Stern Looks for Way Out of NASA's Budget Squeeze Options

[nprev]

Hmm. Very good points, Steve, and they seem to harken back to the central concept of sustainability.

The arguably sporadic nature of UMSF does not lend itself well to sustainability; most missions are either unique in nature or rapidly overcome by technological advances to the satisfaction of no one. I personally think that if we could maintain a steady-state launch schedule (yep...big pipe dream there, I know) then we could realize many economies of scale, not the least of which being stable infrastructures that could be well understood and subsequently improved using lessons learned. Complete new starts, which are the rule rather than the exception damn near every time (since planetary missions usually have to hit the ground running merely to survive) are both costly and risky; gotta stop this.

[Greg Hullender]

Helvick: I agree competely about the JWST.

Steve: So why didn't you guys switch to Gallium Arsenide when that was hot? You take careful, measured steps, and you count the cost. I'm all for responsible risk taking. I just don't think NASA has been as responsible as I'd like lately.

--Greg

[stevesliva]

Steve: So why didn't you guys switch to Gallium Arsenide when that was hot?

I work for IBM, and we've got SiGe Sticking with silicon has its advantages. Certainly there are many "rumors-of-my-demise-are-greatly-exaggerated" moments in the industry. CMOS keeps standing up and declaring, "I'm not dead yet!"

I'm all for proven technologies, myself. I wish we had some more RTGs kicking around, but there aren't, and you just know when more are manufactured there's going to be some sort of technological leap forward. That's just the way it is. Change for change's sake.

I definitely agree that there's got to be something that can be re-used, but it may well only be within a pretty narrow mission description... messenger, dawn, new horizons, MRO... they're so different! The first step would be limited-instrument mars orbiters sharing the same vehicle over several launches, and I'd love to see several outer planets missions sharing the same design, but it just doesn't seem to work that way. It seems the the incremental costs of new development and risk are worthwhile compared to the costs of doing the same thing several times. This could be because the PIs are underestimating development expense, but it's also because missions are required to carry new instruments to new places to guarantee truly new discoveries rather than just more sharper photos. It's got to be harder to pitch "let's just go and take more photos of Europa!" versus, "Let's go take more photos, AND bring radar, AND bring penetrators, AND..." How much more would we learn if we just launched another Galileo? And yet no one is proposing that... it's not just RISK that is being pushed, but the requirement that something's got to be NEW on each mission.

[dvandorn]

Unfortunately, the tendency towards "there must be something NEW each mission" is driven, not always by science, but by "sexiness" and an attempt to sell a project. MARSIS and SHARAD, for example, are good ideas, but the data that comes from them are incredibly hard to interpret. Just as with the Lunar Sounder flown on Apollo 17, these radar "glimpses" into subsurface structures provide data that requires you to know more than we *do* know about subsurface conditions in order to get valid interpretations.

I remember during Apollo, there was such a push to do something new on every subsequent landing, without having any time to analyze the results of the last two landings, that some of the experiments made little sense, or were rushed so much that they had technical or design failures. This was especially true on Apollo 17, where everyone's pet experiment was going to be flown or *never* be flown. So you got such time-consuming and otherwise marginally useful experiments as the Surface Electrical Properties experiment, the Lunar Sounder, and the Lunar Surface Gravimeter. The data returned from the first two was marginal at best, and the speed with which the last one was assembled led to a mistake in the balancing of its central measuring device, a free-floating beam structure, which rendered it useless.

Thankfully, at least for outer planet probes, we have enough time between the last one and the next one that the choice of sensors we fly in the future is at least strongly influenced by the results we've seen from earlier probes. And with lead times in the tens of years between probes to given outer planets, you would naturally expect a lot of improvements in technology from one probe to the next.

You need to have a rapid-fire series of missions to take advantage of volume efficiencies, and it's just not reasonable to send out a probe every six months to a year, as we did during the Golden Age of the 1960s. Back then, if you wanted to send a probe to Venus, well, just borrow a Ranger spacecraft body, adjust its instrumentation, and send it on its way. Build a dozen octagonal spacecraft busses and then outfit them for the mission at hand, a la Mariners 3 through 9. But when it will be 10 years or more between outer planet missions, it makes no sense to standardize your bus -- the technology will advance enough between missions it makes more sense to build new each time. It may be more expensive, but it makes more sense.

-the other Doug

[nprev]

True words, DV; definitely food for thought.

The idea of economy of scale would only make sense if it was also designed around an integrated science campaign for a particular target using Discovery-class missions rather than Flagships. Using Titan as an example, it might make sense to build a dedicated radar mapping orbiter, followed shortly by an atmospheric investigation platform, followed later by a lander/balloon. The two orbiters could probably use similar if not identical busses rated for outer-system operations; this infrastructure then could be used for similar missions to the other gas giants.

Again, though, this only makes sense if the general idea is to accomplish very specific goals with limited instrumentation for each mission rather than to attempt to answer an entire suite of science questions using a single platform.

