O.K., brain-bashing time.
1. The first oven doors opened a substantial amount before sticking.
2. The second oven doors, several sols later, opened only a small amount before sticking.
Does this imply that there has been a progressive jamming of the hinges? A steady accumulation of something in the hinges?
3. The retro rockets exposed and heated a significant area of ice under Phoenix, producing a cloud of water vapor that permeated TEGA.
4. The exposed ice has continued to sublime, wafting more vapor through the lander.
Are there critical opposing metal surfaces within the hinges which could collect water vapor and deposit it as ice, given their ambient temperatures and shielding from sunlight or other heat sources?
Could such an accumulation prevent the surfaces sliding over each other, jamming the hinges?
Has the TEGA door operation been tested under Martian conditions including this vapor flux?
Is there any way of adding heat to the door hinge areas to clear such ice?

Ice disappears from, is not added to, surfaces in the here and now of Phoenix. I don't see how icing can be the problem. A door with little dirt on it failed to open, so I don't see how dirt can be the problem. There may be a tolerance in the mechanism that was too tight for the jostle of launch or landing, but I expect they tested to rougher extremes than were actually experienced. I guess that's just a long way of saying.... I don't have a clue!

I would hope it was designed so that this is not a possibility.

It may be significant that the outer door on the first set of doors opened fully, while the inner door only opened some 40 degrees. The second set of doors only opened around 25 - 30 degrees (estimated). So the common factor appears to be adjacent doors. If this is correct then there are three more ovens with outer doors which should open fully and enable material to be delivered.

I would have thought that if ice deposition is the cause of the door problems then this would result in a go/no go scenario rather than a partial failure. But if failure of the door opening mechanism on the inner sets of oven doors does preclude sample delivery to the ovens (and as Peter59 points out the angle of the ovens to the horizontal means that there is a reasonable triangular gap at the upper part of the doors) then using the arm to free the doors would be an acceptable last resort. Not as a scrape but as a hammer tap. If this fails and results in damage to a non operating door - so what?

Wild guessing: based on the seeming precision with which they dug side-by-side trenches (they're almost exactly the same) in a ground whose position was measured only photogrammetrically, I bet they can do better than a cm in accuracy. (Does anybody know the MER accuracy? I think the joints are the same, although the arm lengths are much greater.) They've got the entire spacecraft (and TEGA) coordinates easily to some fraction of a mm and can reach it. Why not poke the doors down and release with a simple motion? Or slide the scoop down the TEGA topside so the lip is touching just above the top door hinges, then slowly command a wrist up movement? Like popping a bottle cap...

Paul

Machines are beautiful things. Once commanded they will try to move the programmed distance whether it is through air, vacuum, ice, soil, or spacecraft metal. If the programming of a self-touch is off even a fraction of a centimeter it could cause serious damage to the arm or spacecraft or both.

I'd be real hesitant to try such a move before all other options have been exhausted.

-Mike

(I've programmed liquid transfer robots and seen the speed that the needle plunges down to suck up/dispense liquid. The needle doesn't care whether it plunges through air, a rubber septa, or a hand. I always give the robot arm a healthy amount of respect.)

Maybe the most urgent question is this: Are we seeing a progressive weathering process (dust, chemical, thermal, moisture, etc.) that is affecting all of the doors simultaneously and continually reducing the likelihood that they will perform properly when called upon? If not, there is no particular urgency to move on and try the next door. It allows time to test various hypotheses and potential fixes to try to get the #5 oven working as planned. But since #5 seems to be in even worse shape than #4 was, it will be nerve-wracking for the mission managers to have to worry about whether we are in a "use it or lose it" situation with respect to the remaining doors. This makes a case for trying at least one more door, maybe not immediately adjacent to the ones that have malfunctioned, before too much more time elapses. It may throw the intended sequence into some disarray, but if oven #1 works fine and dandy then we can come back and spend ten days trying to get a sample into #5 without concern that the remaining ovens will prove unusable when their turns come. If #1 behaves like #5 or worse, then maybe the TEGA-related operations (including digging) can be accelerated to identify any remaining operable doors and use them as soon as possible. In the worst case, with practice and the proper motivation I am sure the engineers can figure out how to tickle it with the robot arm when it becomes clear that other options are exhausted.

