Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

As always with radio, it's mostly about unobstructed line of sight and the gain in your antenna system.

We're still in communication with Voyager 1, which is operating on a grand total of about 20W of RF power; and is currently about 14.5 billion miles away.



The "path loss equation" goes something like:

At the receiver, you have "minimum detectable signal", MDS, measured in dBm.

At the transmitter, you have power out, measured in dBm. Add transmitting antenna gain, in dB, subtract propagation loss through medium(s), add receiving antenna gain, and if that number is greater than MDS, you win! The Really Great Science in Voyager is the added factor of "coding gain" -- sophisticated error correction codes can give you a many dB adder, at the expense of data rate (nobody cheats Claude Shannon).


And gets up to a whopping 1.4 kbps downlink and 16 bps uplink!

https://voyager.jpl.nasa.gov/mission/spacecraft/


that is amazing for something 14.5 billion miles away


It's also amazing considering it's using a computer built over 45 years ago from discrete components with only 70Kb of memory, while operating on a gradually failing thermonuclear power source. Voyager 1 also holds the world record for the longest continual operation of a computer:

https://www.guinnessworldrecords.com/world-records/635980-lo...


Wow. World record for the longest continual operation of a computer... achieved in a radiation bath.


Pretty much, although I think the RTG is mounted on a limb as far away from the other equipment as possible, but still the computer needs to be able to tolerate a decent amount of radiation. It also has an interesting redundant design where it has double the component count and can either exploit that for extra compute or redundancy in case of component failure.


I think parent comment is talking about cosmic radiation and not the one from RTG which most probably is properly shielded.


Possibly. I wonder which is the more significant source given it's distance from the Sun.


Solar particles become less of a threat, but the sun's atmosphere actually shields the solar system from some types of cosmic rays. Don't know what proportion of electronics faults stem from each, though.


It used a memory called "plated wire memory" which, like the very related core memory is very resistant to changes caused by radiation.

https://en.wikipedia.org/wiki/Plated_wire_memory


> ...the last of the project's original programmers, retired, and it was difficult to find a replacement with such in-depth knowledge of what now seem like ancient hardware and design principles

Man I would not want to be the one to brick that.


That makes me realize, it would actually be an incredibly fun job to be in the team that manages the simulation/emulation stack that people tinker with to learn about this thing.

Going on a lot of presumption here (and a bit of optimism...), I would presume "don't brick it" is backed up by an appropriate level of funding to cover the training people receive before they get to touch production. In an optimal scenario (don't be wrong... don't be wrong...) this would cover component-level electrically-accurate (because radiation) logic simulation that everything gets tested on first.


NASA doesn't mess around with simulators. I've read accounts from astronauts and ground controllers where they say their actual missions feel anticlimactic after all the insane things they had to do in the simulators. Gene Kranz managed to crash a simulated Apollo mission into a mountain.


The code on the computers has been upgraded, and even the language used to prpgram has changed over the years [1]. Which leads me to believe that the computer has been restarted/rebooted many times in all these years.

I think there may be other computers, right here on Earth, that may be able to at least compete given those criteria.

[1]: From the linked page:

> The CCS originally ran software written in the Fortran programming language, but this has been continually upgraded since its launch (software updates can be transmitted and installed remotely). The current software is written in a mixture of C and Fortran.

The following is fascinating.

> The age of the computer and its codebase has caused problems for NASA in recent years. In 2015 Larry Zottarelli, the last of the project's original programmers, retired, and it was difficult to find a replacement with such in-depth knowledge of what now seem like ancient hardware and design principles.


What an incredible project. Does anyone know of any good interviews or written content from the engineers?


There are tons of documentaries and interviews. [1]

[2] https://www.pbs.org/the-farthest/


What's more amazing is how NASA's Deep Space Network continues to make communication with Voyager possible.

https://www.nasa.gov/feature/jpl/nasa-contacts-voyager-2-usi...


I wonder what the ping/latency is...


~23.2B km / c (m/s) seconds, so about 43 hours or roughly the time to get through to an arbitrary customer service representative these days.


ping...

Hello. You've reached the Voyager I spacecraft. We can't answer your ping right now but if you'd like to leave a message, please, do so after the beep. We look forward to servicing your request as soon as possible.


(Current) One-Way Light Time = 21:35:13 [0]

So 43 hours 10 minutes 26 seconds minimum, excluding processing time.

That's probably another record... Although I'm pretty sure it can't process ICMPs, so "ping" in the more general sense.

