NASA Will Not Be Launching InSight To Mars in 2016
Keith’s 10:15 am note: NASA sources report that NASA SMD has decided to cancel the launch of Mars InSight in 2016 due to problems with the French seismometer. There is a design flaw in the instrument – one that prevents it from being able to maintain the vacuum needed to operate on the surface of Mars – or for that capability to be adequately tested prior to launch so as to assure that the instrument will operate. Without being able to certify that this instrument works there is not much point in launching. CNES has been trying to get this instrument to work for most of 2015 to no avail. NASA is not pleased.
You can listen to the press briefing at http://nasa.gov/newsaudio
NASA Suspends 2016 Launch of InSight Mission to Mars
“After thorough examination, NASA managers have decided to call off the planned March 2016 launch of the Interior Exploration using Seismic Investigations Geodesy and Heat Transport (InSight) mission. The decision follows unsuccessful attempts to repair a leak in a section of the prime instrument in the science payload.”
NASA delays Mars InSight mission, Science
“We’re all just pretty disappointed right now. Devastated would be a better word,” says Lisa Pratt, a biogeochemist at Indiana University in Bloomington and chairwoman of a Mars advisory committee for NASA. “Everyone has been waiting to get a seismic instrument on Mars after Viking.”
NASA cancels March launch to Mars, Nature
“NASA chose InSight over two other finalists in the Discovery class, a competition of planetary missions costing no more than US$425 million: a boat that would have sailed on the lakes of Titan, and a probe that would have hopped repeatedly across the surface of a comet’s nucleus. InSight’s delay may change scheduling of other NASA missions, says Elkins-Tanton, who leads one mission proposal, to the metallic asteroid Psyche, which is competing to be the Discovery mission to launch in the series after InSight.”
– NASA Mars InSight Team Addressing Vacuum Leak on Key Science Instrument
– Mars InSight Spacecraft Shipped to California for March Launch, earlier post
– Payload Problems May Delay Mars InSight Launch (Update), earlier post
I would like to think that there would be fallout from this. However, it’s more likely that all we’ll get is this short announcement followed by a significant reduction in the R & A budget and/or a descope from 2 to 1 Discovery selections in the next round, with the savings used to cover the cost of keeping InSight in storage for 2 years.
What kind of fallout? It’s a contributed, foreign instrument. And the latest AO now limits contributed, foreign instruments.
Cancellation? This mission was picked primarily because it was supposed to be ‘low risk’ compared with the two ASRG powered missions it was competing against. However it was always based on a collection of instruments built by foreign ‘contributors’. Now it’s going to have a 2 year delay after CNES cannot build a vacuum chamber?
That’s where the fallout must come in, whether it’s for JPL, HQ, or project management. NASA has already spent $525 million on this mission and the primary science experiment may need a complete re-design over a component that has been replicated in various forms in spacecraft going back 40 years. How was this not identified with enough schedule margin to prevent the delay? There’s really no excuse for it, regardless of what the new AO restrictions are.
I think I agree with much of that, but you may be being too hard about the technology. I’m not sure they are using a vaccum chamber, with the constant pumping that implies. We could be talking about a sealed chamber with no pumps.and no significant gas.leaking in for.a few.years. That’s not easy. So I’d say someone dropped the ball, but I’d go easy on the part about 40 year old technology.
While I think it’s fair to call me on the 40 year old technology line….somewhat….. it is true that sealed vacuum systems are not new. Every CCD camera in space today is effectively an un-pumped vacuum system operating at a lower pressure than the CNES team says they can tolerate. The technology isn’t difficult, though I concede that it is always possible to find a flaw.
However….
My bigger question is why these problems are being identified 4 months before a strict launch window. Environmental testing should have been done long ago. If it had been, then the flaws in the chamber would be fixed. This is why T & E exists. It seems to me that this instrument was pitched and built on a success oriented schedule that required everything to go right. That’s a bad model for developing key spacecraft components, and there’s no way it happens without someone at NASA greenlighting it.
I heard a rumor about this in October, so they’ve known about this (and have been trying to fix it) for longer than you might think. Normally, every single development problem doesn’t instantly generate a press release. They are busy trying to make it work before the problem has a significant impact on the mission. The press releases come out when they can’t, and there is a significant impact.
