The scientific objectives and payloads of Chang’E4 mission [pdf] | Hacker News Reader

The scientific objectives and payloads of Chang’E4 mission [pdf]

epizodsspace.airbase.ru

32 pointsreindeerer9 comments

This is really neat. I worked on a satellite from 2006-2012 that had a RISC processor from ~1993 if I remember correctly. Radiation hardened processors trail behind the modern tech pretty far, but luckily, most spacecraft aren't doing too much processing on board.

What i found intriguing is that Xilinx Q-Pro family is supposedly ITAR-controlled, so how are Chinese running those on the far side of the moon is .. an interesting question.

Further, googling "XQR2V3000" "AT697F" and together returns some interesting results from UK based research ..

Also, Chang'e actually does impressive amounts of processing on board, as they employ vision-based autonomous guidance algorithms for landing. Thats unprecedented

That's really interesting, it hadn't occurred to me that an FPGA would ever fall under ITAR, then again, after quickly scanning Wikipedia, I didn't even realize things like night vision fall under that too, but I guess it makes sense.

Yeah, a lot of devices that are powerful enough (subject to fudging) that are uncommitted but can be committed to Bad Stuff seems to be ITAR regulated. Another thing that might be the trigger is processors with real time capabilities.

I was surprised to see that the Beaglebone (with its very powerful PRU) was ITAR regulated, but the Raspberry Pi (not a lot of real time stuff) wasn't.

AM335x PRUs are exactly the reason why its under ITAR. Hard-realtime running alongside general-purpose OS

Unprecedented, except for this paper[1] from 2010 and the NASA CoNNeCT project[2]. Both of which use the same hardware.

But it is unprecedented in that nobody has landed on the far side of the moon before.

[1] https://www.academia.edu/8538870/A_novel_autonomous_low-cost...

[2] https://www.grc.nasa.gov/doc/space-flight-systems-maintenanc... (Currently down)

Unprecedented in that it has never been used in space before. Studies and projects yes. NASA is working on their own version for Mars 2020 rover, although the system ( called ALHAT ) hasn't been officially included yet

My question in these cases always ends up being how these one off, specialised designs end up not being full of bugs. I mean enterprise servers from tier 1 vendors trail years and years behind gaming desktops in what they ship, allegedly for stability and QA, and yet every new model is followed up by either BIOS or CPU microcode updates to deal with massive, showstopping bugs.

Two things: they normally have a mass market / volume counterpart that is not rad-hard.

Second, they actually are full of bugs, and take a ton of money to qualify. Qualification, among other things, means thoroughly documenting the bugs.

Reply on news.ycombinator.com