"Uptime 15,364 days - The Computers of Voyager" by Aaron Cummings
Vložit
- čas přidán 7. 06. 2024
- The Voyager 1 and Voyager 2 space probes, both launched in 1977, each had a primary objective to explore Jupiter and Saturn. This goal was achieved by 1981. Yet Voyager, NASA's longest running mission, has continued to this day. Both Voyager probes are still operating, and returning scientific data from outside our solar system.
This talk explores the computing systems of Voyager - the systems which enable remote control of the spacecraft, and provide for the recording and return of data to Earth. These systems have proved to be both adaptable, durable, and resilient in support of a scientific undertaking now in it's fifth decade.
What can we learn from the engineering of Voyager's computing systems? Why have they survived for so long in the harsh environment of space? What is involved in patching a system from a billion miles away? And what does the future hold?
Aaron Cummings
@btvaaron
Aaron Cummings is a software developer working in the semiconductor industry, currently leading a team working on tools for building and testing embedded memories. He has had a long term fascination with the space program, and has been interested in Voyager since seeing the pictures returned from Jupiter and Saturn in the 1980s.
Chapter listing
00:00 Background on Voyager
01:16 Talk outline
02:00 About the speaker
02:25 Origins of Voyager
03:18 Initial requirements
08:25 Revised program
10:05 The 3 R's
11:08 Probe design
12:55 Command Computer System
17:40 Attitude and Articulation Control System
20:04 Flight Data System
23:56 Voyager 0
24:35 Voyager 2 mission
29:07 Sustaining power
30:26 The golden record
31:20 Q&A - Věda a technologie
Brings back memories. I was on the mission control team (data systems engineer, DSE) for the 1st 3 encounters, and my duties were to monitor the CCS, FDS and AACS in real time. Best job I ever had. I went on the be a mission controller for Magellan and Galileo. Nice talk.
wow, that must have supercool to be a part of!
Well here is a dumb question ... when you saw Star Trek The Motion Picture was Voyager's (Voyager 6) appearance a surprise or were your bunch in the know beforehand?
Must be really cool to be part of something that will likely be the only remaining artefact of humanity in the very distant future.
I'm envious. Awesome!
Wow! Do you know if FORTRAN was used for the CCS or was it all assembly?
Voyager has been flying for 2/3 of the entire space age, it launched 20 years after Sputnik and now it is 62 years after Sputnik. In 500 years it will be remembered like Magellan's circumnavigation is remembered today. The Great Mission of it's generation.
Thank you SO MUCH for first repeating the first few questions before answering them. Speakers need to remember that the audience is not mic'd, and it's often difficult to discern the question from just the answer.
This is an excellent lecture. Not everybody cares about all these details, but for those of us who do this was terrific. Kudos!
It's amazing that humans can actually build something so delicate that'll last this long being operational without direct maintenance.
Simplicity and redundance.
Imagine being such an unaffiliated fan of a space mission that you get to deliver a conference talk about it. Super cool :)
For being "just a guy", you did a very nice job. Thanks for the very informative presentation, I liked it a lot.
I’d love to see a Voyager 2.0 series of missions. High definition sensors, modern computation and the ability to last 1000’s of years. Then send it to Proxima B.
It would take 30,000 years to get there... so most of us won't sit around and watch it. : )
Thanks, Aaron, very nice. I recently worked with a guy who had the job of testing one type of the Voyager uplink RF bandpass filters, and he found that they would not have survived the vibration on launch. He then personally participated in potting and testing both filters that flew, and his initials are scribed into the test certification stamps on the sides of the actual V1 and V2 uplink bandpass filters. So my friends initials are enshrined on some actual flight hardware, outside the solar system. Pretty neat nerd stuff, eh? Again, thanks for your presentation!
11 KHz clock, 32 interrupts, 64 instructions set, 9 Mb of memory, 62 Mb hard drive. Amazing how far we've come.
Tape drive, not hard drive.
That was a huge amount of RAM for that time period. At that time, 1KB of SRAM was over $100. The real clock was likely 10-20 times the frequency you're thinking. The 11 kHz is instruction cycle time, many clock ticks to make that happen.
@@mikebrown7366 it looks like a supercomputer for the '70s.
Remarkable and amazing they are still sending back data after 43 years in the hostile frigid vacuum. What a testament to the brilliant engineers of that era and the people who made the parts and built them. What a legacy they have left us and what an inspiration to do more with today’s technologies. The robots are far more productive than manned space flight and a whole lot cheaper and safer. They were launched two years after I graduated engineering school and I have seen the planets in a way only dreamed about for thousands of years. Truly a good time to be alive!
