This is not a NASA Website. You might learn something. It's YOUR space agency. Get involved. Take it back. Make it work - for YOU.
Astronomy

TESS Has A Hubble Problem But NASA Will Launch It Anyway (Update)

By Keith Cowing
NASA Watch
August 4, 2017
Filed under
TESS Has A Hubble Problem But NASA Will Launch It Anyway (Update)

Cameras on NASA exoplanet spacecraft slightly out of focus, Space News
“NASA confirmed July 26 that the focus of the four cameras on the Transiting Exoplanet Survey Satellite (TESS) spacecraft will drift when the spacecraft cools to operating temperatures after launch next March. The problem was noticed in recent tests when the cameras were chilled to approximately -75 degrees Celsius. “Recent tests show the cameras on TESS are slightly out of focus when placed in the cold temperatures of space where it will be operating,” NASA spokesperson Felicia Chou said in response to a SpaceNews inquiry. “After a thorough engineering evaluation, NASA has concluded TESS can fully accomplish its science mission with the cameras as they are, and will proceed with current integration activities.” … “The question is how much science degradation will there be in the results,” Boss said. “The TESS team thinks there will be a 10 percent cut in terms of the number of planets that they expect to be able to detect.”
Keith’s 27 July note: Strange that NASA will fly a flawed spacecraft that can only accomplish 90% of what it is supposed to do. Maybe NASA will explain this in a little more detail.
Keith’s 4 August update: NASA Just posted this update about TESS “NASA’s Transiting Exoplanet Survey Satellite Passes Critical Review” This is what NASA says: “Recent measurements revealed the TESS cameras to have slightly reduced focus and image quality near the outer edge of the image when placed in the cold temperatures of space, and better camera focus and image quality towards the center of the image. The difference between the designed and measured focus and image quality will not affect the mission’s science goals.” Last week this was a 10% decrease in capability. Now its no big deal, right NASA?

NASA Watch founder, Explorers Club Fellow, ex-NASA, Away Teams, Journalist, Space & Astrobiology, Lapsed climber.

61 responses to “TESS Has A Hubble Problem But NASA Will Launch It Anyway (Update)”

  1. Donald Barker says:
    0
    0

    “The TESS team thinks..” Well if you are using my tax dollars you had better damn well know. This is not horse-shoes, it is science, and by hand-waiving in your use of language and technical precision, you not part of the solution. Just my opinion.

    • fcrary says:
      0
      0

      At least they are being honest. This mission is supposed to measure things that have never been measured before. Similar things have been measured (e.g. by Kepler), but the requirements for TESS are an estimate and an extrapolation. That means there is some uncertainty involved. So, with the reduced resolution, they can only estimate what the impact will be. Their best estimate is 10%. They were probably also asked for a worst case (or three sigma lower limit.) But “The TESS team thinks…” is just a way to say this is an estimate and they are not absolutely certain the number is 10%.

    • Tim Blaxland says:
      0
      0

      It’s not the best choice of language. I’ve seen engineers, when they actually do know the answer to a high degree of certainty, allow uncertainty creep into their language. It’s an emotional element coming into play. You would be more inclined to be confident in their abilities if they used the phrase “The TESS team’s best estimate is that there will be 10% reduction…”

      As for your comment “you had better damn well know…it is science” – anyone peddling absolute knowledge without estimates and errors is not doing “science.”

      • fcrary says:
        0
        0

        The wording might be a media relations thing. “team thinks” sounds more conversational than “teams best estimate is.” It’s usually safe to assume that any press release will more readable but less accurate than what an engineer or scientist would have actually said.

    • rktsci says:
      0
      0

      This problem is engineering and not science.

    • MountainHighAstro says:
      0
      0

      as a researcher using statistics, I would be more concerned if someone tries to tell me that they “know” something

  2. John Thomas says:
    0
    0

    Actually it’s a 10% cut in what it could do. What it’s supposed to do would be the requirement and they claim it accomplishes that.

    I suspect that since it accomplishes the requirement (they claim) and the cost to delay and fix the problem would be expensive, possibly causing the project to be canceled, this was the decision they agreed upon.

  3. fcrary says:
    0
    0

    If detecting 90% of planned number of exoplanets still satisfies the science goals, not fixing the problem may be the right thing to do. It depends on the cost, but what’s better? Detecting 10% fewer exoplanets or delaying the next Explorer mission by a couple years (to find the money for a fix)?