I don’t know about the pressure sensitivities of the seismometer, but unless it’s something like 10E-6 Torr, people make static vacuum fixtures like this all the time. It’s not exotic tech.
According to the press conference the instrument can operate to 0.1 mbar, or about 0.07 Torr. That’s not much of a challenge really. The leak must be big.
If those numbers are being reported right, that’s only a ~ 60:1 pressure difference with the outside atmosphere, i.e., like having to maintain a chamber at 125 mbar on Earth.
Yep! I could build a chamber that could maintain that using chewing gum as a gasket and only 2 screws per flange.
Well, I stand corrected. The Science article quoted someone on the project specifying a 0.1 millibar (0.075 Torr) requirement. I’m surprised that was a problem. I guess I’m too used to harder vacuum systems: For those applications, you haven’t really pumped down untill you get well below 1e-5 Torr.
I don’t see any fallout happening at JPL. I mean, Curiosity overrun (at around the cost of the whole InSight mission) and NASA HQ’s response was “hey, go build another one!”
I’m actually concerned about the fallout from not canceling the mission. I hate to say it, but having a.credible cost cap is very important to avoiding cost.overruns in the Discovery line. If Insght goes significantly over budget due to the delay and is not canceled, what will that do to future missions? How seriously will people take the cost cap, if they know.it.isn’t.enforced? I don’t think it is a good idea for Discovery missions to have the impression they are “too big to fail.” Which may.mean there are no good solutions to this problem.
The choice that NASA have to make is continue funding Insight or fund a new Discovery mission.
There is no guarantee that the French instrument will actually workfor the next launch window. So the mission should be cancel as soon as possible. Don’t spend any more on this failed mission.
This is exactly the problem that hit us with MSL (Curiousity). NASA had been in the habit of sending a mission to Mars every couple years as we pass Mars on the inside track. When its suspension system had issues that delayed the launch, a lot of folks thought, “That’s unfortunate, but in two years, we can launch MSL _and another mission._” Then we found out that NASA had pushed back all the missions downstream of MSL. It wasn’t clear how much this was technically necessary (too much of a burden on Mission Control(s), DSN, etc.?) and how much it was a cost saving measure that was taking advantage of a bad situation.
I fear pushing back the launch of InSight will result in a domino effect that will push back the rest of the Discovery missions.
Any impact Curiosity had on later Mars missions (or NASA missions in general) was all about cost. Missions with competing DSN requirements are always a problem, and an extra 2011 lander or orbiter wouldn’t have made a difference. Staffing for operations also wouldn’t have been an issue. If there had been another mission in the pipeline, their planned operations staff would have been working phase C/D development not Curiosity operations.
The real problem was cost: There haven’t been many subsequent Mars program missions because Curiosity eat up the program’s budget and the delay, alone, added about $400 million to the problem. At least that was the estimate at the time. I never saw the actuals, but I was working on a different project’s extended mission proposal at the time, and the relative costs made an impression. The standard practice during such a delay is to remain at full phase C/D staffing levels, since those people are needed at and immediately after launch. 26 months at that level is easily 15% of the total mission cost.
I don’t know about keeping full Phase C staffing for a straight delay in launch. You’ve already built the spacecraft and presumably done a lot of the testing. You need the folks for Launch Ops, but that’s usually a fraction of the full assembly. There are lots of missions that “bathtub” the workforce during a delay. In fact, a lot of engineering staff might have already moved on to the next project. This might be a bit different, in that the spacecraft isn’t fully assembled, but it’s not like you need all the folks who assembled and tested the other instruments and subsystems.
Well, I think we’re mixing up what can be done, should be done, and is done by some institutions. Projects often don’t staff down very far during a delay. Somewhere during phase D, they are replacing the people who assembled the subsystems and staffing up operations. But that doesn’t have a huge impact on the number of FTEs. They also tend to keep on a fair fraction of the people who built everything; they are needed for pre-launch (as in on the pad) testing and post-launch deployments and commissioning. Moving people off to other projects is a problem, since institutions like JPL typically don’t have people working on several projects at once I don’t know why, but I’ve seen more people working 100% on a single project that 25% on four projects. I’m not saying you couldn’t ride through a delay on a very low budget. I’m saying MSL didn’t, and it isn’t as common a practice as you might assume.
Damn. There goes the launch window. I hope they can get it to work in time for the next one.