Great presentation, Aaron! "Don't neglect your project after launch". Love it! Legacy application running on legacy hardware and still kicking it! =)
Where has this presentation been for the past 20 years i've been interested in Voyager.
This channel... i watched 2 videos and i totaly got what was said. That is simply amazing that such complex things can be explained in such a simple way... i do aprisiate it !
That was exceptional, thank you :) I've always been fascinated by Voyager 1&2. I'm not above saying I will shed a tear when we finally lose contact, but who knows, maybe they will return with friends one day.
Fantastic talk - I find the engineering, the workarounds, the solution finding for the Voyager probes incredibly inspiring.
All possible thanks to Plutonium-238 , Smokin HOT!
Thank you Mr. Cummings for a great, educational, and informative presentation. I’m still amazed, after all my years on our planet, of the accomplishments made in the space programs. ☺️
@3:55 👍👍👍 Grace Hopper reference!
What an amazing talk, I love reading about these space crafts.
RTGs are built SO strongly, that we've actually had ones crash back to Earth after rocket failures, recovered them, and then re-used them. This happened with Nimbus B-1 failure.
Not every is into this kind of stuff. But for those who are enthusiasts in this field, this is pure gold !!!
I've worked on the Voyager project since 1979 and still gather data from the Voyagers on almost a daily basis. BTW, we *do* receive "high speed" data from Voyager 1 recorded on the tape and played back at a reduced rate. Here's a sample of these data which confirmed the transition into interstellar space: czcams.com/video/CFklHHWbYrk/video.html
That scale is what? 10^-5 watts received at the DSN? That's like hearing somebody whisper in a tornado and still being able to tell what they're saying. I'm surprised the tape drive is still working after all these years.
Thank you for your work
@@BlackEpyon it's around 10^-19 W.
Amazing presentation! Thank you Aaron.
Thank you for the Back to the Future reference. :-)
When I last at Nasa Cape Kennedy Space Center, in Florida, the tour guide mentioned at the time that for every dollar invested in space has returned seven dollars in technical advances. I was quite surprised to hear that, but you see the result with modern computers everyday.
IC’s were developed first for ICBM’s, which I guess are nominally space craft.
When Gerard O'Neil asked his class to estimate the ROI on the Apollo program to everyone's surprise (including his) the found that *just* the taxes on *just* the improvements on *just* the telecom industry would have paid for Apollo by the mid-'80s.
I could not have asked for a better presentation. Thanks.
It is really beyond my grasp the immense quality and quantity of thinking, knowledge, expertise and experience, insight, design, technology, dedication, vision, effort and cooperation that was needed from so many scientists and technicians of various disciplines, workers and volunteers to come up with this ultimately superb for its time earth-machine that is still operating 42 whole years after it left our planet. It is the first palpable interstellar evidence that is sent out to the "great beyond" manifesting the presence of humankind in the universe.
This is an eye-opening speech from a very knowledgeable yet humble scientist ("I'm just a guy..."), whom I respect deeply for his stance.
Thank you very much for this upload!
WOW, didn't know all of this. GREAT JOB!
Thank you for that great presentation!
Great background on the spacecraft and the mission.
Excellent talk. Loved the Voyager missions. "Old school" tech was quite robust.
Amazing presentation! Well done, thanks a lot.
Very enjoyable presentation. Thank you.
This is amazing. Thank you.
As a sys admin seeing that uptime fills me with rage!!!! Also as a sys admin seeing that uptime makes me say "Don't ever reboot it!" 😉
Loved this! Very nice presentation and obviously a very knowledgeable fella on this subject
Absolutely amazing engineering that is from my perspective unparalleled even up until today. Thanks for the upload
www.goonhilly.org/antennas/ghy-1-arthur You should check this out another amazing work
I'm so glad someone asked about the use of the computer system redundancy. As soon as he mentioned they had 5 cores and a management system that could boot cores out if they became unreliable I really wanted to know if any of the cores actually had gone bad or if the redundancy was being used similar to how the radio had failures. Sounds like once the plutonium is nothing but a lump of cold lead and somebody stumbles upon it they'll just be able to power it back up and run it again!
I really liked this!
Fascinating talk. Was always curious about the computer systems on the two oldest and furthest space probes
Very interesting! Thank You!!
I've always been interested in the actual hardware design of the Voyager computers. Wonder whether there is a more detailed breakdown. Thanks for this. Cheers.
