NASA Decides To Reduce Cost/Complexity of WFIRST
NASA Internal Memo: Next Steps for WFIRST Program
“I have reviewed the findings of the independent review team and have accepted them. As a result, I believe reductions in scope and complexity are needed. I am directing the Goddard Space Flight Center to study modifying the current WFIRST design, the design that was reviewed by the WIETR, to reduce cost and complexity sufficient to have a cost estimate consistent with the $3.2B cost target set at the beginning of Phase A.”
Boo! No reductions to the coronagraph instrument, I want to see direct images of planets!
Fomalhaut b (a.k.a. Dagon) doesn’t count? Ok, a jovian planet at 110 AU from its star, that probably isn’t the sort of planet you’re thinking of. It would be nice to see (and get spectra of) ice giants as close as 1 AU, and WFIRST’s coronagraph was supposed to do that out to 10 pc or so (~30 light years.) I don’t think it could quite see terrestrial planet, if I’m reading the project’s web page correctly.
But the mission isn’t even two years out from its formal start, and it’s already $400 million over its $3.2 billion target budget. That’s more than half the cost of the entire Kepler mission. That’s also a very bad sign, since the launch date is still half a decade in the future. Descoping a mission is painful (trust me, I know what the 1992 descope did to Cassini), but letting cost overruns grow without limit and eat up the rest of the program is worse. Sometimes a cost cap, and a real, enforced one, is the better choice overall.
This is disheartening news so early in the mission lifecycle. Taking a quick look, it seems the mission is only in late Phase A. Already $400M over tells me they had no idea how much building the craft would cost when it was proposed. This does not bode well for the remainder of the phases, even with an early descope.
Well, a 12% cost growth so early in a project is very disturbing. But I am more optimistic than you are. I mentioned Cassini. It had a similar level of early cost growth and ended up being an extremely successful mission. I think the important thing is to make the cost cap and the intended goals clear and firm.
That is what Dr. Zurbuchen seems to be trying to do. I did spend a few years working alongside him, when we were both at the University of Michigan, and I can say he is practical and interested in making things work. So I am optimistic about WFIRST.
One can only hope. With a 12% overrun this early in the game, it will be interesting to see what the nature of the proposed descope will look like. It seems to me that given this is a telescope, their options are fairly limited.
This is the ‘post JWST effect’, i.e. can’t have multi-billion dollar extravaganza flag ship missions. Keep cost creep down early.
It’s not quite that extreme. After all, the target budget is $3.2 billion. So a multi-billion dollar flagship mission is acceptable. Of course, if you look it up in a dictionary, a flagship is only part of a fleet. I think that’s the real impact of the JWST budget overruns. It hurt the smaller astrophysics missions, like the Explorer line. That’s down to two in the previous decade, compared to eight in the decade before that (if I counted correctly.)
Field of view is said to be 100 times that of the Hubble; does this mean ten times the diameter? Is this by use of a different mirror curvature, or something else?
Both WFIRST (as planned) and HST have the same diameter mirrors – the difference in field of view comes from the fact that WFIRST has a focal length about 1/3 that of HST.
Another way to think of it is that your eye has a immensely higher field of view that HST (or WFIRST), but also a much smaller diameter.
I’m fairly sure that “100 times” is more-or-less made up. It’s about right, but each of the instruments on Hubble has a different field of view. So is the comparison for HST’s Advanced Camera for Survey, Space Telescope Imaging Spectrometer, Wide Field Camera 3 or what? But for comparison, WFIRST is supposed to have a 0.28 deg wide field of view, and HST/WFC3 is 0.045 deg. (And I’m still annoyed that they dropped the P, for planetary, in the third-generation instrument… WFPC, WFPC2, WFC3.)
In a real sense, this mission – and others like it – may soon become pointless. If the plans of Musk and Bezos come to fruition then there’ll be no need for gold-plated money-pit missions, as the launch cost (and environment) will be both cheap and benign. Some things like mirrors and instruments will remain expensive, obviously, but we should start to see missions which are merely multiples of their terrestrial counterparts in terms of cost. I realise that cheap, easy access to space hasn’t quite arrived yet, but I’d far rather see the likes of JWST delivered to an L-Point by a BFR flight and checked out in situ by a delivery team than trust over-complicated space origami. And, don’t start me on He replenishment!