This is why NASA should build multiple copies of each probe. It’d make it more likely that at least one of them wouldn’t be faulty and thus always available for launch.
It’s.been a long time since NASA did that with a whole spacecraft. The last planetary mission I can think of we’re the MER rovers, and before that, Voyager. But it is more common at the system level. Building two flight-quality copies of an instrument, one for.flight and one for post-launch ground tests to diagnose and solve any problems which come up. That has the benefit of having a.flight spare, if a.last.minute.problem comes up with the flight model.
NASA rarely does it is not argument for doing that way.
The planetary program would benefit greatly by moving away from bespoke custom artisanal spacecraft and moving to a production line of very similar craft — a la planetary resources. Multiple launches ought to be the norm for reliability reasons.
I don’t entirely disagree. I was.describing the current practice,.not what I’d call an optimal approach. But redundancy isn’t a magic cure. Two identical vehicles may have identical problems. That’s, for example, one of the reasons the Ariane 501 launch delivered billion dollar (well, euro, or whatever ESA used in the 90s) science mission to the water off the South American coast.
Identical copies built and launched at the same time can suffer identical, systematic errors of course. I do favor multiple concurrent missions, but I also favor common design heritage across many launch windows (more so than we’ve done previously).
This mission ‘beat’ the other Discovery mission competitors like the proposed Titan surface explorer, with one of the justifications being technically ‘low risk’.
These competed missions typically require the proposer to describe how they would continue with the mission development in the event a ‘contribution’ does not materialize. Wonder what happened in this regard.
Half a billion dollar mission down the drain because the French couldn’t hold up their end of the bargain! Yet another lesson for those that think international collaboration is the be all end all. Pathetic and no guarantee that this mission will ever even fly now.
I’d go easy on blaming the international cooperation here. it’s not like NASA hasn’t screwed their international partners over, ever. ExoMars, ahem.
ISS orbit…Ahem
Have CNES pay the additional costs incurred by their error.
That isn’t going to happen. Collaborations between NASA and other countries under a fairly strict rule, which forbids exchange of.funds. NASA can’t ask CNES to pay for work done in the US. It also means a US mission can’t spend its reserves to help solve.problems with foreign-contributed hardware. I don’t know if that is relevant to this case.
There is foreign policy and then there is space exploration. The two shouldn’t be mixed. This is all because were not willing to pay the freight ourselves.
I’m sure there are things that I am missing but… I don’t see how this can be that hard. I may not be able to weld two gummi bears together, but there’s a guy in our machine shop who can do high-vacuum compatible welding like he was ordering a pizza. How does it take multiple months to fail to fix a weld when there’s a half a billion dollar mission on the line?
We’re guessing, and I’m not sure if that’s useful. I can imagine possibilities. What if the sensors needed a free through across the chamber wall (which they would) and the company making the feed through had a QA problem? What if any number of things? I’ve seen enough flight anomalies to know that very subtle issues can cause problems for an apparently simple engineering job. Let’s not go placing blame before we know the details. We know that something very wrong happened, but I think we don’t know enough to say what.
Don’t worry folks this is the NASA that is taking us on the “Journey to Mars”…Right.
Sure would be nice if these live audio streams were available for replay.
Everyone interested isn’t hovering over their web browsers waiting for the next show. It would be nice to hear what people actually said, unfiltered by reporters.
Interestingly, http://www.nasa.gov/news/me… (“latest audio”) hasn’t been updated since 2013.
As for the snarky comments on what went wrong and canceling the mission: how about waiting until the failure analysis report is available?
Engineering in general, and space in particular really IS hard. This could be anything from a design issue to a manufacturing tool issue to a materials issue to an operator error. Or a test issue. Or maybe it’s not a weld, but a leak around a seal. Or something else. And contributory causes could be political, budget and/or schedule pressures. We won’t know until it’s documented and the fix is documented and verified.
People who work on these programs are dedicated, competent professionals. They often put their work ahead of their personal lives for years at a time. It’s a lot harder than blogging after the fact. Whatever happened, they deserve respect. Yes, even the managers. Remember the golden rule.
The important thing is that when the root cause(s) is (are) identified, everyone learns so that they aren’t repeated. And a rational decision can be made on the future.