[0] https://voyager.jpl.nasa.gov/mission/status/


That must be fun over SSH/Telnet. 43 hours between every key press

Joking, but thanks for the interesting info (and Frost1x also).


Come on, SSH and Telnet? That's ludicrous. It's a job for https://mosh.org/


At the very least, local echo must be enabled.


Neither ICMP nor IPv4 existed when the Voyager probes were launched in 1977 :) ICMP was created in 1981


Yeah, that's what I was implying ;D

I think an older probe also took the first digital pictures before bitmap images were a thing, they could digitise it at the probe end but had to colour in the pixels on a piece of paper at the other end if I recall correctly. Point is, NASA had to solve a lot of digital communication problems long before they had established terrestrial solutions or before computers were even capable of solving the whole problem on their own.

Then again, even if the timing was reversed, a lot of terrestrial communication protocols have no chance of working properly at the ranges required for space exploration - something I've been meaning to look into is the redundancy requirements over those long distances, since re-transmission is so costly in terms of latency that it probably makes sense to pack as much redundancy into the signal as possible, we already do this with terrestrial communication in the form of various codes, but there's a finer balance between increased bandwidth and the relatively low latency cost of re-transmission in the case of unrecoverable sequences... I'm guessing you want a far lower probability of loss for a 43 hour round trip.


And I used to get angry with 250ms ping playing WoW.


Which somehow is a dream on a few networks.


Reminds me of the early days of dial-up Internet... ;-D


The first modems were only 110 bps.


You mean bauds


It's both. 110, 300 and early 1200 were all one bit per symbol.


As always with radio, it's mostly about unobstructed line of sight and the gain in your antenna system.

Indeed. If you grow up with your most common radio interactions being an FM car radio and a dumbphone, you get the impression it's entirely about range. Then you buy a drone and find out one pine needle shaves 50% off of your signal strength.


the mavic 2 series of drones by DJI don't have this issue, but the tradeoff is the controller weighs like 40 pounds, because it's a quite powerful wireless access point. Every other drone i've owned that supported video used an android or ios device to connect to the drone via wifi (drone presents as a WAP).

iirc the claimed range is around 8km on the one i have, about 5 miles. I have assuredly gone well over 2km with no issues with control or video feed. This was over a straight highway. I routinely fly around a kilometer away, and the only issues i have is if i launch from an extremely dense patch of pine trees, and only at about 800-900 meters, i will lose video (artifacting for a second), but not control. It's never had to RTH.

In case you're curious about city usage, i have a friend that has one he launches from a culdesac in Orange County and can fly in nearly any direction for about 8 minutes* before he hits a geofence, the drone still functions normally. If there is any issues, he can just fly higher.

The newest newest DJI stuff claims even more ridiculous range, 15km+ over open water, for instance.

If i hadn't used it myself, i wouldn't have believed it, it sounds like BS.

* this is 5-7km depending on the tailwind


but the tradeoff is the controller weighs like 40 pounds, because it's a quite powerful wireless access point. Every other drone i've owned that supported video used an android or ios device to connect to the drone via wifi (drone presents as a WAP).

That is not accurate. The modified wi-fi some DJI drones use is not reliant on the smartphone/tablet attached to the controller. It's strictly between the drone and the controller, which passes data on to the phone via USB. The drones can be switched into AP mode for faster media downloads, but at that point they lose connection with the controller. Ocusync controllers weigh 390 grams, and that isn't that much considering their build quality and the fact they have two 18650 cells inside.


40 pounds?!?! Really?! That's 18kg.

That's really awesome though.


about 650 grams, actually. I went and weighed it, just now. the controller itself is 390g, plus the phone.


Thanks for sharing those figures. I've always known "it's amazing" and "it's far away", but those numbers really put things into perspective.

Is there anything particularly special about the antennae on the spaceship? They must be rigorously aligned to point at Earth, and even a slight knock would spoil everything? Or is it more resilient than that?


There is a beamwidth associated with a parabolic antenna's size and operating frequency. The attitude of a spacecraft is periodically corrected to keep earth within the beam's central lobe. It's approximate so corrections aren't required too frequently and become less of an issue as Voyager gets further away from earth's orbit around the sun. Once it's out of propellant it won't be able to correct for torque from the solar wind etc and the high gain antenna's central lobe will drift until its no longer pointing close enough to maintain an acceptable signal to noise ratio. Unless they can maintain contact through the omnidirectional antenna, Voyager will be effectively lost.