I remember that it has 16 kilobytes of memory
This one is about Apollo Guidance Computer. Not Voyager, but still - a very interesting sneak peek into the hardware of the era czcams.com/video/2KSahAoOLdU/video.html
If you can't find documents NASA or others have put online, you can also often research those documents yourself through NASA's Technology Transfer Program. For access to Apollo- and Voyager- era detailed documents, you'll need to be a US citizen and resident. They do put a lot of software and information out there for worldwide distribution, but when I researched Apollo documents in March 2017 through the program, US residency was required.
They update their catalog every two years, so the links below are a bit stale relating to the third release covering 2017-2018 made in early 2017. I imagine there was a fourth dump made in early 2019.
___________________________________________________________________________
SOURCES
[1] NASA Technology Transfer Program channel on CZcams, “Download NASA Software for Free!”, 3/1/2017:
czcams.com/video/JIpyc8AfMZY/video.html
[2] NASA Technology Transfer Program website (start here for application and download):
software.nasa.gov
[3] NASA Technology Transfer Program, “NASA Software 2017-2018 Catalog”(PDF):
technology.nasa.gov/NASA_Software_Catalog_2017-18.pdf
[4] NASA Technology Transfer Program, “NASA Software 2015-2016 Catalog”(PDF):
software.nasa.gov/NASA_Software_Catalog_2015-16.pdf
[5] My writeup of the program on the @ScienceVsUniverse group on Facebook:
facebook.com/groups/sciencevsuniverse/permalink/939080342894377/
said information may be available. KLabs has the schematic pack for the Apollo Block II guidance computer (accessible from the UK as of March 2021). Voyager used the same basic computer design as the Viking Mars probes, so you might have luck finding schematics through a search for that. JPL has, in its paper archive, about 15 cubic feet of pretty much every document ever printed concerning the Voyager program pre-launch. A lot of that may still be compartmented under 22 USC chapter 39 due to the fact that launcher hardware is still largely based on military hardware development.
This popped up on my suggested videos list. Very appropriate in view of the latest heliopause measurements by both Voyagers.p
Great presentation on how the whole planning in design and building the spacecraft. I would like to learn more about this. Thank you!
Brilliant well executed talk. I learned a lot.
Thank you for a fascinating talk
thank you for this detailled explaination. as a youth i did lots of machine language programming on the processor of the Commodore64 and actually lots of interrupt-controlled programming. So such old hardware still draws my attention and i still find the design of the machine(s) brilliant.
I wonder how many of the original engineers are still on the project. Mostly retired by now, if even alive.
None.
www.geek.com/news/nasa-seeks-programmer-fluent-in-60-year-old-languages-to-work-on-voyager-1638276/
Is pretty nice infomation about the Voyager's Computer, i also laughed at the added humor of Cassette, 8-Track Tape Format.
800x800 is NOT 640 mega pixels, it is actually .640 mega pixels. I was actually interested in generally how electronics could run that long, and how much of it has been crossed to backups. I'd think some of the reaction wheels and motors may not work any more. Then there was the need to alter the transmission being so far away by at least lowering the baud.
Yeah, he didn't get into the camera mount workarounds they had to use because the tracking platform for the camera on Voyager 2 jammed after visiting Saturn, so they had to rotate the entire spacecraft to keep the camera pointed in the right direction for the multiple-minute exposure times
en.wikipedia.org/wiki/Voyager_2#Encounter_with_Saturn
though they eventually got the camera platform working again. He could easily do another whole talk on the various work-arounds and remote fixes they've done to keep the spacecraft running.
I just loved your presentation on the voyager 1 & 2 probes. I do have a question. how will it react to the cosmic radiation as it leaves the heliopause? wil we still be able to downlink and upload commands?
I would have asked about the mission plans for the earlier (scrapped) mission
Awesome stuff, and amazing talk!
I need to track down what font this slideshow was made in. It's very space-race fitting.
omg its full of stars
Always has been
Thank you for a fascinating lecture. Does anyone know where one can find the schematic diagrams of the Voyager radio transmitter / receiver? I have seen a block diagram of the radio system but no a schematic down to component level. I would be really interested to see this. Hope someone can help.
Some are using Sterling Engine driven generators instead of thermocouples in the RTGs.
500 megabits is about 62.5 megabytes storage on 8 partitioned tape.
I would love some documentation on the actual hardware (especially "the processor" I guess it was more a discrete logic solution, its opcodes etc ... but it's possible that I'm not aware of having this information somewhere already which is available to the "public"). Maybe enough to create an emulator at least for fun. Nos so much sane goal btw, I just like code emulators ;)
General Electric 18-bit TTL CPU.