I'm just catching up on this story after being away from UMSF last night, so apologies if my thoughts duplicate others', but it seems to me that the problem might be a build-up of dust between the two sets of doors. Thinking about it, that dirt has been sat on - and between - the adjacent door hinges for several sols now, and during that time it's been vibrated and jiggled about several times by the various activities - the multiple attempts to send material through the first oven's grille by vibrating it, robot arm operations, etc. So it seems to me that the material piled up over the gap between the doors will have settled quite a lot, with the very fine material gathering down at the bottom of the pile. Now, we've all heard the mystery of the "sticky soil" discussed, so is it possible that all the activities have led to a kind of reaction in the inverted cone of material now sitting over the gap between the doors, turning it into a kind of martian cement? That would bung up the works really fast...

Also a top-of-my-head thought... is it possible for some of the heat generated by the TEGA oven baking the first sample to radiate to the material sitting between the doors and make it react in some way? Probably not, just thinking outloud.

Boy, those TEGA guys are earning their money right now... good luck to all of you reading this. We wish you well.

From this site: http://www-robotics.jpl.nasa.gov/publicati..._2005_final.pdf

Guess too the only move they could do with the scoop is to poke the door down several times. It shouldn't be, but there may be already too much soil in the mechanism of the doors to allow that move with the doors.

Great paper, Tman!

A demonstrated accuracy of 0.8 mm was cited for the Opportunity arm. I don't think the extra lengths of the Phoenix arm would make a big difference. An important issue might be the flexing of the arm while moving into a final precise position. I really bet they could do it (after exhausting more prudent options)!

Paul

On Mark's SSI SOL 25 page, near the bottom, he has a stereo pair of the scoop at TEGA with this intriguing header:

"1302-E: SSI Documentation of RA at TEGA Teach Point & Closed Door"

Paul

Those oven doors look awfully small and flimsy next to the RA and scoop.

Seems to me this sequence (and the arm move) doesn't make sense unless using the scoop on the doors is a serious possibility in the next few Sols.. I can think of three ways to try:

• press down and release, hoping the spring will work properly given another chance
• get the lip of the scoop into the gap and flip one door open
• a side-swipe manoeuvre to push one door open, moving the scoop along parallel with the oven-door surface of the TEGA wedge
• ...others?

EDIT: do my eyes deceive me, or does it look as if the right-hand door (as we look at it) of the current oven has opened more fully? (Or is it completely closed?) I can only see the triangular shape of the door closest to the RAC:

Click to view attachment

That doesn't jibe with the mention at the most recent briefing that the robot arm stopped digging at the Snow White 2 trench, as programmed to do, when it hit ice. They probably use the current draw for the actuators as a sense of touch.

I don't know how fine that sense is, but there is some ability to have the arm not try to pass through something too solid. It's already encountered very hard stuff and it's still functioning.

This image was taken before the attempted opening so all your seeing is oven 4.

Tosol (26) we have: "SSI doc TEGA, 1318" which will be our first look like that since the opening.

To me, "teach point" means a rehearsal of where the scoop will be positioned while shaking the next sample into TEGA. The arm has a memory and if the position is fine, it can be told to return to it with the actual sample. Looks like standard routine.

Okay. Maybe I am reading too much into the header. But "closed door" is mentioned...

Paul

That imaging sequence was designed before the door was opened so the caption isn't likely to be referring to the failure to open - there was no way to know that beforehand.

Right. I just checked Emily's page and Door #5 was opened between 12:45 and 14:30 on Sol 25.

Paul

Sol 26 images started to come down and it doesn't look good...

Shouldn't be an issue. The arm is very solid and the final move to touch the doors could perform the scoop alone. Btw. this animation shows an accurate motion with at least two actuators http://www.nasa.gov/mission_pages/phoenix/...ress/13339.html
It looks like the distance to the MECA was rather stable and the used "sprinkle technique" sounds very interesting to vibrate on the doors too.

I make Door #5 to be standing up about 6-7 mm. Lots of "cushion" :-)

Paul

For those in a pessimistic mood, here are the minimum mission success criteria.

