Slashdot Mirror
Design, Hardware, Software Errors Doomed Japanese Hitomi Spacecraft (scientificamerican.com)
Reader Required Snark writes: The Japanese space agency JAXA said its recently launched X-Ray observation satellite Hitomi has been destroyed. After a successful launch on February 17, contact with the satellite was lost on March 28. Off the 10-year expected life span, only three days of observations were collected. Preliminary inquiry points to multiple failures in design, hardware and software. After the launch it was discovered that the star tracker stabilization didn't work in a low magnetic flux area over the South Atlantic. When the backup gyroscopic spin stabilization took control, the spin increased instead of stopping. An internal magnetic limit feature in the gyroscope failed, causing the spin get worse. Finally, a thruster based control started, but because of a software failure the spin increased further. The solar panels broke off, leaving the satellite without a long-term power supply. It seems that untested software had been uploaded for thrust control just before the breakup. This is a major loss for astronomical research. Two previous attempts by Japan to launch a high-resolution X-ray calorimeter had also failed, and the next planned sensor of this type is not scheduled until 2028 by the ESA. Just building a replacement unit would take 3 to 5 years and cost $50 million, without the cost of a satellite or launch.
it sounds like everyone is starting from scratch every time a project like this is built.
regardless of success or fail, wouldn't it be best for everyone to release the engineering and software so that the next one is an improvement over what went before.
it also might reduce the reduce the life cycle of the next project.
just my .01999 USD
Those are called political and budget pressure by managers who have no clue on engineering ---
Software uploaded with out testing ? There is no way they could have gotten this far with out testing. I am sure there is no engineer in Japan that does not test thoroughly. Actually Japanese code is famous for being of the best quality -
This was caused by politics, bureaucracy and plain bad management.
It was probably running Linux, first mistake.
Nah; it was probably running ITRON. It may well have included a POSIX library, but that wouldn't qualify it as a version of linux, even if some linux code is included there.
I haven't actually bothered to dig up the info, but that's what anyone acquainted with how such things are done in Japan would guess for a situation with serious RT requirements. Maybe it'd be interesting to investigate, to get an idea whether the OS and system libraries might have had anything to do with the failures.
Those who do study history are doomed to stand helplessly by while everyone else repeats it.