AAS Reacts (10 Days Later) to FY 2013 Budget

1. Well, I’m sure they carefully considered how disappointed the President would be at their delayed response …

   But seriously, I rather suspect that there was some serious disbelief that the Astrophysics Division budget didn’t crater, like the Planetary Division budget did. Surely we’re paying for JWST mismanagement, they’d be saying. We’ve gotta be getting whacked somewhere! It’s gotta be in here somewhere!! Took a few days to reassure themselves that it wasn’t.

2. It’s kind of refreshing to know that somewhere in the 15-minute news cycle world, there is a group of intelligent people who prefer rational discourse and adequate time for discussion, first by a public policy group and then by the AAS Council, before mouthing off. This stands in favorable distinction to the AAS staff’s knee-jerk “too big to cancel” reaction when it looked as though JWST funding might be stopped. What the new statement fails to address, however, is how badly poor management of a very large project can kill not only new missions in its own discipline, but also in other disciplines (e.g. planetary) in the same Division. Since there’s no reason to believe that NASA will ever learn that lesson, there’s a very strong argument to be made that we should never have another project the size of JWST. Sad, and limiting to science, but maybe it would insure that more, and more diverse, science gets done for the taxpayer’s buck.

   1. “there’s a very strong argument to be made that we should never have another project the size of JWST“

      sunman,

      I agree with your statement, but I would also argue for a movement to another concept to help get around it, that, at first, might seem to contradict this — what I think of as a “program of programs.” We can accomplish “big” results by developing incrementally, in the same manner as you design large systems. It’s the same reason why prime contractors use subcontractors. And, of course, the first step in solving any large problem is to break it into manageable pieces. Every engineer on the planet will quote, “top-down design; bottom-up testing.” (But unfortunately too many of them don’t actually practice it on a day-to-day basis.)

      The ISS was built and launched one module at a time and grew incrementally. The Chinese are reportedly planning the same approach for their space station. The same concept has been successfully applied to thousands of programs and projects. But, is it possible to apply this approach to something like a large, complex space telescope, particularly when it includes significant levels of new technology and cutting-edge manufacturing and deployment processes? I believe it is, but I suspect others would argue.

      Something like JWST would have to be launched all at once because the precision and equipment required preclude “assembling” it in space (even the “deployment” phase of JWST is a major risk). But the actual designing and building of it can be done two different ways: with subsystems designed and tested either in parallel or sequentially. Doing it sequentially would obviously make it a much longer program (on the planned schedule), and would quite likely increase the cost (in the planned budget), but it would, I believe, very significantly improve risk management, and therefore improve the actual schedule and budget. If the riskiest and least known elements were done first, we could see things getting too far out of control without spending near as much money as was the case with JWST by the time people started raising a fuss over cost and schedule overruns. It could have been reassessed and even canceled before the sunk costs factor started gaining weight.

      I recognize that Congress’ tendency to cut and even withdraw funding on long-running programs increases the risk, especially when elections can redraw the power structure so thoroughly, and the new guys kill off the old guys’ programs. But overall, I think it would make program management more effective and more accountable, which has become a major issue at NASA. Programs don’t get late or over budget all at once; it happens day by day, but this is easy to overlook, or even hide, over the entire program hierarchy, when a program is above a certain size and complexity (this is true everywhere, not just at NASA). A sequentially executed program would greatly help reduce this lack of visibility. It could potentially also reduce the number of management levels, with all the benefits which that would imply.

      The major challenge I see in converting to sequentially executed programs is that each division would require developing and using its own new planning methods, since you don’t build and fly a satellite the same way you do a rover, a telescope, a manned mission, etc. But the hardest part of all, I suspect, would be getting the various divisions and centers to comply with the changes, since they all typically seem to childishly reject anything that challenges their autonomy. And, of course, there’s the old “that’s not the way we do it here” syndrome. But if Congress can mandate the details of NASA programs and how NASA does business, then surely NASA senior management can do so as well.