4. If TEGA, analyze at least 2 soil samples to create a profile of H2O (in the form of hydrated minerals,
adsorbed water, or possibly ice at the deepest level) and mineral abundances near the surface. It shall
also analyze an atmospheric sample in its mass spectrometer.
5. If MECA, analyze the wet chemistry of 2 soil samples.

Steve M

Success is better defined by the taste it leaves in your mouth than by some bureaucrat's list of ex-ante bullet points.
Sure, the biggest threats to mission success are in the launching and landing phases, so if you draw up a list of success criteria before launch, you'll settle for getting to the destination safely, taking a couple pictures, and doing a little science. But at this stage you can bet nobody's going to take refuge in the pre-launch criteria if the soil analysis terminates prematurely due to "door failure". Not on the heels of MRO. Anyway, they're going to work it out.

Notice that is either TEGA or Wet Chemistry, not both for minimum mission success.

For those in a optimistic mood, we only need 3 TEGA measurements for full mission success. No evidence we won't make that, one out of two so far.

Sol 27 has the interesting sequence description, a-posteriori of what is supposed to be an open door:

"SSI Documentation of TEGA (Open!?) Door" - ActId: 1338

Poor TEGA...

I suppose a true optimist would have to say there's still hope for oven 5 (not that I'd put myself in that camp)...

Interesting. SSI took a image of TEGA yestersol with the partially open door 5. Maybe they are trying to open one of the doors on the other (visible properly with SSI) side tosol? I'd be temped to try this if I were them.




True.

I'm more inclined to think this is just for documenting the door No. 5. We'll see soon enough.

An optimist would say the small opening may allow just the right amount of material through to sift through the screen without clumping up in a cohesive pile (The pessimist would say the soil may clump in a cohesive pile over the small opening).

This card-carrying pessimist is wondering if surface-film ice could be the cause of the clumpiness that prevented the sample getting through the screen of the first oven, in which case the screen will act as an perfect barrier to ice (mixed with regolith, anyway) ever making it into the oven. Anyway, let's hope purer ice collected with the scraper or rasp will be less sticky than the first soil sample, or that the soil mechanics are different with the sprinkle method.

What about the actuator itself: The soleniod that releases the locking pin? Is it possible the doors are still dragging on the pin, or is it positioned so that a partially opened door is (obviously) clear of the pin?

Why don`t they just use another oven? let`s say oven no.6. We have 6 ovens left! why do they want oven no.5? we could leave it for the end of the primary mission, and if it works then OK.

They may end up doing that. But not before they spend some time trying to figure out what is going on. There are nine weeks left in the primary mission. That's one week for each of seven unused ovens with two weeks left over. Time enough to do some testing and decide on a course of action.

Right now, I'd be happy to know which oven is number six, and how the rest are numbered, just out of curiosity.

Good TEGA article here...

I could imagine that the oven door problem is temperature related, for example. What if the chance that a door gets stuck partly open is greatest at the warmest part of the day, and if they open them first thing in the morning they will open fine?

Of course it may have nothing to do with temperature, but it's definitely worth trying to figure out what the root of the problem is before opening another oven, in case there's anything that can be done to improve the chances that they open fully.

I'm imagining pretty much the reverse, Fred. I'm thinking that the springs may have become very stiff in the Martian cold, and they're just not pushing the doors far enough to have them achieve the stable "open" position. In which case, we'd have best luck during the warmest part of the day.

However, if it's not the springs that are binding but the door hinges themselves, it's possible that the metal of the doors has shrunk in the cold, just enough to make the hinges bind. If that's the case, it's possible that, again at the warmest time of day, a good sharp rap on the side of the entire TEGA device might indeed jar the doors open.

Of course, the only thing you have available to produce a good sharp rap is the robotic arm...

-the other Doug

I imagine the team is scrambling to test door opening here on Earth as well, trying to isolate possible failure modes in a similar environment. I can't imagine them not testing the opening mechanism under martian temperatures even before launch, though.

I have to believe there is a physical impediment blocking the doors -- something out of alignment. The spring is released, putting constant pressure on the door over the entire course of a Martian day with no change in the opening size. If temperature related stiffness or tightness is the cause, some change should be seen as the temperature changes throughout the day.


They could just place the scoop in contact with the TEGA superstructure, or even on an adjacent area, and vibrate it. This vibration may be stronger that that provided by the TEGA vibrator alone.

