Hacker News Reader

Comment by whartung | Hacker News Reader

This is one of my favorite reasonably recent Voyager anecdote, from a recent special about the program.

Paraphrasing, the person said "You carry more computing power in your pocket than what we have on Voyager. And by that, I don't mean your smart phone. I mean you car key fob."

View on HN

The one I often think about is: Your USB-C charger is more powerful than the Apollo 11's computer.

This blew my 12 year old mind when someone said this about that Furby toy back in the late 90s.

... and yet its travelled 12 billion miles from Earth

The power of persistence!

and radioisotope thermoelectric generators...

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

Touché!

Could you power a house with a few dozen of these RTGs?

Or are they just so obscenely expensive?

If you could use these then you wouldn't really need batteries, which you do need if you go solar.

Is it even possible for a mortal to buy a RTG?

You can steal some from russian lighthouses....on your own risk.

Generally it'd be more efficient to build a nuclear power plant with the materials from the RTG, and power many houses.

The voyager probes had 450 watt of total electrical power at launch, supplied by 3 RTGs. The plutonium alone in one of those RTG units (4.5kg) would easily cost $10M.

And it's the type of thing that is only used by government programs, very dangerous to handle, and which requires a whole specialized industry to produce. The accounting gets complicated. I wouldn't be surprised by a $100M cost per RTG unit.

The Cassini probe used about 60 pounds of plutonium-238 for its RTGs, at a cost of around $200 million for just the plutonium.

Computing power is nothing without a high Delta-V rocket under your backside.

I remember reading that the SIM card in your phone was equivalent [or superior] to a Commodore 64..

Your SIM card is running full blown Java applications in a “secure enclave” like chip. Good luck getting a 6502 system to do that.

Interesting related tidbit - Most sophisticated computers would instantly fail when exposed to outerspace. The process of making them rad tolerant is expensive and complicated. Using simpler chips makes it easier and means there is less that can go wrong. This is why most space bound computers even today appear on the surface spec sheet to be out of date.

That never made sense to me until I really learned about bit flips, and how the smaller the process node, the more susceptible it is.

They are at least up to PowerPC with the most recent Mars missions, probably not a terrible tradeoff of computing power vs. radiation vulnerability.

https://www.macrumors.com/2021/03/02/nasa-mars-perseverance-...

The Mars Helicopter Ingenuity actually has a 2014-era 2.2GHz ARM phone SoC, the Snapdragon 801 (same chip as the Galaxy S5 and OnePlus One)

https://rotorcraft.arc.nasa.gov/Publications/files/Balaram_A...

Wow that's crazy, since I'd imagine SoCs to be the most vulnerable to bitflips.

Here are details on the Voyager computers:

https://www.allaboutcircuits.com/news/voyager-mission-annive...

That was popular understanding until Spacex used normal intel chips for most of their flight hardware. They have the computers generally in triplicate do a bit flip in one doesn’t cause issues, and it is many times more cost effective.

That’s ... probably correct on the facts, but won’t a key fob be using something along the line of an ATtiny or an STM8S0? Those have pretty decent computing power as far as historical microcomputers go (about a TRS-80 or an Apple ][ or a somewhat downscaled AGC) and may even be clocked at 12 or 16 MHz, yet routinely have absolutely pitiful amounts of RAM (to be distinguished from program ROM), down to 256 bytes in some cases. An 8-bitter with 2 or 4K of memory and one with 64K of memory afford very different programming styles and environmental conveniences.

And that exact anecdote is contained in the article. > In fact, the smartphone comparison is not quite right. “The Voyager computers have less memory than the key fob that opens your car door,” Spilker says.

A better way to phrase this would be "Look what they managed with only 70 kilobytes of memory. Now what have you done with all the terabytes you have available to you?"

I've got to say that I am very doubtful of that. A key fob just generates one time passcodes, and offers the ability to set the seed to be used. I've implemented this sort of algorithm on SIMs (to be used for VPN authentication, not mobile authentication). You don't need 69kB for that. Even 1kB would be luxurious. So why would a car key fob have more than the 69kB on Voyager computers?

Yeah, probably a bit of hyperbole :) I'm no auto guy, but this seems to be an example of a microcontroller used for key fobs: https://www.microchip.com/en-us/product/ATA5700 - 8-bit AVR core - 32 KB flash - 2 KB EEPROM

The datasheet mentions SRAM, but no indication how much, and no indication of clock speed AFAICT.

But which one is better ? Voyager is on since 1970, your car key since max 10 years.

Yeah, but my key doesn't have a nuclear generator strapped to it.

Pfft, I thought this was a forum for hackers.

We had remote key fobs in the nineties.

I think their comment was regarding battery life, not year of invention. A 90s key fob would be dead by now without a battery change even without use.

Reply on news.ycombinator.com