Anybody who has set up a satellite TV antenna will tell you that alignment is critical when picking up a signal from space. With a satellite dish it’s a straightforward task to tweak the position, but what happens if the dish in question is out beyond the edge of the Solar System?
We told you a few days ago about this exact issue currently facing Voyager 2, but we’re guessing Hackaday readers will want to know a little bit more about how a 50+ year old spacecraft so far from home can still sort out its antenna. The answer lies in NASA Technical Report 32-1559, Digital Canopus Tracker from 1972, which describes the instrument that notes the position of the star Canopus, which along with that of the Sun it can use to calculate the antenna bearing to reach Earth. The report makes for fascinating reading, as it describes how early-1970s technology was used to spot the star by its specific intensity and then keep it in its sights. It’s an extremely accessible design, as even the part numbers are an older version of the familiar 74 logic.
So somewhere out there in interstellar space beyond the boundary of the Solar System is a card frame full of 74 logic that’s been quietly keeping an eye on a star since the early 1970s, and the engineers from those far-off days at JPL are about to save the bacon of the current generation at NASA with their work. We hope that there are some old guys in Pasadena right now with a spring in their step.
How would it be built today?
The whole thing could be implemented in a small microcontroller with a built-in A/D converter. One of the ATTINYs could do it.
What it’s basically doing is looking at an analog input and seeing if it falls within a programmable range of values for a period of 2.56 seconds. As long as it is within that range a 12-bit counter counts up, and outside of that range it counts down. The main CPU can read the value of the counter to determine how far off it is from the target.
Would that last half a century and more in extreme conditions? Usually miniaturization is inversely proportional to longevity, and you could do it with a whole lot less than an ATTINY
Not a good choice for deep-space use. The tiny size of the transistors etc. make contemporary devices a poor choice. AFAIK, the most recent space-rated processor is a PowerPC device, specially fabbed.
SOS CDP1802, think there are some already on board and I hear you can still buy them from Renesas, but am sure there is some new old stock around.
It wouldn’t be. Uncomfortable but it’s the truth.
Yeah because there’s not much going on space-wise these days…
JavaScript.
They’d use a cloud service.
nebula service.
Almost certainly it would be many interconnected computers using SpaceWire to communicate. The computers would likely use RAD750 processors and some ASICs or possibly a space-grade FPGAs.
Dang it, this was a reply to the comment above.
Darn kids! always flailing away, punching in commands with all the participation trophy earned self esteem they can muster. Well I better start to get dad ready, that October reset will be coming up fast.
This only works in 1 axis. Does the spacecraft have two of these; one for each axis ?
The other axis is fixed using a sun tracker located in the extreme of the main antenna.
I worked with one of the guys responsible for the Apollo computers, and they had nothing but RTL gates in 14 and 16 pin packages. Pretty awesome.
Interesting. Thanks for the summary and the link, Jenny. ^^
I still don’t understand how this works.
Assume that you can unambiguously find the sun. That gives you one axis.
Now, you have to find Canopus. You probably have some idea of the angle between the sun and Canopus from wherever you are, so you don’t have to search the whole sky, but you *do* have to search a ring around the sun axis.
Depending on how much uncertainty your guess at the sun-Canopus angle contains (how good and how long ago was you last fix?) that ring might be a degree or two wide.
When you design your sensor, you’re either going to use a very narrow angle, in which case you risk the star *just* missing your FOV, or you’re going to use a wider angle. But then you’re going to see a *lot* of stars and there’s no guarantee that Canopus is going to cross through the center of your sensor. If it crosses near the edge, the transit could be brief and register as a fainter star. Or, since you’re only looking at one number, two dimmer stars might cross the sensor at the same time and add together to look like a bright star.
I suppose there could be something special about the space around Canopus that makes it easy to pick out against the stellar background, but this really feels like there’s some voodoo at work.
Yeah, star tracking has always seemed ridiculously… too hard…
Like, really? A few white dots amongst a million other white dots, and you can track off that? They all look the bloody same!
Let’s hope that the software for the antenna realignment was not written by one team calculating the angles in degrees and another team assuming the angles are in radians.
Unlikely (I think). The people in the 70s were different. The mentality was different, I mean. It was the hippie era, with people being thoughtful, into spiritual stuff and philosophy. At this time, the US was very open minded, too, I think. With reasoning almost on European level. That’s at least the impression what documentaries and YT videos provide. The 70s were a bit before my time.
Back to JPL/NASA: Those engineers at the time probably worked conscientious and double-checked everything. While they had felt time pressure, too, they still took more time for themselves to check and prepare things. In the 70s, I bet they had access to “stuff”, too, which helped them. ;)
Also, I assume, it’s not certain that the Star Finder worked with metric/imperial numbers. Maybe these were just increasing/decreasing values, without any relationship to a measuring system. The main task of each tracker was to find SOL and Canopus, rather. They swept over a certain field of view and compared minima/maxima of brightness information and sent a control signal if there was a close match with a pre-programmed value, while the computer was including the current rotation of the probe into the formula. Or the otherway round (ongoing signal up until there was a match). Or something like that. I have to admit that I didn’t read the documents very carefully. What I wrote is very simplified, thus. 😞
PS: I got the reference. ;)
Your thoughts on American culture being similar to American culture back in the 70s is not quite true. The media you watch is famous for pointing out the things that stood out back then. It wasn’t the norm, it featured what was different. Sure things look pretty cool, SST travel was going to be the norm, it looked like women were getting treated equally, you were seeing the beginnings of things.
At that distance, what is the angle between the Sun and Earth?
If less than 0.5 deg, then you’d only need to find the Sun?
That 0.5° might mean 10dB or more in the link budget.
And at those distances one want’s every scrap of dB they can get.
For the records:
From NASA:
UPDATE, Aug. 4, 2023: NASA has reestablished full communications with Voyager 2.
The agency’s Deep Space Network facility in Canberra, Australia, sent the equivalent of an interstellar “shout” more than 12.3 billion miles (19.9 billion kilometers) to Voyager 2, instructing the spacecraft to reorient itself and turn its antenna back to Earth. With a one-way light time of 18.5 hours for the command to reach Voyager, it took 37 hours for mission controllers to learn whether the command worked. At 12:29 a.m. EDT on Aug. 4, the spacecraft began returning science and telemetry data, indicating it is operating normally and that it remains on its expected trajectory.
My father was the inventor of the Canopus Tracker used on Voyagers and others! His name was Gerald W. Meisenholder. I have a model of the Canopus Tracker in our home here and I have some of his plans and patents. He has passed on but he would be absolutely thrilled to know it is still working! Thank you writing about it. Puts a smile on my face. :-)
It would be seriously cool to see some pictures! You could send them in on the Hack A Day tip line.
Sure! I would love to! Where do I find the “tip line”?
It’s here: https://hackaday.com/submit-a-tip/