Spirit Sol 41

On my way in I stop at McDonald's for a big soda -- 44 oz of Coke for 64 cents (including tax); who can resist?

The line is slow, and some guy is taking advantage of this to go from car to car, saying something to the occupants. When he gets to my car, I find out what it is.

"Can I have 80 cents?"

"No," I tell him, "I haven't got it."

He's very intense, very worked up about something, and standing very close to my rolled-down window. "Man," he says tightly, "I just want 80 cents for a cheeseburger."

"Sorry," I tell him. "I've already ordered, and I've only got enough to pay for it." That last bit is a flat-out lie, of course, but something about this guy is triggering alarms.

"You need change?" he asks me, digging in his pocket. "How much you want change for?"

I don't know how he has change for whatever I've got, without having 80 cents for a cheeseburger, but I let it go. "No thanks," I reiterate, "I've only got enough to pay for what I ordered."

He's clenching his fists over and over and talking through his teeth now. He starts ranting about what a shitty day he's had, and how he just wants 80 damn cents for a cheeseburger. Flecks of spittle start accumulating at the corners of his mouth. He's a crackhead.

I casually rest my elbow on the door lock, glance in my rear-view mirror to check that there's nobody behind me, and slip the car into reverse. Just in case. If he grabs for me, I'm going to spin the wheel to the right and floor it, which will knock him across half the parking lot.

Then something intellectually interesting happens. His rant cuts off abruptly, and his program apparently resets. "So ..." he begins again, relatively calmly, "can I have 80 cents?"

"No," I tell him.

"God DAMN it," he says. He stalks off, ranting to himself now, pacing the parking lot, going up to the drive-through windows to harass the poor minimum-wage slaves, to the bus stop, walking around the other cars. I roll up my window and watch him. Eventually, the frightened employees hand him a bag, and he goes to the bus stop to eat his cheeseburger.

This is why I prefer to work on Mars.

Yestersol's activities on Mars were not terribly dramatic, but there were some fireworks here on Earth. The plan was a "go-and-touch" sol, something we've never done before. The term means to drive, deploy the IDD, and use the IDD on a target. Previously, we've always gotten post-drive imagery before deploying the IDD, so that we'd know it's safe. The drive was a short one, which should mean that there would be very little uncertainty about our final location, and hence that the pre-drive imagery would be good enough for planning the IDD work in advance. So that was the plan.

Until the top IDD developers, Bob Bonitz and Eric Baumgartner (themselves rover drivers), got wind of it and freaked out. Touch-and-go sols -- use the IDD, then stow it and drive -- are fine, and we've done them many times. But go-and-touch is new, and -- in their view -- runs a serious risk of damaging the IDD or its instruments.[1] So they put the kibosh on the idea, but because they didn't find out about it until so late, they had to act after the sequences had already been uplinked to the rover. The only reasonable fix at that point was to uplink a real-time command to tell the rover to disallow use of the IDD until further notice. We'll clear that flag before today's sequences. Meanwhile, Bob and Eric are PISSED.

And we're going to need a new software delivery. Over on Opportunity, they've had a couple of unsuccessful sols traceable to a bug in RSVP: the currently installed version of RSVP doesn't know that the IDD will refuse to execute certain motions, so you can command them and they'll simulate fine on the ground, but the real rover will fault out. The bug has already been fixed in RSVP, and would have been installed already but for an error in the previous delivery attempt. So we need to try again.

As a result of yestersol's aborted plan, not much happened on Spirit; just a 90cm drive to a ring of rocks nicknamed "Stone Council." The big topic of debate in the downlink assessment meeting is how long to spend here -- is it worth hanging out for another two or three sols, or should we push on to Bonneville? They're especially interested in a flaky rock called "Mimi," which is just out of the rover's reach to the right. Its flaky appearance may indicate that it's a sedimentary rock, unlike most of the other rocks we see around us, which are volcanic. (I feel weirdly out of touch -- just from missing one day, I barely know what's going on. Until I get a chance to look at the telemetry later in the day, I don't know exactly where Mimi is, or what the surrounding terrain looks like, or how Stone Council fits into it.)