NASA has released a trove of mission documents related to Apollo and later missions, so there's likely to be quite a bit of stuff available through NASA, especially through their Technology Transfer Program if you're a US citizen and resident (they have a more limited set of documents and software available worldwide).
I was able to receive all sorts of amazing Apollo-era PDFs for the design of various subsystems back in March 2017. You might have similar luck regarding Voyager designs if you're willing to apply to the program and do a bit of digging through their massive catalog.
The links below are a bit stale, corresponding to the third release of the catalog for 2017-2018. They release the catalog every two years, so there should be a corresponding fourth dump for 2019-2020.
_________________________________________________________________________
SOURCES
[1] NASA Technology Transfer Program channel on CZcams, “Download NASA Software for Free!”, 3/1/2017:
czcams.com/video/JIpyc8AfMZY/video.html
[2] NASA Technology Transfer Program website (start here for application and download):
software.nasa.gov
[3] NASA Technology Transfer Program, “NASA Software 2017-2018 Catalog”(PDF):
technology.nasa.gov/NASA_Software_Catalog_2017-18.pdf
[4] NASA Technology Transfer Program, “NASA Software 2015-2016 Catalog”(PDF):
software.nasa.gov/NASA_Software_Catalog_2015-16.pdf
[5] My writeup of the program on the @ScienceVsUniverse group on Facebook:
facebook.com/groups/sciencevsuniverse/permalink/939080342894377/
The data format is the same as was used on the Apollo flight control computers. I am guessing a strong relationship.
How does JPL NASA train new engineers about programming for Voyager. It has been now two generations of people since Voyager launched.
At 37:35. Wikipedia says the Voyager RTGs use bi-metallic thermocouples (not semiconductor).
See: en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator#Life_span
@@gordonrichardson2972 That makes sense, the radiation from the plutonium would tend to p-type dope silicon by knocking holes in the crystal lattice. It would probably degrade faster than a thermocouple. Maybe the concept would work better with a wide bandgap semiconductor, not available in the 1970s.
A question at the end was asked about an alternative to RTG's, well in the near future there may be one. Using a stirling engine instead of thermocouples would be more efficient, but of course has issues with wear and working fluid leakage. Nevertheless it is being worked on:
en.wikipedia.org/wiki/Stirling_radioisotope_generator
time for new probes, we can probably catch up to the old voyagers and reach much further?
We did. And have, still, probes around Jupiter/Saturn, etc right now... Crack open yer Googler...
I was wondering myself if New Horizons has the potential to overtake Voyager in terms of distance from the sun, but it turns out that NH is travelling slower than either Voyager and that the difference in speed is widening. So any chance at catching either V'ger in terms of distance from the sun would be with some new mission yet to be launched.
Does NASA or JPL publish incoming data from either Voyager as it comes in? Moreover; what does that data look like? Can we use web browsers to view the actual data, even if it's incomprehensible to amateurs?
you can check here: nssdc.gsfc.nasa.gov/nmc/spacecraft/displayDataset.action?spacecraftId=1977-084A
just freaking impressive how they managed to do all these things with the limited capability's of the hardware from that era. now look at technology. one thing acts up and it will bring a whole system down.
Yeh definitely impressive! Not the case at all for spacecraft nowadays though - systems are just as redundant as they've always been, depending on the mission requirements of course
I don't know why you're antagonizing domestic and space technologies. They have different design goals. If that thing was designed to not act up, or the system was designed to be resilient, its cost would be 10-fold if not 100-fold. You would simply not buy it.
Actually, there's just the same attention to detail today in military, aviation and space systems today; hardening and redundancy of various types depdending on the engineering constraints which is what Aaron Cummings is focusing on.
I've been working a lot with FPGAs, which can be thought of as the 'next level' beyond GPUs for high-performance computing. When you're planning to code an algorithm into FPGA hardware that is designed to run for perhaps weeks or months at a time, if not years, you have to take things into account like cosmic rays coming along and performing random bit flips in any register or memory location. So redundancy and recovering from errors is key to detecting such issues, complicating the design but making the results far more reliable reliable and statistically predictable.
I also started working with digital electronics and computers professionally beginning in 1977, first in NYC as a high school teen and then at MIT working for Dr. George Ricker who's the chief scientist on NASA's TESS mission which just wound up (and replaced the Kepler mission most folks have heard about). Coincidentally, that was the year the Voyager 1 and 2 missions were launched.