  4. DJE51 says:
    0
    0

    Has to be a cost / benefit (of fixing it) analysis. If fixing will cost too much, then launch as is. But someone – the contractor or whoever is ultimately responsible for this – should be penalized in some fashion, either monetary or in future contract bids. Would be interested in a follow up regarding who is responsible, and what was done about it!

  5. Zen Puck says:
    0
    0

    Unlike JWST, or many other NASA Missions, TESS is a cost capped mission. Go over the cap – get cancelled. Their hands are tied.
    What might be interesting to learn is what was the Technology Readiness Level the TESS team stated in the proposal for the Cameras? What level or risk did they think they had going into Phase A with the cameras?. Perhaps they ‘missed this’ in the proposal, and the NASA review panel that evaluates the proposals missed it too.

  6. rb1957 says:
    0
    0

    I agree with Keith, it’s odd for NASA to do this. Fixing the problem “properly” may be unnecessary if the satellite can still complete it’s science mission (that’s another odd statement). Possibly more reasonably position is knowing the problem then prepare a calibration curve now … if the camera focus drifts with temperature, then record the temperature when it takes the pictures (maybe not part of the data collection now) and pre-launch measure the effect of different temperatures on the cameras so you can photoshop afterwards.

    • sunman42 says:
      0
      0

      It’s not an “odd statement.” Part of the systems engineering approach to any mission is to establish first what the top-level, scientific requirements are, and everything else flows down from that, including specific engineering requirements for hardware and software. That can all be subject to iteration, say because of launch vehicle or technology readiness or budget or other factors, but at some point before hardware is cut, the project management finalizes the scientific requirements.

      Sometimes, issues are discovered farther down the path, and that may lead to descoping of the mission, which requires modifying those top-level requirements. No one’s ever happy about that, but it does happen in some fraction of missions. If the science team is willing to accept a mission that does “90%” of what the originally agreed requirements called for, as a taxpayer, I can live with that. If there’s a quick fix that doesn’t threaten schedule (in which case it’s always, 100% of the time, the science that is hit with budget cuts), great, but sometimes such fixes involve too much risk to schedule (and thus cost, and thus science), and it’s better to go ahead with what you know works.

      • fcrary says:
        0
        0

        I’m not sure about Explorer missions, but the current requirements for Discovery and New Frontiers call for both nominal mission requirements and “performance floor” or “minimum” mission requirements. Instead of the scientists being asked if 90% (or whatever) is acceptable when a problem comes up, they are have to provide the limit in advance.

        As far as science taking the hit if there are budget issues, I’m not sure I’d say that happens 100% of the time, but it’s pretty close. However, this is another requirement problem: Actual success requires more than simply making the measurements. Someone has to look at and analyze the data. I’ve never seen a formal requirement on how many scientist-years must be funded for data analysis. It’s every bit as critical as “24 deg. x 24 deg. field of view”, but it isn’t captured in the requirements.

        • Michael Spencer says:
          0
          0

          If data analysis isn’t in the requirements, how is it funded?

          • sunman42 says:
            0
            0

            At the risk of seeming glib, I’d have to answer, “Always at a lower level than spelled out at the start of Phase B.” From experience, the best science is not done until extended phase E (that is, the first or following mission extensions beyond the project lifetime carried in Phases A – D), though there are some exceptions. In general, though. careful and thorough analysis takes time.

            And much of the good science is done not exclusively by members of the initial science teams, but when the data become available to larger scientific communities. That’s why NASA’s funding of participating scientist/guest observer/guest investigator programs is crucial to getting the most out of what we taxpayers pay for the missions’ development and ops.

          • fcrary says:
            0
            0

            That’s one of my concerns about JWST. With Hubble, we’ve had 25 cycles (years) of guest observer time and as much as 15 years of experience with some of the instruments. People have found ways to use the observatory and the instruments which were not even imagined at the time of launch (and to overcome problems which were not originally imagined.) JWST has, more-or-less, a maximum lifetime of ten years. I’m worried we’ll still be learning how to get the most out of it at the end.

          • sunman42 says:
            0
            0

            Good reason to learn how to service missions at L2?