The main plan they're discussing is to spend today and tomorrow here, then move on. Today we'll IDD the dunes and/or soil immediately in front of us, and turn to face Mimi in the afternoon. Tomorrow -- after getting imagery showing that the drive completed successfully, of course (go back to sleep, Eric and Bob) -- we'll IDD Mimi.

The alternate plan is to investigate the dunes more thoroughly now, using the APXS and/or MB on a nearby drift overnight. This would mean spending an additional day at this site. They put it to a vote -- letting the Red Rover students in the room vote as well. The Red Rover students vote for the fast-track plan, and they're in the overwhelming majority. The losers, who wanted lengthier investigations of the dunes, are mollified by the observation that dunes are all over the place; they've been within reach about two out of every three sols, and we can expect that to continue. So we'll get a shot at them, even if it's not today.

I really should catch up on what I've missed the last couple of sols, but I make time to stick around for part of the science talks. The post-RAT PANCAM spectroscopy on Adirondack yielded some puzzling results. It's not pure olivine; the spectra they got from it don't quite match anything in the library. Analysis is in progress.

The other talk I stay for covers the short MB integrations we've been doing. The short version is that the results are naturally not as good as those for long integrations, but you get a lot of value very quickly. This is graphically demonstrated by plotting the results of a 30-minute observation against the results from a 24-hour observation. The 24-hour version is much cleaner, but even the 30-minute observation has the outline of the spectrum clear -- more noise, but the same big spikes. The MB, they say, gives the mineralogical context for the MI and APXS observations and will help to map significant changes in mineral composition of the soil along the traverse. So we'll probably be doing a lot of these short MB touches as we proceed.

I don't have much time before the SOWG, and I spend most of it trying to get up to speed. I don't really feel I've succeeded by the time the meeting starts, but they're not going to hold it up for me. I get one pleasant surprise when I arrive: they have a microphone by the Rover Driver workstation now. Previously, we were the only station without a mike. You betta recognize!

They plan ambitiously. Touch-and-go sols are ambitious anyway, and this is the biggest one yet. They'll use the APXS, MB, and MI on each of two different spots on the dune -- one spot on the crest, one in the trough between two dunes. Six instrument placements total, and that's before we drive. Not that the drive will be much; Mimi is just off of our front wheel, so the only question is whether to turn in place, or back off 10cm before turning in place.

Despite all the work they're trying to achieve in the meeting, they manage to have a little fun. It's a Friday the 13th ("not on Mars," one of the scientists hastens to point out) -- but not just any 13th; it's the day before Valentine's Day, and many of us are facing, shall we say, mild spousal tension. One of the scientists named a rock "Sarah," in honor of his wife, and placed a target called "Be_My_Valentine" on it. (At one point during the discussion, he says, "Sarah's bigger" -- then thinks better of it. "Sarah fills more of our field of view," is the phrasing he ends up with.) They rag on him for this -- but not much. I think we all understand. And when one of the scientists falls asleep, a couple of the guys across the room repeatedly shine a laser pointer in his open mouth.

Well, maybe you had to be there.

Before I can get to work, Art Thompson pulls me aside for a discussion about our upcoming "megadrives." This is our plan for covering much more distance per sol than we've been doing so far. His idea is to drive "blind" (a faster driving style in which the rover drives exactly as directed from the ground, without imaging for hazards) as far as possible, then turn the autonav loose. In the middle of the day, the rover will shut off for a couple of hours to cool down, after sending us images of its new location. We'll use this to plan a second drive for the afternoon, following the same pattern -- blind-drive as far as possible, then autonav until we reach our daily limit. We're hoping to cover 50m per day, which would take us to Bonneville in less than a week if all we do is drive.

I'm for it.

The kink in the plan is how to do the afternoon drives. We'll have only a very short time to plan them, because the rover will sleep for only a couple of hours, and that's cut even shorter by downlink and uplink times on either end of its nap, plus we have to wait for the images to be turned into terrain meshes we can plan with. After this overhead is subtracted, we might have as little as thirty minutes. So the best approach is to first create some building blocks -- sequences that will do blind drives of 20m in various headings that will take us in more or less the right direction. When we get the terrain meshes, we'll pick the sequence that takes us as far as we safely can go, slap an autonav sequence in, and send it up.

