The scientific objectives and payloads of Chang’E4 mission [pdf] (opens in new tab)

(epizodsspace.airbase.ru)

32 pointsreindeerer7y ago9 comments

9 comments

6 comments · 1 top-level

codezero7y ago· 5 in thread

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.

reindeererOP7y ago

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

codezero7y ago

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.

1 more reply

ficklepickle7y ago

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)

1 more reply

technion7y ago

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.

reindeererOP7y ago

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.

j / k navigate · click thread line to collapse

9 comments

6 comments · 1 top-level

codezero7y ago· 5 in thread

reindeererOP7y ago

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

codezero7y ago

1 more reply

ficklepickle7y ago

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)

1 more reply

technion7y ago

reindeererOP7y ago

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.

j / k navigate · click thread line to collapse