[mchan]

I'll have some news on the JWST over on the main JWST thread later but for now the headline points to refute the implication that JWST and GPB are comparable - JWST is a long term multi instrument observatory that will provide 10-1000x the sensitivity of current observing instruments for a 10 year operational period. Comparing that to a one off mission with one experiment is not fair. It has overrun its budget badly but in today's dollars it's about 2x its original budget and that includes a decade of operations - if that remains true then it will come in at 50% of the LTD cost of Hubble. _And_ $300million of that overrun was caused by foot dragging outside of the project over the decision to use the Ariane-5 launcher. The recent schedule re-formulation exercise seems to indicate they are pretty serious about keeping things tight now that they know how to do this thing.

My comment about JWST was a bit unfair in comparing it to GPB. I do support its broad science objectives. That it is a such a mega-project makes it a target the comes up first when massive budget overruns and schedule delays are discussed.

Don Merritt (cndwrld) described well the underestimating that goes on in project proposals. I hope the Dawn experience give future project hopefuls some caution to include more detailed analysis of development and operational costs in their estimate of how much money the project requires for mission success.

One other recent lesson was in the descoping of NH and Dawn. Both lost their magnetometers to try to stay under cost caps. The tactic of adding an instrument to gain wider support with the knowledge that it could be dropped if total project costs start going over budget should be done with integrity. If an instrument is really required for mission success, then that instrument cannot be dropped without risking cancellation of the whole project. One possibility here is for inclusion of an aspirational category. It gets a line in the budget. The instrument team knows the risk when they sign on. Depending on interteam dynamics, it could be an incentive for all teams to work harder to stay within costs so everybody makes it to the gate.

[Littlebit]

An even broader upscaling of cost is the distribution of workload amoung congressional districts to bring on congressional support. Frankly, I think this is acceptable, but everyone needs to understand that the cost of space missions is driven by the need for broad congressional support - the situation is not unlike the requirements the ESA runs into to integrate components from many nations.

[dvandorn]

You point out a key factor in understanding how these things happen, Littlebit. People who don't care to evaluate the relative desirability of various funding programs refer to the process simply as "pork" -- money distributed amongst various congressional districts in order to gain broad support for the funding.

In reality, you're often talking about pumping millions, if not billions, of dollars into the private sector to build, fly and manage these spacecraft. It makes sense to distribute that money as widely around the country as possible, to avoid giving all of the economic benefits to a small segment of the population. A majority of the dollars spent on a spacecraft generally goes into labor costs, and that money gets spent in the communities where the workforce lives. So, even though you're always going to be giving the money to the same types of workers -- engineers, scientists, factory workers, etc. -- if you gave all of the contracts to a few companies in southern California, that money wouldn't circulate as widely, and provide economic support for as many communities, as if you spread it around to contractors all over the country.

That's one reason why the Shuttle is so expensive to fly -- it serves not only as a space transportation system, but also as a mechanism for distributing federal funds into the private sector via an overly-large (for the task) workforce. While this does nothing to reduce the costs of getting into LEO, it does provide jobs and pump money into the communities where Shuttle processing and management facilities are located.

So, while this type of funding process is looked down upon by purists who believe that "pork" is always a bad thing, it actually has some sound economic reasoning behind it.

Oh, and BTW -- there are other factors involved, too, not the least of which are the inter-Center rivalries and jockeying for new project funding amongst all of the NASA Centers. APL winning the management of New Horizons was a major coup, since most every other planetary probe flown by the U.S. has been managed out of JPL. It's a good precedent, encouraging each Center to work lean and mean and out-bid the other Centers for new projects.

-the other Doug

[nprev]

APL winning the management of New Horizons was a major coup, since most every other planetary probe flown by the U.S. has been managed out of JPL. It's a good precedent, encouraging each Center to work lean and mean and out-bid the other Centers for new projects.

-the other Doug

Well said, Doug. I think that competition to fly Discovery-class missions is indeed healthy, and fosters innovative mission concepts as well as cost conservation. Perhaps it would be wise for NASA to increase Discovery program funding and thereby get more missions flown by more people.

[Littlebit]

Well said, Doug. I think that competition to fly Discovery-class missions is indeed healthy, and fosters innovative mission concepts as well as cost conservation. Perhaps it would be wise for NASA to increase Discovery program funding and thereby get more missions flown by more people.

There are other downsides: The instability in aerospace, combined with the earning potential in other areas has driven many potential engineers into other fields. In the US, while other professions have shattered the 100K$ salary ceiling; space-related engineering salaries have doddled. The best and brightest have more income potential and job stability in urban planning, accounting and pharmacology than space engineering. Factor in the inflated CEO skim, and it is difficult to woo enough capable engineers into industry centers featuring short term projects. The lose of NASA's glitter only makes matters worse: It is going to take serious money to keep exploring space.

[dvandorn]

You make some truly excellent points, Littlebit. It seems that the one area in which aerospace holds a greater appeal than other, more highly compensated and stable engineering disciplines is in its intrinsic appeal to sense of wonder. It's "cool" and "sexy" to work on spacecraft systems. For some engineers, this is enough to lure them into aerospace engineering. For others, it just doesn't outweigh the disparities in compensation and job security.

