[X] My Assistant

[mike]

To be fair, it seems that Japan doesn't allocate much money to their space program, and it's better that they tried at all than not. Surely they learned something about implementing basic failsafes..

And it's not like the USA and USSR haven't had their even more spectacular failures.

[Guest_BruceMoomaw_*]

Yes -- but, again, the problem is that they consistently try to do too much with their limited funding, as even their official reviews have stated. Hayabusa, for instance, is a much more complex mission than NEAR -- being done with about half the money. The same thing has happened over and over again with very complex Japanese Earth satellites that they've tried to build on a shoestring.

[tedstryk]

This failure really does smell like the MPL/MCO failures. However, Hayabusa is still going to attempt sampling, and has already returned spectacular coverage of the asteroid. Calling it a failure at this point would be a huge stretch.

[hugh]

I don’t understand the attitude in some of the posts here since the apparent loss of Minerva, which was a minor element of the mission anyway- if anything far less important than Beagle-2 was to Mars Express. Considering the spectacular images received, the mission is already a qualified success, whatever happens from now on. And if these results were achieved on a budget half that of NEAR, I would have thought that ought to earn some respect.

[djellison]

Hayabusa, for instance, is a much more complex mission than NEAR -- being done with about half the money.

And they've achieved much of what NEAR achieved on much less money. That deserves a lot of credit.

Doug

[tedstryk]

I think this has to do with the Bruce-All-Mighty's attitude toward the Japanese space program.

[Joffan]

And just think - whatever happens on sampling now can be good news. If the sampling goes ahead successfully, that is terrific; and if it doesn't, Hayabusa can stay next to Itokawa and do an astounding mapping job on it, practice lots of station keeping and close approaches etc. Obviously there's no need to return to Earth. Maybe it can even transfer to another asteroid ....

[odave]

There are valid points on both sides of the discussion.

From Oberg's article, it sounds like they were trying to "time it":

Kyodo quoted Kawaguchi as saying the mission team tried to make sure that the deployment signal would arrive while Hayabusa was descending. For some reason, this did not work.

Just relying on timing for things to happen is bad-bad-bad when doing real-time process control. They certainly blew it on the Minerva release software not having some kind of interlock (if indeed they didn't have one - I don't think we know that for sure - but if they did, it was buggy). They probably short-changed the software development process, either in design reviews or testing. That is consistent with my experience with similarly constrained software development efforts, (Japanese or otherwise), which tends to be "Just In Time", or rather, "Just Too Late".

But, the loss of a little technology test probe pales in comparison to the loss of MCO due to the english vs metric unit thing. In my opinion, that blunder was embarassing and unforgivable. I'd give the guys responsible for that one a lot more time in the hairshirt than the JAXA guys who forgot to put in the software interlock.

Restating what others have already said, Hayabusa has returned a lot science without Minerva. Even if the sample collection & return fails, it will still beat MCO, which never had a chance to return anything.

Yes, JAXA has had plenty of opportunities to learn from failures in the past, and they should take a hard look at what went wrong at all points of this mission. They should also be congratulated on what they have accomplished so far.


Edit: Reworded the end of my second paragraph, as the former version was unfair to Japanese software developers in general.

[chris]

I agree that trying to release the probe by pure timing alone was a mistake. Is it possible that this would have been a valid strategy when all the reaction wheels were working? The loss of Minerva may reflect the fact that they are now struggling to get the fine control they really need.

One place that a limited budget might really hurt them is in their ability to respond properly when the reaction wheels died.

Chris

[RNeuhaus]

I am afraid that the timing for firing and sucking sand samples from Itokawa is of the same software logic as for releasing Minerva mini-probe (not interlook timing software between the altitude and velocity direction).

Rodolfo

[odave]

At the risk of beating a dead micro rover:

Note that Minerva is not even on Hayabusa's mission success criteria (scroll 3/4 the way down the page).

That's not an excuse for sloppy software and losing Minerva however. But still, at this point I'd say they're pretty close to achieving more than half of their goals.

[djellison]

I think this has to do with the Bruce-All-Mighty's attitude toward the Japanese space program.

He has phases, 6 months ago it was the Discovery program on the recieving end - despite giving us Deep Impact, NEAR, Pathfinder, Genesis, Stardust, Lunar Prospector and Messenger, and only one total failure, Contour.

I'll probably be ESA next, not sure what the angle will be though

Doug

[The Messenger]

I agree that trying to release the probe by pure timing alone was a mistake. Is it possible that this would have been a valid strategy when all the reaction wheels were working? The loss of Minerva may reflect the fact that they are now struggling to get the fine control they really need.

One place that a limited budget might really hurt them is in their ability to respond properly when the reaction wheels died.

Chris

Yes - there are really at least three issues here, maybe more: 1) Software that did not properly address a possible error mode (bad timing). 2) What caused the timing problem in the first place? 3) Reaction wheel fritz outs(?).

Reaction wheels have historically been a bug-a-boo. Roughly a third of the Galileo planetary science runs were compromised by reaction wheel problems. Cassini was designed with double-redundant reaction wheel control systems, with an extra wheel thrown in to boot. Even so, there have been ~2 incidents a year where the reaction wheels have gone south due to breaker trips - thought to be caused by cosmic rays. There has been little mention of this during the Cassini mission - most of what I just stated, I did not know until this week's Cassini Event log was posted.

My question is: If NASA would have been more open about the puxxling reaction wheel problems plaguing missions, would JAXA have been more conservative in their own reaction wheel circuit designs?

Which brings me back to issue two: What caused the timing problem? The rough surface, a bad control feedback system (PID), or a bad estimate of the mass of Itokawa?

Bruce just pointed out to us Psyche is made out of nickle, but the density is calculated at an unbelievable value of ~2g/cc - Likewise, Deep Impact estimates yield an unrealistically low density for Temple 1. There is too much of this going around.

Edited to add: Plowing through the Cassini Event logs, there is mention of previous events - especially during the encounter with Jupiter. Likewise, the Galileo reaction wheel events appeared to occur during closest approach of Jupiters moons. So cosmic rays? It would seem to me, that this 'usual suspect' makes its rounds too close to close encounters.

Recent articles indicate gamma rays actually originate in thunderstorms via an unknown mechanism - that the gamma rays known to be associated with lightning are not triggered by known cosmic events. Is there an triboelectric loading during closest-approach of a body that increases the probability, or intensity of cosmic ray discharge/encounters?

[tedstryk]

Well, most, if not all, NASA missions with reaction wheels had backups. Hayabusa doesn't. I think the lack of redundancy is the problem.

[ilbasso]

[Tongue In Cheek Mode]
I think the problem is obvious. Just try to make sense of those dreadful Google translations of the Japanese blogs and press conferences. Now imagine the poor little computer in Hayabusa trying to comprehend and translate the Japanese commands it is being sent. No wonder it is confused! EVERYONE knows that computers speak English, right?
[/Tongue In Cheek Mode]