Not to be overly dismissive, but there are so many factual inaccuracies with both the modeling and information I'm honestly not sure this is useful. From the Antenna feeds, lack of sun detector, incorrect modeling of the bus, internals of the RTG being hallucinated, mislabeling of computers and components. There are a plethora of really high quality drawings and diagrams from the original design phases that could have been used here.
Based on what you listed been accurate (I don't know enough about Voyager to know one way or the other), I think it's perfectly reasonable to be dismissive in this instance.
Me too! The only other probe heading out now is New Horizons. It should still be working when it reaches interstellar space in the early 2040s, if NASA keeps funding it.
I'd be fascinated to know the exact nature of the failure. What analog processes cause a CMOS memory IC that we like to think of as being discrete to be consistently wrong at reading its stored values? I'd like to know all the way down to the interaction between the gates and traces. Did some pathway erode due to constant interchange of charge carriers? Did thermal cycles or stress cause one of the mechanical interconnects on the package to fracture and no longer conduct? If so, which one?
According to AI, one type of failure mode in those early aerospace CMOS chips was sudden parasitic current in the PNPN junction being caused by a high energy particle, shorting paths to ground melting metal. Though that sounds rare. Another option is a total ionizing dose of radiation over time, which creates electron-hole pairs in the gate oxide, leading to trapped positive charge. That makes NMOS get closer and closer to turning on when they shouldn't, and PMOS requiring more voltage to turn on. It's fascinating to me that there came a specific moment in time when the chip managed to work, and the next clock cycle it didn't. I'd love to see the I_D vs. V_GS charts showing a before and after. So I had Claude make a plausible approximation. You can too by asking your AI agent of choice by basically just pasting in my comment!
I tried to find the exact root cause. NASA said they could not determine the cause of the chip failure with certainty, so we can only speculate.
This is what I found as part of my research:
NASA: "Engineers can't determine with certainty what caused the issue. Two possibilities are that the chip could have been hit by an energetic particle from space or that it simply may have worn out after 46 years." https://science.nasa.gov/blogs/voyager/2024/04/04/engineers-...
IEEE Spectrum, citing project scientist Linda Spilker, reported that the memory readout showed "a stuck bit, fixed at the same binary value" in that 256-bit chunk: https://spectrum.ieee.org/voyager-1
EE Journal reported JPL's chief engineer naming the part as an RCA CD4061A, a 256 x 1-bit CMOS SRAM. Losing it took out 256 of the data computer's 8,192 words, which matches NASA's 'about 3%': https://www.eejournal.com/article/jpl-software-update-rescue...
This is neat, even if it's missing some things. I actually really like how Voyager is displayed. Thinking about putting something like together in the late 70's is pretty wild.
Seeing the spectrometers reminded me that nearly my entire physics degree only covered what was known up until the 50's. I always underestimate how much we understood even 75 years ago!
Nice! This appears to be vibecoded, but what was the process? Was there any original research or AI did most/all of it? Was this double checked against any authoritarive source?
Thanks! Yes, Claude did a lot of the heavy lifting. For research, I mainly used NASA's archives on these probes, and in total I gathered about 3 GB of files.
I didn't really think about how it has to adjust for earth's travel around the sun to point the dish, it's crazy that it is constantly gently adjusting every so very, very carefully. What a great tiny detail
With reportedly over 90% of code at even Anthropic written and tested by AI, what is not vibe coded in 2026? The term has lost all meaning when most applications are mainly AI generated code. Are people still writing While loops by hand?
Execution is becoming all that matters. This is a great project and congrats to the creator on putting it together, whether he did by punchcard coding or prompting.
It’s possible to use LLMs as part of your process while still having editorial or design control over the output. I usually take “vibecoded” as a pejorative to mean you’ve let LLMs handle almost all elements of a project. Using LLMs themselves does not make something vibecoded.
My projects aren't. "Execution" is what my question is about. This is a great project, but we need more information to have confidence on the output. I'm just struggling with AI right now in getting a makefile right. So AI output, no matter how good, can be wrong - sometimes very wrong.
I normally see it as pejorative or condescending these days, although I don't think that's what the commenter intended.
I read "was that vibe-coded?" in the voice one might speak to an infant in. "did you codey-wodey that all with your friend clooooood?". And it's super cringe. As you noted, nobody cares if you hand-coded things anymore. Execution counts, and this was well done.
Other missions have probably returned much more data but I will always have a soft spot for Voyager 1 & 2. Their pictures from the gas giants and their moons were so fresh and inspiring back in the day.
Together with the longer lasting Mars rovers I think they have offered some of the best returns on space "investments" we've ever seen.
Sorry for the late reply. It's not the traffic. The site needs WebGL, and Chrome on Linux often has it turned off. The site now shows a message when that happens.