Y'all have to remember that in the late 1950s through the late 1960s, America so highly encouraged its best and brightest to specialize in aerospace engineering that we created a crop of people who were literally able to put men on the Moon. And then we thanked them, gave a few of them some medals, took most of their jobs away, and said "We don't need what you do anymore -- go find something worthwhile to do."

The engineering community hasn't forgotten the massive layoffs of the late 1960s into the early 1970s, and I can't blame them.

So -- what do we do to make aerospace engineering a well-compensated and nationally critical discipline once again? As much as people don't want to hear it, I think the first step is that more money needs to be spent on aerospace projects, and those projects need to have enough committment from the people and the government to make the discipline seem worthwhile to those who are making their career decisions today.

How do we accomplish that? I wish I knew.

-the other Doug

[gndonald]

You need to have a rapid-fire series of missions to take advantage of volume efficiencies, and it's just not reasonable to send out a probe every six months to a year, as we did during the Golden Age of the 1960s. Back then, if you wanted to send a probe to Venus, well, just borrow a Ranger spacecraft body, adjust its instrumentation, and send it on its way. Build a dozen octagonal spacecraft busses and then outfit them for the mission at hand, a la Mariners 3 through 9.

Slightly off topic...

In the post-Apollo period some consideration was given to taking the hardware developed for Viking and modifying it for lunar use, a sound, low risk concept, which would probably have had some major scientific payoffs. For example the Viking landers were fully the equivalent of the Block II Surveyors. But for whatever reason, probably a case of "we've already photographed it/sampled it, so why do it again", the idea does not seem to have been carried beyond the proposal stage.

I still think that NASA missed a wonderful opportunity when they elected not to replicate the Pathfinder mission to Mars, just think of the science return we could have had from a series of simple landers sent to Mars at every launch opportunity in addition to the high complexity probes such as the MRO...

It's interesting to note that the latest MEPAG report is recommending a Viking style Orbiter/Lander mission for the 2013 launch opportunity.

Hopefully, Alan Stern will find a way to send two such probes.

[Littlebit]

While I am on a role, there are two more major cultural impacts:

1) End of the cold war.

2) Video Games and Big trucks.

In the sixties and seventies, there was a very real war mentality in space techology: We had to catch up to, and then surge ahead of the Soviets before they quite literally conquered us on earth as well as in space. We were frightened, and would work all kinds of ungodly shifts and hours in the committment to be first to the moon. (I have heard war stories about parts being smuggled out of Detroit in campers shells during a Chicago trucking strike.)

It was much easier for a manager to ask an engineering team to put in extended shifts, and easier for spouses to tolerate it, when the fate of the world was at stake.

Which is oddly, just as true today as it was then, it is only the urgency that is less.

It was also true that a computer nerd only had access to computers at work. I have a nephew who plays computer games for a living - providing online hints for game addicts. (My first experience with computer games was a lunar lander game, written in Fortran and programmed into a $150,000 first-generation FTIR.) I loved to get stuck late into the night, just waiting for a sample...

[monitorlizard]

Maybe space science could benefit from the Air Force example, where constant technology upgrading is required, but they can't afford to design a new aircraft every couple of years. So they use block upgrades,
where nearly every system is modular and new technology is introduced by more-or-less plug-in modules that fit in the same volume as what they replace. Then you only need a new aircraft design every twenty or thirty years. A few different aircraft for different missions (fighters, bombers, intelligence gathering)--a few different spacecraft designs for different missions (orbiters, landers, rovers).

I know NASA is doing some of this (AFL rover in 2016 should be very similar to MSL in all but payload), but I think it could do more standardizing of designs, and block upgrades.

As an aside, I heard once that up to 25% of the Apollo program work was done by people donating overtime with no pay, such was the patriotism/nationalism/Cold Warism of the 1960's. When I worked as an intern at the USGS Astrogeology Branch in 1980, all new employees were told at the orientation meeting that it was illegal to work overtime without pay. All of the secretaries scratched their heads, saying "why would anybody do that?", while all the scientists were laughing their heads off, like "go ahead, try to stop me." Still, it's not something you can depend on to keep costs down, but I think a lot of you know what I mean.

[nprev]

Still, it's not something you can depend on to keep costs down, but I think a lot of you know what I mean.

Yeah...love what you do, do what you love...that's saved more projects then I think we really want to know...

Definitely not something that can be depended upon, though, especially nowadays. The Baby Boomer retirement crunch is upon us, and up to 50% of the current S&T workforce (at least in the US government) will probably be gone by 2012. GenXers and later cohorts have a different orientation with respect to work; it's a means, not an end, and this is an artifact of not growing up in economically difficult times when a job was something to achieve and keep.

There will be exceptions, of course, and a certain amount of heroism is always needed to make projects succeed. In order to get that, though, there will have to be good leadership and management combined with inherently interesting problems to solve. Things are changing; heads up, and be adaptive.