          • fcrary says:
            0
            0

            In a sense, the data analysis is funded by implication. The formal requirements come in different levels. Level 0, “owned” by NASA headquarters, is very high level. For TESS, it might include things like “Conduct a full-sky survey of transiting, Earth-like exoplanets.” To satisfy that requirement, some data analysis is clearly required. But for the hardware, there are lower level requirements. A TESS level 1 requirement might specify the precision and signal-to-noise required to see transiting, Earth-like planets, and a level 2 requirement might specify cooling the focal plane to -75 C in order to achieve that signal-to-noise.

            If, during development, the project management wants (or needs) to shift resources around, those detailed requirements constrain (or guide) how that happens. You can ask an engineer, “If do X instead of Y, can you still meet your level 2 requirements?”

            For the science, it’s much vaguer. Everyone knows some funding for data analysis is required for to satisfy level 0 requirements. But there aren’t any derivative, lower level requirements. So if the project asks, “Can we cut your budget from X to Y?”, the scientists can’t really say, “No, our level 2 requirement is for Z people for the necessary work.” It’s a much more free-form negotiation than the engineering work that goes into hardware development.

          • sunman42 says:
            0
            0

            I don’t recall an instance in which project management ever asked the science team. They tell them what the cut is going to be, and if people are really upset, they use the usual channels to complain and attempt to get the blow lessened. Not too surprisingly, in the majority of cases, NASA management sides with project management.

          • fcrary says:
            0
            0

            I’ve always seen at least some level of negotiation. But perhaps that depends on what you mean by “management” and “scientists.” Is an instrument PI on a flagship mission a scientist or a manager? In practice, they are both. On Cassini, there is definitely some negotiation between the PIs and the Project Scientist and the Project Manager.

            Even at lower levels, where the scientist often is simply handed a budget number, there is still some back-and-forth. I’m used to seeing the scientists asked to provide a “shopping list” statement of work. The scientist, when asked says, “You should fund me to do X for $Y, Z for $W, Q for $U, etc.” and the project comes back and says, “Yes to X, no to Z, yes to Q, etc., so your budget will be $V.”

            But you seem to be working on the astrophysics side. I’m a planetary scientist, and I’ve noticed that those different fields do have different approaches to this sort of thing.

        • sunman42 says:
          0
          0

          I’m just taking a WAG here, but I’m guessing the 90% wording was thought up by Communications (the Oragnization Formerly Known as PAO), not the engineers.

          Phase D typically ends 60 or 90 days after launch, or whenever a more extensive, on-orbit checkout has been completed. After that comes Phase E, the project phase in which the project management for A-D is not involved. And I’ve never seen a quantification of the science FTEs needed to achieve minimal or fully successful requirements, only what the PIs negotiate with the project.

          • fcrary says:
            0
            0

            Technically, for selected missions, there is a PI, he’s the most senior member of project management, and that doesn’t change until (possibly) deep into an extended phase E. But yes, there are significant changes, both in management and the approach to formal requirements, between phase D and E. I just wonder if that makes sense.

            On the other hand, I’ve always been uncomfortable with the way some of the requirements are written. It’s fine to say “measure X with Y resolution.” But how can anyone sign off and say this requirement has been satisfied before launch? That is what the process expects. To be honest, you shouldn’t sign off on satisfying such a requirement until the measurement has actually been made. Or write a requirement to “be capable of measuring X with Y resolution” and sign off on that before launch.

          • sunman42 says:
            0
            0

            The pre-launch reviews aren’t signing off on having made the measurements, but on certifying that, to the best of all concerned’s knowledge from testing, the instrumentation should be able to accomplish the mission objectives. The, after the end of prime Phase E, there is an additional review to determine whether the objectives were, in fact, met. In principle, that’s when NASA declares a mission a success, though it’s been known to happen sooner if some subsystem is failing, or fuel/cryogen/detector throughput is going away faster than expected, but the requirements have already been met.

          • fcrary says:
            0
            0

            I know what the pre-launch reviews are about. I was just objecting to the wording. Technically, literally, the requirements signed off on say “measure” not “able to measure.” (At least for the ones I’ve been involved with.) Obviously, that isn’t what they mean, since those requirements are signed off on before launch (and well before the actual measurements.) I just wish they would rephrase them, so the words matched the meaning.

  7. Paul451 says:
    0
    0

    in a little more detail.

    Ha.

  8. John Thomas says:
    0
    0

    In the real world, we (engineers) get yelled out for including more in the design than what’s required. I glad that the NASA engineers and subcontractors are able to put in as much extra capability as they can for minimal cost.

