One comment that caught my attention, though, was that the computers onboard the rovers are similar to home computers from about 15 years ago. I knew the shuttle's computers use old technology but was surprised to hear the rovers suffer from the same problem.
It's not nearly as much of a problem as you might think. An awful lot of a home computer's energy is 'wasted' on tasks like managing a display and other chores which aren't necessary for a typical embedded computer. I've spent quite a few years writing software for embedded processors and it's really quite surprising how much you can get done with the right software architecture. A consideration for a spacecraft's computer which is every bit as important as power is survivability. These things must survive extremes of temperature and radiation--not to mention a fair bit of physical abuse. Some processors are 'hardened' by special aspects of chip construction, others by their enclosures. When you think of the absolute hell that some of these craft fly through, it's amazing they work at all. Given the expense of a service call, it's not surprising that a flightworthy processor gets used for quite some time.
--- Michael Carnes <michaelcarnes@earthlink.net> wrote:
I knew the shuttle's computers use old technology but was surprised to hear the rovers suffer from the same problem.
I am not familiar with the specific processors used on the Rovers but I would be a bit suprised if it was really seriously out of date. The company where I work provided Inertial Measurement Units (IMUs) for the Rovers. IMUs measure acceleration (including gravity of course) and rotation in all three axes. There were two per mission, one on the backshell and one in each Rover (the last I heard these are still functioning very well). The units that JPL got were pretty much the standard current generation product with fiber optic gyroscopes. About the only modification we did for them was to replace the standard RAM with a nuclear hardened chip (and did a bunch more testing). We delivered them to JPL not all that long before Spirit's launch, (I'm quite sure it was less than a year). Point being, JPL went out an found a current product that met their requirements and used it rather than spending years developing something that was out of date by the time it was used.
It's not nearly as much of a problem as you might think. ... I've spent quite a few years writing software for embedded processors and it's really quite surprising how much you can get done with the right software architecture.
Michael is very correct on this point. We also happen to make the guidance systems/computer for the Navy's Tomahawk and the Air Force's Air Lunched Cruise missles. The use very old technology processors (similar to the original DEC PDP-11) to get very good results.
A consideration for a spacecraft's computer which is every bit as important as power is survivability. These things must survive extremes of temperature and radiation--not to mention a fair bit of physical abuse. Some processors are 'hardened' by special aspects of chip construction, others by their enclosures. When you think of the absolute hell that some of these craft fly through, it's amazing they work at all. Given the expense of a service call, it's not surprising that a flightworthy processor gets used for quite some time.
The standard IMU that JPL bought is qualified and calibrated to function correctly from -54 to +71 deg C and in vibration environments of up to 15 g's (rms). So it works over a lot bigger range of conditions than your household computer would. ( I already mentioned that we added the nuclear hardened memory that most customers don't have much need for.) p.s. - If anyone is curious about what these IMUs look like or are spec'ed to do, the product brochure can be found at: http://www.nsd.es.northropgrumman.com/Html/LN-200/brochures/ln200.pdf ____________________________________________________ Start your day with Yahoo! - make it your home page http://www.yahoo.com/r/hs
participants (2)
-
Ivo Stutznegger -
Michael Carnes