That said, a common critique is about how long it takes to get things done. Much of that time is due to long checklists of “don’t repeat that mistake”, some mistakes going back to the 1800s. (Yes, and under-resourcing, oscillating program demands, poor scheduling, etc)
So to get anything done, programs have to take, and manage risks. Sometimes they materialize and the available mitigation strategies are insufficient.
However much fun it is to throw stones at what happened, remember that the targets are real people who did their best.
(And no, I’m not associated with NASA or this program in any way other than as a taxpayer.)
I agree about not second-guessing the engineering and failure analysis. We don’t have enough details for that. But I do worry about the management.
Specifically, Discovery missions (or, the 2014 selection in particular) require a 25% cost reserve at the time of selection (i.e. if their best estimate is $400 million, they have to have an extra $100 million in the budget.) That’s supposed to cover any unexpected problems which come up. Unfortunately, it is a common practice to spend reserves down to zero at the time of launch, even if there aren’t any unexpected problems. I don’t think that’s a great idea for various reasons.
I’d like to see InSight find a way to hold until the next launch window, and do so without exceeding their cost cap. If they’ve spent down their reserves, that may not be possible.
Of course, the international contribution is another complication. That 25% reserve is on the US side. I don’t know how it works for international contributions. I do know the project manager couldn’t send any of his budget reserves across the Atlantic. That must make problems more difficult to resolve.
It’s also important to keep in mind that this is THE primary science instrument and it was so far behind schedule that NASA was actually planning to let them install it on a spacecraft that had already been delivered to the *launch site*. That is bordering on insane.
The post-mortem on this really needs to focus on how the schedule margins could have been so far off and that NASA could find itself in a situation where it wasted so much money sending the payload and, presumably, the booster to the launch site when their go-no go instrument didn’t work.
“insane”
Could you clarify this for those of us not involved directly? Aren’t there extensive clean room facilities at the Cape?
One could imagine the need for specialized tools, is that the issue? Or is it actually handling the spacecraft. or?
When an instrument is installed on a S/C there are many systems integration tests to be run including things like electronics, thermal, mass etc. Many of these test require unique setups like a large thermo-vac chamber and ground test equipment. Installing a prime instrument at the launch site is not a good idea in general. I agree it’s a sign this train went off the rails long before we heard about it.
In this case, it’s Vandenberg, not the Cape, but that’s a detail. Instruments have been delivered to the launch site before (one on Cassini, for example) but it’s not a good idea. There probably isn’t one, single problem. It’s just that you’d need lots (dozens to hundreds) of little things to all go right. That’s somewhere between optimistic and insane.
The two examples come to mind. First, I don’t think the Cape or Vandenberg have test facilities for vibration or thermal vacuum. That means skipping full-up testing for the instrument. For the rest of the spacecraft, you’d have to test before delivery using a mockup of the instrument, which isn’t ideal, and then you’d have to partially disassemble it, install the instrument, and reassemble without retesting. Again, that’s not ideal.
Disassembly and reassembly is my other example. Spacecraft are designed to be put together in a particular order. A big example is the instrument needs to go on the lander before the lander goes inside the entry shell, which has to happen before the lander/entry shell is mated to the carrier vehicle. Small examples go down to the cable for an instrument going on before it’s tested (obvious) and that should happen before the antenna goes on, since it blocks access the the cable connections (which you might have to fiddle with in case the tests show a loose connection.) I’m making that one up, since I don’t know how the cables on InSight are routed. But that’s the sort of thing that comes up in assembly. If you deliver a system late (e.g. to the launch site) then you’re doing things in the wrong order. That definitely means some extra work, and possible lots of extra work if everything doesn’t go as planned. And you wouldn’t be doing any of that in the first place, unless you were already behind schedule.
Thank you Boardman and fcrary.
” NASA was actually planning to let them install it on a spacecraft that had already been delivered to the *launch site*. That is bordering on insane.”
That comment is completely idiotic and nonsense. Instruments being installed on a spacecraft at the launch site is common place and SOP for large spacecraft.
Comments like this show the lack of knowledge of the poster.
Could you give some examples? I’ve been involved in five planetary missions (Galileo, Deep Space One, Cassini, Juno and MAVEN), and followed a larger number of them fairly carefully. I can only think of two or three cases of instruments delivered to the launch site. In all of those cases, it was considered unusual, a source of concern, and very definitely not what the original schedule called for. Perhaps Earth-orbiting missions work differently; I probably wouldn’t know about that.