To turn it on, paste chrome://settings/system in the address bar, turn on "Use graphics acceleration when available" and restart Chrome. Or try another browser.
Inside a nuclear battery: https://rshekhar.com/voyager/v1/rtg-1
The data computer's memory chips: https://rshekhar.com/voyager/v1/bay-4/memories
One bit of the command computer's memory: https://rshekhar.com/voyager/v1/bay-3/memory-1/plane-1/bit
The Golden Record's cover: https://rshekhar.com/voyager/v1/apart/golden-record
How Voyager 1 was fixed by radio: https://rshekhar.com/voyager/story/the-fix
The journey past the planets: https://rshekhar.com/voyager/story/the-voyagers
> Antenna feeds, incorrect modeling of the bus
Which parts do you think are wrong, and which drawings would you use?
> internals of the RTG being hallucinated
It follows NASA's 1980 cutaway: https://ntrs.nasa.gov/api/citations/19810001583/downloads/19.... One thing you are right about: the fuel spheres should be stacked in pairs. I'll fix that.
> lack of sun detector
fair, it's not named on the site. I'll add it.
> mislabeling of computers and components
The layout and names come from JPL's own packaging drawing: https://spdf.gsfc.nasa.gov/pub/data/voyager/vgrmag-data/raw/...
Based on what you listed been accurate (I don't know enough about Voyager to know one way or the other), I think it's perfectly reasonable to be dismissive in this instance.
I wish we had other long running deep-space exploration experiments to take their place!
Here is my little tribute to the Golden Record on both Voyagers https://goldenrecord.voyage/
Also a decoder to turn their raw audio into the images https://gourneau.github.io/golden-record-images/
According to AI, one type of failure mode in those early aerospace CMOS chips was sudden parasitic current in the PNPN junction being caused by a high energy particle, shorting paths to ground melting metal. Though that sounds rare. Another option is a total ionizing dose of radiation over time, which creates electron-hole pairs in the gate oxide, leading to trapped positive charge. That makes NMOS get closer and closer to turning on when they shouldn't, and PMOS requiring more voltage to turn on. It's fascinating to me that there came a specific moment in time when the chip managed to work, and the next clock cycle it didn't. I'd love to see the I_D vs. V_GS charts showing a before and after. So I had Claude make a plausible approximation. You can too by asking your AI agent of choice by basically just pasting in my comment!
This is what I found as part of my research:
NASA: "Engineers can't determine with certainty what caused the issue. Two possibilities are that the chip could have been hit by an energetic particle from space or that it simply may have worn out after 46 years." https://science.nasa.gov/blogs/voyager/2024/04/04/engineers-...
IEEE Spectrum, citing project scientist Linda Spilker, reported that the memory readout showed "a stuck bit, fixed at the same binary value" in that 256-bit chunk: https://spectrum.ieee.org/voyager-1
EE Journal reported JPL's chief engineer naming the part as an RCA CD4061A, a 256 x 1-bit CMOS SRAM. Losing it took out 256 of the data computer's 8,192 words, which matches NASA's 'about 3%': https://www.eejournal.com/article/jpl-software-update-rescue...
Seeing the spectrometers reminded me that nearly my entire physics degree only covered what was known up until the 50's. I always underestimate how much we understood even 75 years ago!
These were the main sources:
JPL's Voyager requirements book: https://spdf.gsfc.nasa.gov/pub/data/voyager/vgrmag-data/raw/...
JPL's report on the electronic parts and radiation: https://ntrs.nasa.gov/api/citations/19780007219/downloads/19...
Positions (JPL Horizons): https://ssd.jpl.nasa.gov/horizons/
The dishes (NASA's DSN Now): https://eyes.nasa.gov/dsn/dsn.html
Where the sources don't say, like which memory chip failed, I've taken the liberty of choosing one, with a disclaimer.
The failed bit is the fifth bit (bit 4 in 0..15 numbering) at addresses 0x1400-0x14FF (module 3). See Voyager Spacecraft and Fortran 5 - FDS Memory Problems: http://www.geonius.com/writing/other/voyager.html#FDSMemoryP...
Execution is becoming all that matters. This is a great project and congrats to the creator on putting it together, whether he did by punchcard coding or prompting.
I read "was that vibe-coded?" in the voice one might speak to an infant in. "did you codey-wodey that all with your friend clooooood?". And it's super cringe. As you noted, nobody cares if you hand-coded things anymore. Execution counts, and this was well done.
Together with the longer lasting Mars rovers I think they have offered some of the best returns on space "investments" we've ever seen.
this should be the link if you want to take apart the craft
To turn it on, paste chrome://settings/system in the address bar, turn on "Use graphics acceleration when available" and restart Chrome. Or try another browser.