Launched in 1977, the Voyager 1 and 2 probes were both cutting-edge pieces of technology for their time. The computers at the heart of their operations consisted of three systems, each with dual-redundancy, that worked together to enable the probes to journey to Jupiter, Saturn, and beyond: the Computer Command System (CCS), the Flight Data Subsystem (FDS), and the Attitude and Articulation Control System (AACS).
What is amazing is that even after four decades of traveling through the harsh, sometimes unpredictable, environment of space, both probes continue to function and call home with new insights and data. It is taking longer and longer to be able to communicate and upload new routines to the probes, but the fact it is still possible with technology from a bygone era is a testament to the quality of engineering put into these spacecraft.
The story of how the Voyager computers took shape is a fascinating one. Coming together during NASA budget cuts after the excitement of the Apollo era faded, and overcoming challenges not yet encountered by engineers in exploring some of the most interesting places in our solar system, the details will surely give one a greater appreciation of the Voyager computers.
My favorite recent Voyager computer anecdote was from I think a recent (past year or so, āThe Farthestā I think it was called) show about Voyager where someone commented (paraphrasing):
āMany people today carry more computing power in their pocket than what is on Voyager. And Iām not talking about your cell phone, Iām talking about your carās key fob.ā
But that computing power comes a cost, buggy hardware and software.
At one time rockets exploding, killed a space project now it is software
and hardware problems.
I donāt know if youāre aware of this, but they pretty entirely reprogrammed the Voyager crafts once they pass Jupiter. They uploaded large amount of new mission profiles since they were able to change the actions of the probes with 10, 20, 30 years of hindsight for new missions.
Software has bugs. Hardware has bugs. Theyāve always had bugs. Itās why so much time is spent in the Engineering part of software engineering, especially on critical systems like this.
But even with all that, stuff slips through. Cāest la vie.
Indeed. A recent case involved the New Horizons probe as it approached Pluto. Just days before the encounter, New Horizons went into safety mode due to an programming bug (CPU got overloaded), which meant that they not only had to recover the craft but also reload the fly-by program schedule. And they only had one chance to get it right, with the hours-long time delay between Earth and the probe.
The details are in Alan Sternās book, āChasing New Horizonsā. This excerpt gives a taste of the urgent task the team had to do (note: the excerpt is behind a āplease remove your ad-blockerā notice).
As Voyager 2 has been in the news - a scheduled series of commands caused it to lose orientation and therefore lose contact for a while - I thought Iād add some more resources.
Contact was regained, as I understand, by increasing transmitted power at this end so new commands could be received even with the antenna mis-pointed. The backup plan was that the spacecraft would autonomously re-orient itself, which it is set to do on a regular schedule, and this would have happened in October. The means of doing that involves no computer, but specific digital logic which seeks the star Canopus. More here:
Thereās a book on the computers used at NASA from the beginning through to when it was written - late 80s, I think. Whole thing is available online
It has a chapter titled āVoyager - The Flying Computer Centerā. It gives a high-level overview of the computers and software. Three different processors, each dual redundant. 18 and 16 bit machines. Comparable to early 1970s minicomputers.
I used to use the Voyager bit error from 2010 as a topic in an introductory lecture from a freshman seminar that I was involved with, as an example of good solid embedded systems design.
In May of 2010, a cosmic ray or other disturbance unexpectedly flipped a bit in the flight data system computer on Voyager 2, which caused it to stop sending valid scientific data back to Earth. By diagnosing the data that was coming back, NASA engineers were able to isolate the bit, and send the command to flip it back, then reset the spacecraft. At that time, Voyager was about thirteen hours from Earth at the speed of light, so the reset-command-to-confirmation-of-success loop was just over an Earth day. What a harrowing window that must be.