Sorry, but it is far from “SOP”.
Spacecraft require Thermal vacuum, vibration, acoustics, and EMI/EMC testing prior to launch. Most of these tests are required by the customer before OK to ship (JPL, Goddard, etc.) and CANNOT be done at the launch base.
Spacecraft I&T I have participated in: MGS, Mars 01 Lander, Mars 01 Orbiter, Stardust, Genesis, MER rovers, Deep Impact, Kepler, NPP, JPSS.
of all of those missions Only Stardust had ANY science related piece installed at the launch base and that was the Aerogel collector and it was done that way intentionally to minimize the risk of contamination.
So your “SOP” statement is way off base.
3 of 5 cases is more than half and hence: common. Add MRO, LRO, MMS, RBSP, Juno, DSCOVR, and others. And it is not “far” from ‘SOP” nor off base, it happens a lot.
Many components have been removed from spacecraft after Thermal vacuum, vibration, acoustics, and EMI/EMC testing to be reworked and reinstalled at the launch site. Sometime only qual components are installed for the spacecraft tests and the flight components are installed at the launch site (after the components going through the testing apart from the spacecraft”
And another strike against your argument,
spacecraft like Cassini, Galileo, MER, MRO, MSL, Juno, etc, were disassembled after testing at the “factory” to allow for shipment to the launch site.
Sorry, but it is still far from “Standard” as you suggest above.
For me it is 0 for 10.
And YES we have installed flight components at the launch site. Stardust had almost ALL of the flight avionics removed before shipment and reinstalled at KSC due to a late breaking GIDEP on dozens of transistors.
I have still NEVER had a flight science instrument NOT go through full system level environmental testing prior to shipment.
This story is a perfect example of WHY. If you screw up, you can get cancelled.
I’m missing something here. I said instruments delivered to the pad had happened in two, possibly three cases, out of the five I’ve worked on (including Juno) and about an equal number of missions I didn’t work on personally but followed closely or had friends working on (i.e. where I probably would have heard about such it.) SJG_2010 listed ten more he’d worked on, none of which had instruments delivered to the pad. Isn’t that 2-3 out of 15 to 20, and therefore rare?
I’m also unaware of instruments delivered to the pad for the missions you listed, with the possible exception of Juno. That’s the maybe on my list, because the radio science/gravity experiment’s Ka band transponder had a problem. (The Italians provided it, and the L’Aquila earthquake of 2009 literally brought the roof down on their flight model. They delivered, but I know it was tight, and I don’t remember if they had to deliver to the pad.) Could you list some specific examples for the missions you mention?
I’m also not talking about a post-testing refurbishment. That’s not unheard of, although it usually gets redelivered to the spacecraft manufacturer rather than the pad. That’s a planned part of the schedule and can be skipped if necessary without undue risk. I’m talking about a real, the spacecraft-never-saw-the-instrument-before delivery to the pad.
If the vacuum leak in the CNES SEIS instrument forces extra costs on the Discovery program, possibly reducing its number of future missions, then CNES (or France) should be held financially liable.
I looked at a couple french newspapers with the highest readership, in both they wrote that CNES was taking the hit for the loss of launch.
Could you be more specific about that? My French is limited, so I can’t really follow their newspapers. What sort of hit is CNES taking? A public embarrassment over the problem? A budget hit to cover the costs of the launch slip? If so, is that the costs in France, the cost to all ESA.partners, or the cost to everyone (including the US). The former is quite possible, but the later two strike me as unlikey. It would contradict inter-government agreements and my understanding of ESA funding policies. Since you can search the French press better than I, additional information would be appreciated.
ESA wasn’t involved at all. One must clearly distinguish between ESA itself and its member organisations, which also act on behalf of their own as national space agencies.
You are quite right and I should have been more careful about the distinction. The european contributions are from France, Germany and (I think) Spain, not ESA. But government policies on exchanging funds would still interfere with financial compensation for the delay.
There’s also a contribution from the UK; already delivered.
The papers were reporting that CNES acknowledged that it was their fault the launch date was missed. Taking a hit from the public would be my take… Neither article went into how it was going to finally play out budget wise as it was still early in the reporting.