I refine the idea a little, suggesting to Art that we also have a matching set of 10m drives, because there may not be a single safe 20m drive from wherever we end up, but it's better to get 10m than nothing. Art agrees, but he wants to keep it simple, so we cut it off there -- we're not going to add in 5m, 15m, or 25m variants. I don't know when I'll have time to do it, but I volunteer to create the building blocks anyway. Because that's cool.

Art is a funny guy. It was he who shepherded both rovers through ATLO[2], supervising the entire process of putting the pieces together and getting them safely off the launch pad. Now he's a Tactical Uplink Lead, supervising the daily uplink process. He's a Pathfinder veteran, enthusiastic, energetic, yet surprisingly laid back. He's one of the people pushing for a rover rivalry; as part of that, he refers to the Spirit team -- Spirit being otherwise known as MER-A -- as the "A-Team." Every time he says it, I unconsciously start humming the theme song from that show. Damn you, Art![3]

I'm still humming it as I go downstairs to start sequencing. Irritatingly, there's one more distraction before I can get started. The Sun server that runs SEQGEN for us is hosed and needs to be rebooted. This problem has come up a couple of times before, and I still don't know what's really wrong with it. It started happening just before landing, right after they switched to a different NFS configuration. It seems that SEQGEN gets as far as trying to open a lockfile, and hangs there. Once the system gets into this mode, nothing short of rebooting seems to cure it. I suspect the problem is with nfslockd, which tells other NFS clients about the lockfile. But I have no proof, and no way to track it down further. At some point, this is going to happen at just the wrong time, when we're rushing to complete our work at the end of the sol, and we'll end up losing a day over it. I'll have to talk to the SAs about it. Meanwhile, I just call the responsible parties and request a reboot.

At last I can actually start to get work done. Somehow there's never enough time for this part of the process. I manage to get most of the work done by the time Bob shows up as RP-2, but I haven't started on the drive yet, and I'm having some problems with the IDD sequencing anyway -- the simulation is complaining about some moves that look fine to me. Bob explains the problem: there's a known limitation (not really a bug) in the flight software, which RSVP is accurately reporting. In some circumstances, the IDD software effectively can't figure out how to get the arm "there" from "here." He shows me a trick to help it -- breaking one move into two different moves -- which mostly fixes the problems I'm having. We end up splitting the work anyway; I take the drive, and he finishes the IDD.

While I've been working on the IDD sequencing, the MI PUL has made a simple request: he needs to know what shadowing to expect to see in the MI images. This is not hard to do, but we need to get the sequencing right first, and the trouble I have with the simulation forces me to keep putting him off. Eventually he gives up and leaves. I feel really badly about this and decide to teach him how to do it himself, which will help both them and us. But it won't happen tonight.

Just about when my shift is supposed to end, a cameraman shows up. I'd completely forgotten about this. The JPL Media Office keeps getting requests from TV stations for footage of us driving the rovers. But they don't want to keep sending camera crews into the sequencing MSA, because it's distracting for the people who are trying to get work done. So instead, they've sent their own cameraman to get some footage that they can just hand out to stations as "B-roll" (the footage they show while the commentator talks about what you're doing). As it happens, I have more work to do anyway, so he films me doing it. For the sake of the cameraman, I put on the 3-D LCD goggles for a while, which is what everybody likes to see even though we don't actually wear them most of the time. After he leaves, I still have to put together the shadowing data for the MI PUL, so I end up staying very late. I get home barely in time to sleep. Happy Valentine's Day.


[1] Knowing what I know now, I'd make a stronger statement: Eric and Bob were exactly right; this idea was insane. Since then, we have uplinked awesome new software to the rovers (written by Chris Leger) that enables the rovers to do a go-and-touch autonomously, and safely. But as the rovers were then? Insane.

[2] Assembly, Test, and Launch Operations. As I go on to say, that's just as it sounds: putting the pieces together, testing everything, and getting it off the launch pad.

[3] Re-reading this just now, I did it again. Damn you, Art!

1 comment:

Anonymous said...

"...the spectra they got from it don't quite match anything in the library. Analysis is in progress."

That's one seriously awe inspiring statement.