Well, actually SpaceX is used now as the baseline launch provider for missions like this. The launch costs aren’t the dominant driver here, it’s building a complex machine that can survive in the space environment. AFAIK WFIRST isn’t headed to L2, but going out there also really blows up the cost. Even ground-based observatories (big facilities, anyway) are headed into billion dollar territory.
Is it planned for GEO? Or some other orbit?
The baseline orbit for WFIRST is L2. The most recent WFIRST mission parameters can be found at https://wfirst.gsfc.nasa.go…, at the “WFIRST Observatory Reference Information” link therein.
I was confused by that as well. When I was last paying attention, WFIRST was slated for L2. But then I ran into this:
https://wfirst.gsfc.nasa.go…
“The current WFIRST design assumes the observatory will operate from geosynchronous orbit.”
I will admit my own web site is a mess, so I assume there are some out of sync web pages. I’m not sure if this “serviceabilty” requirement means manned or robotic, but if manned I assume this would tie into being at GEO.
The page you mention is probably out of date. Go back to the website I reference, and at the bottom you will see archives of previous design cycles. The orbit was changed from GEO in cycle 5 (July 2015) to L2 in cycle 6 (April 2016).
The launch costs aren’t necessarily the issue, so much as the specific cost (cost per kilo) and payload shroud. JWST would be much less expensive if the mirror didn’t have to fold up before launch, or (if it still had to fold up) if there were astronauts around to make sure if unfolded properly. If launch mass weren’t such a design driver, there would easier (and cheaper) solutions to surviving and operating in the space environment. I can also imagine cheaper designs because they were less ambitious. Not every telescope needs to be a “great observatory.” Arguably, a series of telescopes, each a factor of a few improvement over the last, would be more cost-effective than a single one going for an order of magnitude improvement.
Well, a large part of the budget creep can be traced straight back to the “gift” of the NRO mirror. Oops – sorry, that’s a “national asset” transferred from “another agency”. The whole cloak and dagger thing is stupid, given that we all knew from the beginning what it was and where it came from (and yes, I worked on this).
But anyway, the larger mirror enabled a lot of additional science. Science you really needed to in order justify the project given the existence of Euclid. But that science added complexity in operations and the instrument package. So really, the simple fact is that the project was underscoped in terms of enabling science when it was pushed by the decadal survey, which is why many of us (and yes, I worked on the JDEM and SNAP predecessors) had the reaction of “say what?” when the decadal survey was revealed. When it grew into something more palatable from a science perspective, it naturally got more expensive.
Thanks for a clear explanation for the high cost.It sounds like the main reason for the cost growth was _not_ that the project itself was poorly managed but that the original cost estimate was inaccurate based on the stated goals. It’s a common problem for NASA since realistic cost projections tend to reduce the chance of a project being selected.
I’m not sure of that. The $3.2 billion cost estimate was from the start of phase A. Wasn’t that in February, 2016? That was after the decision to use the NRO 2.4 m mirror. I have no doubt about what you’re describing, but I think that’s the cost growth from the $2 billion considered by the decadal survey and the $3.2 billion at the start of phase A. This recent decision seems to be about $3.2 billion growing to $3.6 billion since then.
yes, but if you descope the project for the original budget, then you’re aren’t meeting the budget … you’re just not asking for more money and delivering less.
NASA project, and this probably applies to all projects, budgets are little more than guesses. Each project has an original component for which we can’t estimate the cost with much certainty. So each project has an associated risk and an equivalent contingency budget. Companies bidding on firm, fixed pricing include some risk factor … sometimes they make money, more often they lose it (in development, but gain it back in production). NASA could budget jobs with Congress controlling the contingency budget, or budget as firm fixed and either return the under-run profit to Congress or retain it for other project over-runs.
Missions have budget reserves at all levels. But they almost always get spent. Not because things always cost more than expected, but because you can always put more into risk reduction or enhancing capabilities. Many places consider it poor management _not_ to spend reserves down to zero. I’ve even heard half-good justifications for that. The money does not, in general, get returned to headquarters for use on a different project.
I see now that Euclid — a mission perhaps as centrally important to our understanding of the universe as COBE — is having similar instrument problems that will delay the project at least a year.
Not very long when compared to 14B years, but too long when you are 67 years old.