What about having only one of the #5 doors fully open? That should give them enough room to get something in, no? I'm thinking placing the scoop lip just below the height of the open doors, then gently rotating the azimuth north, pushing down slightly the more southern door #5 but catching the more north side door and pushing it up. I think the geometry works. I don't care if it goes too far into #4 since I guess they're done with that. You could only do it in one direction so don't open any doors south of it until you've tried.

Paul

I can't wait to see -- after mulling over all our suggestions -- which one they choose.

I'd fall off my chair if they used any of them!

Paul

Hey! Whats' going on here?
I check Emily's Planetary Society blog and see this:
"Phoenix successfully delivered a sample from the Wonderland area to the Optical Microscope on Saturday (sol 26), and managed to leave enough soil in the scoop to deliver a sample from the same location to the wet chemistry laboratory and to TEGA."

Excited, I rush to UMSF to find.... nothing. I'm starting to think maybe I'm missing something, or Emily made a mistake. Then I check Mark Lemmon's image page for sol 28 and find this:
Click to view attachment
Is that a pile of dirt I see covering the partially opened doors of oven five?

So... what's going on?

No, that's still the dirt on the door #4. See sols 25 &26:
http://www.met.tamu.edu/mars/i/SS025EFF898...9_1302EL1M1.jpg
http://www.met.tamu.edu/mars/i/SS026EDN898...4_13180L1M1.jpg

The Phoenix team hasn't said much what are their current TEGA plans, just that a TEGA sample should be delivered eventually. I guess they let the MECA and wet chemistry deal with the samples first, as they figure out what to do with TEGA.

Well, Emily says that enough sample is left for delivery to WetChem and TEGA. Not that it was actually delivered there. We'll have to wait and see.

The way I read the recent activity.

Sol 25: Acquire sample to be delivered to all three instruments Microscope, Wet chemistry and TEGA. Open TEGA doors for the sample which fails.

Sol 26: Delver sample to Microscope

Sol 27: Recover from safe mode

Sol 28: Deliver sample to Wet Chemistry

Sol29: [tosol] ???

Nobody seems to have mentioned this bit from The PLS report:

Certainly, there is a design spec to ensure that everything works at Mars-ambient temperatures. Whether metric degrees or english degrees...

One simpel gotcha is the coefficient of expansion of materials. At +20o you have one clearance in the mechanism, but at -60o you have a negative clearance (ie, it binds). I once built a nice camera focusing mechanism for my telescope which had ball bearings with a perfect interference fit in the housing. At room temp. But I found that below 0o F the difference in expansion between the aluminum of the housing and the steel of the bearings yeilded a loose slip fit clearance, which resulted in inconsistent focus problems.

Also, I don't know if there is a "temperature coefficient" with springiness, but materials can become more brittle in cryo-environments, and this might affect the spring's ability to open the TEGA doors.

Mars is an alien world...

--Bill

Young's Modulus, which is one of the measurements of the tensile elasticity of a material, definitely exhibits some temperature variation and generally (always?) reflects the fact that materials get stiffer as temperature drops. My understanding of the effect is that lower temperature should create a stiffer spring though - and unless the temperature is far higher than they expected I can't see how the springs would end up being too weak to carry out their intended task. For most normal metals the change in effect over the range of temperatures experienced by Phoenix doesn't seem likely to have any serious effect (to me) see http://www.engineeringtoolbox.com/young-modulus-d_773.html.

Young's Modulus (and the Bulk\and Sheer Moduli which are two other relevant coefficients in terms of overall elastic behavior) only applies within the strain region where Hooke's law applies and that region itself will vary with temperature. I'd be very surprised if the TEGA spring opening mechanisms hadn't been tested at all potential temperature and pressure regimes though so I'd be surprised if that was the case.

I still think it's far more likely that the debris from the previous sample delivery event has found it's way into places that weren't anticipated and what we're seeing is a hinge that is jammed as a result.

This would not explain the partial opening of oven four's door. So you would be saying that the two door opening problems have two different causes.

So much for the idea that the springs had sproinged. We've been using latches and springs to open things in a space-environment and that technique has the bugs worked out. It'll be interesting to eventually determine what caused the TEGA door problem.

--Bill