    • rb1957 says:
      0
      0

      and for writing specs that are overly demanding, but no congratulations when this provides buffer to overcome shortfalls and save the mission.

      welcome to the engineering “lose lose” !

      but we’ll all talking without knowing the full problem. maybe the focus drifts slowly so that for 90% of the time it’s within tolerance ?

      • fcrary says:
        0
        0

        Rereading the story more carefully, it look like the camera’s focus is better than expected in the center of its field of view, and worse than expected at the edges. That’s not unusual for telescope optics (lots of work goes into minimizing this, but if something’s a little off, this is what you’d get.) Presumably, the 10% refers to the fraction of the field of view where this is significant enough to compromise the measurement.

    • fcrary says:
      0
      0

      There are some real problems with the way requirements for flight missions are handled. If a scientist wants “a factor of ten” improvement over previous measurements, he doesn’t mean 10.000. He probably means about ten, nine wouldn’t be a disaster, and eleven would be great. Sometimes it turns out there are good, technical reasons why cost/mass/power/whatever go up dramatically at 9.5, or why 10.5 isn’t any harder than 10.0. But the current system doesn’t really have a way to capture the scientific flexibility. Most scientists won’t say 9.0, because they know the engineers might get yelled at if they exceed requirements.

    • Michael Spencer says:
      0
      0

      Your comment, and the ones responding below, bring to mind discussions here many times about costs and cost-containment, specifications being the other side of the coin, I suppose.

      But specifications and contracts are often written in a way that sets up an adversarial situation. But that’s not what is wanted. The project is best served by a team all pulling together.

      I know it sounds like pollyanna.

      But having written many sets of specs over the years (none for anything like a scientific instrument), I’ve learned that the project invariably benefits if one can gain insight and experience from the vendor.

  9. fcrary says:
    0
    0

    MIT Lincoln Laboratory
    It sounds like a materials problem. Thermal contraction when cooling down to -75 C seems to be greater than expected, and it shifts the optics around a little. I’m not sure if I’d call that a design flaw, or a parts problem. The story does mention “recent tests”, so I don’t see evidence of any delay reporting this.

    • Daniel Woodard says:
      0
      0

      So there is no actual capability to focus the optics during operation?

      • fcrary says:
        0
        0

        You’re pushing the limits of my knowledge, both of TESS and telescope optics. But… Space-based telescopes may or may not have focusing. For small ones, no focusing and very, very solid, rigid mechanical designs are popular. That’s basically getting the alignment right on the ground, and bolting it down so firmly nothing shifts during launch. For larger telescopes, some sort of focusing on-orbit is, I think, typical. I don’t know about TESS.

        However, most focusing mechanisms involve the position and orientation of the detector and mirrors. That changes focusing over the entire field of view. As described, the problem with TESS is that the focus is better in the center of the field of view and worse on the edges. That sounds like something related to the shape, not the position, of the detector or the optical elements. Except for adaptive optics (which TESS isn’t), I don’t think people try to bend or deform the detector or mirror in flight.

      • cb450sc says:
        0
        0

        Most projects are highly allergic to adjusting focus. The issue is that you have a moving mechanism that potentially could move the wrong way and then jam completely out of focus. And hence you’ve just wiped out your science mission. For Spitzer, we had a focussing mechanism that we moved just once. And we invested a lot of effort in modeling so that we could predict exactly how much we needed to move and in which direction based on the initial image quality. Racking the focus back and forth like you do on the ground is just considered too dangerous.

  10. JB says:
    0
    0

    Apparently this has to due with phase transitions happening in the lens mount glue at the lowish operating temperatures. Discovered during test. The criticisms comparing to Hubble’s problems and “launching anyway” with a 90 % reduction in sources are less than helpful IMHO, these guys have enough to worry about already. One might worry about photometric stability.

    TESS has lots of capability, and I’m sure they wouldn’t be going ahead if it impinged on science requirements. And yes, TESS is cost capped. That reduces the team’s ability to fix a problem like this late in the game. On the other hand, the explorer approach has benefits for launching relatively cheap competed missions, and staying in the cost box.

    Good luck TESS!

    • Roger Jones says:
      0
      0

      Yeah, the admins of this site don’t seem to know much about flight hardware development – though they’re excellent at creating clickbait headlines. Reading the article, it’s clear that this is a rather mundane development issue typical of complex systems.