The three seismometers in the instrument, sensitive enough to detect vibrations as slight as the width of an atom, require a near-perfect vacuum for precise measurements.
The seismometers sit within a sphere about nine inches in diameter. Bruce Banerdt, InSight’s principal investigator, said that during tests of the instrument, still in France, air was pumped out to a pressure of about one ten-millionth of a millibar, or less than a billionth of the Earth’s atmospheric pressure of about 1,000 millibars.
Over the course of the mission, the vacuum would gradually rise by a factor of 10,000, to about a thousandth of a millibar, because of gases released within the instrument. Dr. Banerdt said the instrument would still function if the pressure were 100 times higher, at a tenth of a millibar.
But the leaks were large enough that the pressure inside rose to two tenths of a millibar over the course of a few days, “which is by most standards a pretty darn good vacuum,” Dr. Banerdt said. “But for our purposes, we needed a better vacuum than that.”
Thanks, Yale, for finding and posting these details.
Does anyone wonder why such accuracy is specified? I wonder about the utility of such rarified data. Surely even the wind blowing on Mars- quite slight by any measure, even at high speeds- could upset the instrument.
I think I’m asking this: wouldn’t dropping the accuracy by a couple orders of magnitude cause only a very slight degradation in the data?
Placing the sensor unit on the ground achieves better contact with the ground (let’s hope the thing could actually sit on rock), and second, it is down in the boundary layer at the surface, here wind effects are minimized.
Well, you’re certainly right about the wind. I found a description of the Viking lander seismometers which reads, “The seismometer on Viking Lander 1 would not work after landing, and the seismometer on Viking Lander 2 detected only one event that may have been seismic. Nevertheless, it provided data on wind velocity at the landing site to supplement information from the meteorology experiment.” Those were, however, deck mounted.
I’m not sure about the details of the InSight instrument, but it does use three seismometers. One possible reason would be noise reduction. When it comes to the details, there are many ways to do this, but it’s possible to use signals from two or more, identical, co-located sensors to eliminate extraneous signals. That’s basically what noise-canceling headphones do. For this, the higher precision and quality the measurements, the better you can remove noise.
I’ll also guess they are expecting very weak signals. They are trying to study the core, and that involves observing seismic waves which have propagated most of the way through the planet. On Earth, that sort of work generally relies on signals from relatively major earthquakes and (in the past) nuclear tests. The former aren’t likely on Mars due to the planet’s low level of geological activity. The later isn’t likely since, to the best of my knowledge, Edward Teller is the only person who ever suggested that sort of active experiment. So InSight is probably planning on the signals from whatever weak earthquakes may exist and from meteor impacts. Measuring that, after the signal has propagated all the way to the core and back up to the surface, would take very sensitive measurements.
Well, OK, and I understand the idea of cancelling by comparison. My question was the accuracy- measuring the width of an atom? really?
I really have no idea what that “width of an atom” business means. seismometers measure acceleration, not distance. Someone probably did not dubious calculation involving acceleration and frequency of the oscillations, to create an impressive-sounding sound bite. As far as the actual requirements go, they are looking for faint signals, and I doubt they are sure how faint. So they would err on the side of caution. They don’t want to risk a non-detection.
details of the instrument (in English)
https://insight.cnes.fr/en/…
Something is very strange about this. France’s engineering talent is top shelf. They’re sending a laser cooled atom trap clock to the ISS in 2016. It has a large ultra high vacuum chamber. The leak/outgassing rates discussed here make it sound like an open hole or a gigantic fingerprint somewhere.
It is strange, and I’ll be interested to see the analysis. But outgassing and fingerprints would surprise me. That’s a known problem, and a mistake.the even excellent people can make. I know of.one case,.in the US, where a.highly competent group accidentally left a ballpoint pen in a calibration vacuum chamber while.testing a flight instrument. That was a mess. But excess outgassing from contamination is a known and easily diagnosed and corrected.problem. That does.not seem to be consistent with what we know about the InSight problem. It sounds.like a.problem which was hard to diagnose and fit. I’d really like to know what that was: I’m a collector of past.mistakes, because I try to learn from them.
Agreed, wholeheartedly. I was speaking somewhat facetiously about an actual fingerprint. I hope the analysis does get published sometime. Maybe a pumpout hole was left out of one fastener or some subtle glitch like that…