      Making programs longer would also require the NASA PAO (and those associated with it) to both revise their methods and get better at what they do, which likely means increased PAO/media costs. So, one issue to be addressed is: does this extra cost, if it exists, get charged to the individual programs or is it an overhead?

      All in all, I think sequentially executing NASA programs is worth giving serious consideration. Yes / No ?

      Steve

      1. “Something like JWST would have to be launched all at once because the
         precision and equipment required preclude “assembling” it in space (even
         the “deployment” phase of JWST is a major risk).”

         While I strongly agree with most of this post, this particular statement is flat out wrong. JWST itself couldn’t assembled in space because the architecture is not integrated in a way that would easily permit it. But the idea that “something like JWST”, as in a large space telescope with the capabilities of JWST, couldn’t be assembled in space, is just daffy. One could use the same handwaving argument to express why something like ISS could never be assembled in space.

         The “precision required” argument is nonsense. The precision attainable in in-space assemby, for use on zero-g, is much higher than that attainable in 1-g assembly. The deployment phase of JWST is certainly a major risk, but that’s not because there are people or independent telerobots there. It’s because there aren’t!

         Future progress in space science (and especially space astrophysics) is going to be absolutely dependent on being able to assemble large structures in space. That’s the fallacy of using Ares V or SLS to loft a large, pre-deployed space telescope. If that’s what we learn how to do, then telescope size is forever going to be limited by launcher size, which is a box that space astronomy desperately doesn’t want to find itself in.

         1. Hallie,

            Gravity is not the only consideration. I honestly don’t see how you can equate assembling a high-precision device (“Something like JWST”) in space to assembling the ISS (by adding on one self-contained module at a time) in space, but never having done either one, I can’t argue the point.

            Steve

            1. Let me help you out a little here. JWST as deployed and operational is a “high-precision device”. But the high precision is achieved not by assembling, but by aligning the precision optical segments (which no one is proposing be “assembled” in space) once they are assembled. That alignment is achieved with precision actuators that are bolted onto the back of the optical segments. When JWST is deployed, you kinda hope it’s close to being in alignment, but to the extent it isn’t (and it won’t be) you adjust the alignment with those actuators remotely.

               So assembling a “high precision” device in space (or on the ground) isn’t about assembling it with that high precision, it’s about assembling it with subsystems that allow hands-off precision alignment to be achieved.

               Look at it this way. There are good concepts for space interferometers that involve formation flying of optical segments. Those segments are individually highly precise, but when you kick the segments out of the launcher, they are nowhere close to being a telescope. You have to bring the segments into formation, which you don’t do with human hands or tools.

               So adding one self contained module at a time is PRECISELY how these interferometers would be made operational.

               I think you’re assuming that a telescope has to screw together at a nanometer kind of precision. That doesn’t happen on the ground, and it won’t happen in space. That’s what the actuators are for. It’s just the way it works …

            2. Once again we run out of nesting.
               In response to Hallie Wright, February 27, 1202, 05:22 PM

               Hallie,

               I understand what you’re saying, but we seem to be at cross purposes. I wasn’t worried about the post-launch deployment or alignment that you are, nor station keeping alignment of detached elements of a system. I see the on-Earth assembly of the “device” as a precision task, and by “assembly” I mean putting together the various subassemblies and large components, like they do in a JPL clean room, for example. Some of the subassemblies can be as large and massive, by themselves, as a commsat, so one might consider launching them as smaller, multiple payloads, on smaller, cheaper LVs, and then collecting them and assembling them together in orbit, and that, I’m saying, I don’t think can be done. It’s not just whacking together nuts and bolts; there’s a level of craftsmanship involved that, I believe, would be too demanding in free fall vacuum. Also, if things like sealants, thermal compounds, thread locking, internal gasketing, etc. are required, you’re not going to handle those in vacuum; you’d need an adequately large, controlled-pressure environment (one of those things that Shuttle and ISS were supposed to give us, but we don’t seem to have). For the final package to have sufficient “precision,” its component systems must have it as well, as must their assembly. That’s how I see it, but like I said, I’ve never had to do it myself.

               Steve

Comments are closed.