      • kcowing says:
        0
        0

        Actually I used to work at NASA and was involved in all aspects of flight hardware design. You?

        • Michael Spencer says:
          0
          0

          SO! A trained and NASA-experienced biologist AND an expert at click bait! 🙂

    • fcrary says:
      0
      0

      Actually, it’s a 10% reduction, or a reduction to 90%. I’m sure “90% reduction” was just a typo, but I don’t want people to get the wrong idea.

  11. cb450sc says:
    0
    0

    This really just gets to how their requirements are written, which is critical to containing costs. And I have never see the TESS requirements documents. The level 1 requirements are often pretty shocking. I have been on missions where the level 1 requirements literally allowed a non-functioning instrument. In other words, if one camera didn’t work, well that was just how they would launch.

    • fcrary says:
      0
      0

      Actually, I can think of at least one example like your non-functioning instrument being allowed. But that was a really convoluted interpretation of the level 1 requirement Specifically, multiple sensors intended to simultaneously give the specified time resolution and field of view, one of which failed during cruise. But technically, that satisfied the level 1 requirements, since those requirements didn’t actually say, “simultaneously.”

  12. HammerOn1024 says:
    0
    0

    “Strange that NASA will fly a flawed spacecraft that can only accomplish 90% of what it is supposed to do. Maybe NASA will explain this in a little more detail.”
    Not if to get that last 10% costs 5% or more of the cost of the satellite as is.
    You make a lot of hay about people not thinking before speaking… Look in the mirror.

  13. fcrary says:
    0
    0

    And the story went on to quote the project scientist: “The TESS instrument, as built, is going to achieve all the Level 1 requirement goals it was built for, and it’s going to be able to address a wide range of exciting additional science.”

    In a sense, this is how things are supposed to work. The lower level requirements (a 45 deg. field of view sounds like l level 2) are all derived from level 1 requirements. In this case, it might be something about the number of stars observed over the course of the mission. To satisfy that requirement, someone said, “We’ll do that with four, 45 deg. field of view telescopes,” and that became level 2 requirements.

    Now something didn’t quite work out, and the full field of view isn’t useful. So they reported this at their next review, and explained to the external review panel that a 42 deg. usable field of view was still good enough to satisfy their level 1 requirements. The review panel agreed. If the panel hadn’t agreed, the project would have been told to go back and fix it.

    [Note added in later edit: A usable, 42 deg. is not in any of the official statements. 90% of a 45 deg. wide field of view is, and I just did the math. And rounded incorrectly. It’s closer to 43 deg.]

    That actually speaks well of the TESS project management. I know some PIs who would (and have) choreographed their reviews very carefully to sweep things like this under the rug. The TESS project reported it and justified continuing on as is.

  14. Mark says:
    0
    0

    10% reduction in performance while still meeting 100% of science requirements. Not hard to parse through the seeming contradictions, Keith…

    • kcowing says:
      0
      0

      They are launching something with 10% less capability than it was designed to have.

      • david reich says:
        0
        0

        But apparently still 100% of the capability that it is required (tier 1) to have. “Better than” is the enemy of “Good enough”
        This is like complaining that your flight is only landing 20 minutes early when the pilot claimed that they would land 25 minutes ahead of schedule.

        • kcowing says:
          0
          0

          And yet NASA asked for – contracted for – and paid for – 100%.

          • fcrary says:
            0
            0

            I don’t think that is the usual practice. NASA contracts and pays for the level 0 requirements, and they are “owned” by NASA headquarters (i.e. they are the only people who can change them.) The level 1 requirements are owned by the project, but changes need approval from headquarters. We’re thinking about level 2 requirements, which are mostly internal to the project. As long as changes to level 2 requirements are approved by an external review panel (who decide whether or not the level 1 requirements will still be satisfied), NASA feels that they are getting what they pay for. Not in the microscopic sense of engineering details, but in the big picture sense of the project’s overall science.

          • ML says:
            0
            0

            Just like NASA paid for 90-day-long MER missions. How wasteful of them to build rovers that last for several years!

          • Michael Spencer says:
            0
            0

            You meant your comment to be taken with humor, I think, but I’ve wondered about that myself many times. So many projects go into extended lifetimes— desirable, to be sure.

          • ML says:
            0
            0

            Partly. But Keith also implies that NASA and/or the mission scientists and engineers somehow possess sufficient knowledge to build an instrument that will deliver 100% of the requirements – not less, not more. Remember when the Kepler mission figured out that the Sun is not an average star, as it was assumed, but a quieter one? It was one of the basic assumptions of the mission during planning, but in reality they needed a mission extension to beat down the stellar noise and find the desired transits. Performance margins are needed for input science inaccuracies too. (Same goes for the unexpectedly effective dust removal of the Martian winds, as far as I remember that was a welcome surprise.)

          • kcowing says:
            0
            0

            So – you sign a contract to provide a product and then you fail to meet the specs – and it is OK?

          • ML says:
            0
            0

            That’s exactly not what I said, Keith. You suggested that designing an instrument with capabilities that exceed the baseline mission requirements is plain waste of money, whereas I said it’s not that simple.

          • kcowing says:
            0
            0

            Why don’t you ask NASA about this. I did.

          • fcrary says:
            0
            0

            MER was exceptional, because they had not expected gusts to clear dust off the solar panels. But in most cases, the long extended missions are due to reliability requirements. Take Cassini, by doing everything they could think of, to make sure it would work for eleven years (seven year cruise and four year prime mission), they also made it very likely it would last for twenty years. In point of fact, it could probably last longer except that we’ve almost run the fuel tanks dry.

          • Michael Spencer says:
            0
            0

            As a point of speculation, and closer to the subject of Keith’s post, I wonder if measurable money could have been saved in the construction of Cassini if it were less robust? Not knowable, I suppose. But looking at the stunning cost of these missions it’s a footnote at least.

          • fcrary says:
            0
            0

            You can certainly save money by making a spacecraft less robust. How much you could save and how much less robust is debatable. For Cassini, like almost all actual missions, the numbers are hard to estimate. But for other missions, it’s at least possible.

            GRACE, GRAIL and GRACE-FO are example where we can try to make an estimate. They were (or will be) two-spacecraft missions to do extremely high precision gravity mapping. GRACE has done things like measure month-to-month changes in the depth of the water table during a drought. It was developed as a “pathfinder” mission, with development and demonstration of the technique being on par with the science. I think that makes it what NASA would call a class C or D mission (i.e. less effort to make it as reliable and robust as possible.) GRACE cost $127 million (about $170 million in FY17 dollars), designed for a five year prime mission and it is in the 15th year of operations.

            GRAIL was, in effect, a GRAIL reflight, only sending it to the Moon as a Discovery mission (and therefore probably developed at a Class A or B mission, where greater effort is taken for robustness and reliability.) It cost $375 million (about $405 million in FY17 dollars), had a three month prime mission and was deorbited after nine months of operations. (Note that, unlike the Earth, there were no serious expectations of month-to-month variations in the lunar gravity field, so continued operations would largely have been redundant.)

            GRAIL-FO is, literally GRACE Follow-On. An improved version of GRACE, but a pure-science mission and not a “pathfinder”. It will launch later this year and cost $432 million, if I found the correct number. (It is a collaboration with Germany, and I’m not 100% sure if the $432 million number includes their 77 million Euro contribution.)

            So, in this one case, I guess you could say that pulling out all the stops to ensure reliability (i.e. Class C/D versus Class A/B) is about a factor of two in cost. I’ve talked to people involved with other missions, and, although they didn’t have any specific cost comparisons to point to, a factor of two or three was a reasonable ballpark figure. A National Academy of Sciences study recently said the difference, in terms of reliability is roughly 80% versus 90% chance of achieving prime mission goals.

            Of course, this isn’t a complete picture. You could argue that, by striving for high reliability of prime mission success, you are also assuring a longer or better extended mission. In fact, the Cassini project argued that the extended mission was equivalent to a follow-on mission. (E.g. four Enceladus encounters in the prime mission versus 19 in the extended mission, motivated by the prime mission discovery of how interesting that moon is.) But, if it is a factor of two difference in cost, the counter argument is that you’d save enough for an actual follow-on mission, and a more modern one with more state-of-the-art instruments tailored to original discoveries.

      • fcrary says:
        0
        0

        I think that’s called performance margin. Designing the system to be a little more capable than it needs to be, so if (when) something doesn’t quite work as planned, the system can perform well enough to satisfy the higher level requirements.