At MIT when I worked for Dr. Ricker, circa 1980-1981, we were making the same transition from mid-1970s computers based on TTL processing boards and core memory (we had 32Kx16-bits in our Data General Nova 2 hand-me-down from another, better funded, lab). From the presentation, the Nova 2 looks very similar to the RCS and AACS computers except for the 4-bit nibbling of the fetch/store to memory; the Nova 2 would fetch/store 16-bits at a time bit with 4x the amount of electronics. But the Nova 2 was insufficient for the huge 128x128 images [ !! ] we were getting from the first Texas Instruments CCD devices we were mounting on telescopes for astronomical measurements in the visible light and X-ray spectrums. So we hand-built a custom 256Kx16-bit frame buffer using S-100 boards with CMOS memories on them, and for a while we gained some bragging rights for having the most memory in any single system across all of the labs in MIT's Center for Space Research (which I should mention is where the Apollo-era computers were designed and built a decade earlier). That frame buffer, though it lacked a CPU and was controlled by the Nova 2 as an IO device, was roughly comparable to the FDS imaging computer on Voyager, which may have had up to perhaps 512Kx18 of CMOS memory to hold a single 640x640 image, perhaps more (though that would make for more power requirements and circuitry).
It's true this is nothing compared to what is possible today; TESS has gigabytes of memory on board (I think perhaps 10GB), but then having such tight memory constraints leads to good memory-efficient coding practices, error recovery, and fault-tolerant designs.
If I understand, the plutonium core produces not only heat, but a large amount of radiation. Depending on the kind (and intnsity), the radiation damages the crystal structure of any materials, including metals. The damage is mentioned with a term "dislocation". That sooner or later causes enough of migration from one to the other metal which form the thermocouple. That then means local shorting or leakage and reduction of the thermocouple output even at the same temperature difference. I hope this simplified description is not too much off the real thing.
I wonder... in the future it will likely be possible and even cheap to send vehicles out to grab these two titans of the space age and return them to earth for the hero's welcome they deserve. Should we do that if we could? Or do we let them go on their way as intended? It sure would be magical to stand in a room with one and consider it's achievements. (We could even send on replacement craft on the same speed and trajectory so their role in history continues regardless.)
Is there any possibility to send a "chasing" spacecraft to intercept and amplify the signal from the Voyagers and is there any point in that?
This is about capturing and returning, but similar problems should arise:
what-if.xkcd.com/38/
The relay craft would need to be fast enough to close to a reasonable distance in a reasonable time and then slow down again, so that you dont overtake Voyager, as that wouldn't do you any good . . .
And it would need sufficient communications gear and power supply. I have a hard time seeing a 34m oder 70m dish mounted on a spacecraft (that's the size of what the Deep Space Network uses).
And then the question is for how much longer you' get usefull data from the Voyagers, even if you manage to pull all that off.
My guess would be that a purpose built interstelar exploration spacecraft would be more usefull. You could use the results from Voyager to select relevant instuments (i.e. a camera probably wouldn't be of all that much use) and (maybe?) put it on a faster trajectory, as you wouldn't need to visit any planets, except for gravity assists.
I don't know much how NASA implemented computer system on Voyager, but an uptime from embedded system is something totally different from flipping a switch on a workstation, boot it up and then just hope there will be no need to upgrade, there will be no crashes and the electrical company will not shut the grid off for maintenance for 15.000 days. Those systems 'reset' constantly, it is more like an array of microcontrollers like Arduinos than an advanced PC
Could someone explain how Voyager send the data back to Earth, even though the distance is big?
Basic radio data communications - if you have a big enough dish pointing in the right direction you can receive data. Doubling the distance reduces the power received by a factor of four. The transmitter power is slowly declining but the receiver sensitivities have improved by several orders of magnitude since the probe was launched (and we can replace the receivers).
ASA should have used 'plug in Plutonium-238 fuel sources', so a high speed refueling robot could be sent to repower the two probes.
Stirling engines can last decades for power generation. NASA is considering these for future probes.
I wonder if a moving parts inside an engine would knock the satellite off course ?
Can ground signals still be heard by Voyager? If so, when will Voyager no longer be able to hear from Earth?
The Deep Space Network (DSN) ground stations have been continuously upgraded over the decades with larger and more powerful components. The communication failure will ultimately be the downlink from Voyager (whisper), not the uplink from the DSN (shouting).
👈👽we found the disk, and we are coming to earth.
If there is no security, then someone better start figuring out how to get it to play Doom.
About the V-s communication system czcams.com/video/FzRP1qdwPKw/video.html