theory for the blurring suggested a simple solution: change how the camera operates so that the TCS stays on during imaging. We implemented that fix last summer, but since Mars was near its closest point to the Sun we could not test whether or not the blurring would re-appear when the environment is colder. Mars’ range to the Sun has been increasing since last spring and the environment is now cold enough to produce blurring unless the fix worked. We have seen no blurred images, so it now appears very likely that the blurring problem is gone (hurray!).
The NASA press update about MRO said “To prolong battery life, the project is conditioning the two batteries to hold more charge, reducing demand on the batteries, and is planning to reduce the time the orbiter spends in Mars' shadow, when sunlight can't reach the solar arrays. The spacecraft uses its batteries only when it is in shadow, currently for about 40 minutes of every two-hour orbit.” MRO has operated near a local mean solar time of ~3 p.m., which provides optimal lighting for topographic shading while maintaining a high signal level for imaging. However, with battery degradation, spending ~40 minutes in the shadow of Mars on every orbit is risky, and indeed the spacecraft experienced a couple of scary episodes. One solution, already applied to Mars Odyssey, is to move the orbit to a later time of day near the terminator (day-night boundary) so that the spacecraft is in sunlight for most of every orbit. In such an orbit HiRISE would need to acquire mostly 2x2 or 4x4 binned images to acquire sufficient signal, degrading the spatial resolution. We could still acquire “full-resolution” (unbinned) images but the poor signal-to-noise ratio effectively reduces resolution anyhow, and with 2x2 binning we can cover almost 4 times as much of Mars’ surface for the same data volume to be returned to Earth.
The good news is that the battery reconditioning has been very successful, pushing the need to change MRO’s orbit off at least several years into the future. This means that HiRISE can continue to acquire images at 25-35 cm/pixel scale, and at similar lighting angles to prior images, which is best to detect changes from active processes.
MRO has enough fuel for at least another decade of operations at Mars, and the battery situation has improved to the point where another decade of life seems likely. Possibly MRO will eventually need to move to a later time of day, but not for at least several years. The bit-flip noise continues to worsen, so we will continue to increase the warmup imaging length to reach the FPE temperatures we need, and estimate that we can continue this mitigation for at least 10 years before reaching temperatures that are not considered safe. The high FPE temperatures lead to high CCD detector temperatures, beyond the range of our current radiometric calibration. We should derive new calibrations and reprocess the images, which is difficult with constantly shrinking support NASA gives to extended missions, but this can happen anytime in the future.
In summary, there are no known reasons to expect that good HiRISE imaging will end in the next decade. MRO and HiRISE are well beyond their design lifetimes, so there are no guarantees, but note that Voyagers 1 and 2 are still alive almost 42 years after launch, and GOES-3 operated for 38 years in geostationary orbit (but lost its imaging capability 10 years after launch).