I would think the biggest part of being able to transmit that far is the perfect vacuum of space. There is almost nothing, not even air particles between us and the probe.

If you threw a beach ball from the distance of voyager straight to earth it would eventually make it here.


Vacuum is good, but inverse-square law still applies.

There is almost perfect vacuum between Earth and every star in the galaxy, and yet they don't appear nearly as bright as the Sun.


But you can see the Andromeda galaxy, which is several quintillion miles away.


Sure, but the andromeda galaxy doesn’t have a power budget of 20 watts.


Some photons would make it, just not frequently enough to ascertain any information.


Well, orbital mechanics would mess up a straight toss but at least it wouldn't be stopped by air resistance.


Jupiter might have other plans once your beach ball gets close enough.


So by the articles calculations voyager 1 is a steal.

At $865 million[1] and 14.5B miles that about 0.034 Euros/Km. 1/20 of what he did in 2005.

1: https://voyager.jpl.nasa.gov/frequently-asked-questions/fact...


For some two-way wireless protocols (like wifi) you have to take into account the guard interval, slot times and interframe spacing which are all values set in time (~1-50us). For long distance transmissions your speed-of-light limited signal propagation time can exceed these values.

In terms of size usually guard interval < slot size < inter-frame space. If propagation exceeds guard interval AND have a channel with lots of echo any communication will be difficult. If propagation exceeds slot timing then coordination between more than 2 devices will be different (high retries/low throughput). If propagation exceeds interframe spacing a two-way wifi connection will not be possible as both stations will think every frame timed out waiting for an ACK.

More info here: https://en.wikipedia.org/wiki/Guard_interval https://en.wikipedia.org/wiki/Distributed_coordination_funct... https://en.wikipedia.org/wiki/Short_Interframe_Space


It's also handy when you only need a low data rate and can make your channel bandwidth was wide as you like without worrying about licensing restrictions.


That's...not how that works. The FCC requires power be dropped to compensate for antenna gain.

https://afar.net/tutorials/fcc-rules/


For a space probe? I'm saying to get that performance NASA has some operating freedoms the rest of us don't.


Besides, I didn't mention power level. Yes, effective radiated power is regulated.


Would this mean it's possible to communicate with someone orbiting another star? Or do the sheer distances become too great?


doesn't atmosphere (or the lack thereof) plays a role in power losses too ?


Yes, radio waves are attenuated in the atmosphere. This is highly frequency dependent - for practical applications the lower the frequency, the less radio waves are attenuated. In comparison with attenuation from obstacles in non-line-of-sight situations, the atmospheric component is not significant.

For really long range propagation on earth, reflections on atmospheric layers are the dominant factor (as there is no line of sights due to the curvature of the planetary surface).

See https://www.sciencedirect.com/topics/engineering/atmospheric... for some nice graphs.


Does that mean that there is an electromagnetic frequency that, no matter how much power we can feasibly put into it, it simply will not transmit through earth's atmosphere due to its' instantaneous attenuation outside of a vacuum?


Yes, in the terahertz band range for example you have very high attenuation of more than 1000 db/km. It's a spectrum (no pun intended), so depending on where you draw the line on "through earth's atmosphere" in terms of how far you want to transmit, you'll have some frequency where this becomes infeasible or you'll need more directed antennas and/or more transmit power.


Atmosphere does play a part but just freespace losses are going to be massive, probably at least 250dB just for the distance voyager is at. Atmosphere could add another 60-100dB. Voyager's antenna has about 40dB of gain but the DSN network can array multiple antennas to make up for the losses. They have up to 2 70m dishes and several 34m dishes that can point at Voyager, quite a massive antenna gain.


it does, but it is not as significant in normal weather conditions in the frequency ranges we're dealing with here (Likely sub 2.4 Ghz) [1], (General rule of thumb, Atmosphere absorbs everything except visible spectrum and 10 cm - 10 m wavelength), compare this to effect from the inverse square law. :)

Now this might be significant enough in directional waves with a huge constant multiplier (like a 'ideal' laser with no divergence). Someone can probably give insight on it here.

1. https://www.researchgate.net/publication/270512069_Propagati...


I'm curious why you ordered the wavelengths the way you did, you totally potholed my entire comprehension. 10 meters (28mhz) through 10cm (2800mhz) to me reads "left to right".


Oh I just wrote it like that cause that's what I remembered off the top of my head. For some reason my brain prefers using wavelengths for Radio waves (short wave radio, long wave radio) while frequencies when looking at the whole spectrum.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: