NASA is struggling to fix a computer glitch that has sidelined the iconic. While the problem is an uncomfortable reminder of the aging observatory’s eventual mortality, engineers are confident they’ll have it back up and running soon.
“Hubble is arguably the most important asset in the NASA’s astrophysics portfolio, and it’s been doing world-breaking science for over 30 years now,” Paul Hertz, director of astrophysics at NASA Headquarters, said Wednesday. “And we’re counting on it operating for many years more.”
The problem cropped up June 13 when the telescope’s instrument-overseer payload computer suddenly stopped working. That triggered protective “safe mode” software that halted operations and effectively put the telescope in a state of electronic hibernation pending analysis on the ground.
Since then, engineers have been working to identify which component in the Science Instrument Command and Data Handling — SI C&DH — system is not working properly, determine if it can be recovered or whether to switch over to a backup.
“The great thing about Hubble is all the redundancy we built in,” Hertz said in a telephone interview. “We’re all confident that once they’ve identified which component it is that isn’t working, we’ll switch over to the redundant version of that component.”
While it sounds straightforward, NASA is taking no chances with one of the most scientifically productive spacecraft in the agency’s inventory. Hertz said he expects it to take a few weeks to fully resolve the problem and returnto science operations.
“Because it’s so important, we’re making sure that the troubleshooting team is taking every opportunity to be careful,” he said. “And so, they are over-analyzing all the data they have, they are developing decision trees and then working through a list of tests … so they can isolate which particular component it is.”
Once the culprit is identified and a fix implemented, “we’ll bring the instruments back up and get back to science,” he said. “Because of all that redundancy, we’re all very confident that we’ll get Hubble back to operating.”
Astronomer Adam Riess, who shared the 2011 Nobel Prize in physics for work he did using Hubble to help confirm the accelerating expansion of the universe, certainly hopes so.
“HST is THE workhorse for me and the kind of research I do,” he said in an email exchange. “When it goes offline it makes you reflect on just how invaluable it has been (as an interlude between your prayers that it comes back to life soon!).
“Hubble has been a cat of nine lives, and we hope it’s got a couple left in its tank.”
The SI C&DH payload computer system is made up of 14 components that control the telescope’s science instruments, format and temporarily store their data and then help route that data to the ground. Two fully redundant systems are in place, either one of which can operate Hubble’s instruments.
The original 1980s-era hardware operated for 18 years, from launch until 2008, when one of its two processors failed. The telescope was successfully switched to the backup, but NASA managers decided to replace the entire system during the final shuttle servicing mission in 2009 to restore the lost redundancy.
One data processing channel, or “string,” worked from then until the malfunction earlier this month. Components in the backup system have not been turned on since launch 12 years ago.
“This is the computer that queries all the instruments, grabs the data that they’ve taken out of the instrument memory, packs it up into the SI C&DH memory, and then hands these packets off to the spacecraft, which puts them in the telemetry stream and sends them down to Earth,” Hertz said.
“If it’s not working, you can’t get any data out of the instruments. Because it’s so necessary, that’s why it’s fully redundant.”
Engineers initially suspected the problem involved one of four 64K complementary metal-oxide semiconductor — CMOS — memory modules shared by the computer strings.
Only one module is used at a time and shortly after the shutdown, engineers attempted to restart the computer using a different memory bank and then again with the original. But the computer failed to start.
Diagnostic data from the attempts indicated the memory issue may be symptomatic of a problem in a different part of the computer system.
One possibility is a standard interface unit, or STINT, that routes data between the computer’s central processing module and other components. Another possible culprit is the central processor itself. Engineers currently are running tests to pinpoint the problem area and come up with a fix.
Depending on the results, they could attempt to switch over to the STINT and central processor in the backup payload computer string, also built using 1980s-era hardware. The backup components all passed their initial pre-launch tests and engineers are confident they will work if needed now.
“We’re professional worrywarts,” Hertz said. “So I think it would be overstating to say nobody has concerns. But what we do know is that a dozen or so times during the lifetime of Hubble (protective software) has turned components on and off because of a problem that is autonomously detected.
“We haven’t had any problems like that with turning things on and off on Hubble, and so we don’t anticipate that problem this time.”