JWST: Albatross of SMD now $8.8 billion

[as58]

Would it be possible to put a wide but flattened like aeroshell on a falcon heavy?  Ie a fairing that would allow an 8-10m diameter monolithic mirror to be launched in a side-on configuration for assembly into a big telescope in orbit.

The availability of heavy lift (assuming falcon heavy in 2013) will massively change the most economic method for achieving the the performance capabilities of JWST.  Given JWST's rising 8.7billion price it is time to down tools and think again.  With all the problems that L2 imposes it might be preferable and cheaper to go for a LEO telescope that can be serviced to extend life into decades, and utilise stored (and replenishable) cryogens or cryocoolers for cooling. 

Building an 8-10 metre actively cooled telescope would be very expensive and radiative cooling in LEO would require heroic efforts in thermal shielding. It's not exactly easy even at L2... And what problems does L2 entail, anyway?

Furthermore, a monolithic mirror does not mean easy and inexpensive. I think JWST's individual segments are the largest cryogenic mirrors ever built after Herschel's primary (and Herschel has much less stringent requirements due to its longer operating wavelengths).

[Jim]

Would it be possible to put a wide but flattened like aeroshell on a falcon heavy?  Ie a fairing that would allow an 8-10m diameter monolithic mirror to be launched in a side-on configuration for assembly into a big telescope in orbit.

The availability of heavy lift (assuming falcon heavy in 2013) will massively change the most economic method for achieving the the performance capabilities of JWST.  Given JWST's rising 8.7billion price it is time to down tools and think again.  With all the problems that L2 imposes it might be preferable and cheaper to go for a LEO telescope that can be serviced to extend life into decades, and utilise stored (and replenishable) cryogens or cryocoolers for cooling. 

Too much assuming for heavy lift availability and bad assumption that it will save costs.

No, it is not cheaper to go to LEO.  LEO is the worst place for telescopes.   HST was grossly compromised to work in LEO and to be serviceable.

[baldusi]

Is there a detail of which parts of the program got overbudget? Because if it was 100% through all parts, it would have been a management problem. There's a lot of discussion here between optics, instruments and space bus.

[oldAtlas_Eguy]

Is there a detail of which parts of the program got overbudget? Because if it was 100% through all parts, it would have been a management problem. There's a lot of discussion here between optics, instruments and space bus.

Just generic NASA mismanagement of all elements of the program, mostly because of:

1) Too much NASA involvement with the day-today engineering decisions and
2) not having the NASA staff to perform #1 in a timely manner.

This resulted in the contractor teams to sit idle a lot wasting time and money, sometimes even discarding hardware for a contractor implemented solution which would have been OK but was not approved by NASA engineering.

[joek]

Is there a detail of which parts of the program got overbudget? Because if it was 100% through all parts, it would have been a management problem. There's a lot of discussion here between optics, instruments and space bus.

Just generic NASA mismanagement of all elements of the program, mostly because of ...

This went way beyond generic mismanagement (or subcontractor management issues) with the most serious problems at high levels of management, with ill-formed, inaccurate and incomplete budget and schedule from the very start, and going down hill year after year from there; see:
James Webb Space Telescope (JWST) Independent Comprehensive Review Panel Final Report, Oct 2010

[oldAtlas_Eguy]

Is there a detail of which parts of the program got overbudget? Because if it was 100% through all parts, it would have been a management problem. There's a lot of discussion here between optics, instruments and space bus.

Just generic NASA mismanagement of all elements of the program, mostly because of ...

This went way beyond generic mismanagement (or subcontractor management issues) with the most serious problems at high levels of management, with ill-formed, inaccurate and incomplete budget and schedule from the very start, and going down hill year after year from there; see:
James Webb Space Telescope (JWST) Independent Comprehensive Review Panel Final Report, Oct 2010

Ok, so a little micromanagement problem causing a slight cost growth and delay then because of insuficient reserves caused further delays to manage current year costs to meet budget caps caused even more cost growth that caused even more delays that caused even more cost growth ...

Suddenly you have a program that cost 3X or more what it should have cost.

Uh. The SLS program seems to be headed this same direction unless they significantly overbudget for the schedule they start with.

[baldusi]

Could something smaller, less ambitious, like a 4m monolithic mirror have been a good compromise, with less technological risk? And then do something like JWST as the next generation craft, after having mastered and used at something like half the critical technologies?

There are stories of alernatives that were less ambitious and why they were not selected.  I do not have reliable reference for any of them.  However, a less ambitious project would cost less.  ESA's  Herschel is currently at L2 iwth a 3.5 m primary mirror.  It certainly cost less thand $2B.

I know that measuring just by the primary mirror size is an over simplification. As I understand it, there had never been a design with a shade. Which implied to maintain cryogenic temperatures all through the mirror truss. Also, that the level of wavefront control over the mirror was beyond anything (non classified) ever developed. Ditto for the sensors. Etc. So I would love to see a summary of all the technological breakthroughs that they expected to make. Which also brings me to a second thought. What is the usual level of over subscription for this type of telescope? What would be the cost of a second copy? If the project is successful,  and since it has a five year life, could a second copy be manufactured with minimum improvements only on the sensors? Could it still offer some useful science once it's depleted of helium?
Could a second unit be build the same, but an occulter added?
The way I see it, it's not only the amount of breakthroughs that they took on one step, but also if this will make cheaper and safer projects in the foreseeable future. So a similar version could be made on the Hubble spectrum with minimum investment? Could an even bigger but fundamentally same technology version also be done? What about add-ons, like an occulter?
Or will the next step also be such a jump?

[jgoldader]

The science goals of JWST concentrated on the IR, a region where you need cryogenic temperatures.  Discovering and measuring galaxies at high redshifts (z>~10, say) where all the spectral features like the Lyman break are way past 1 micron wavelength, simply requires a big, cold telescope.

The decision was made early in the program to choose a mirror coating that would optimize the IR properties of the telescope.  This meant that UV observations weren't feasible and even optical observations would be iffy.  I remember that in the early stages, there was hope of imaging down into the red optical bands, but am pretty sure that was lost in one of the first cost-overrun descoping rounds.

There were options; but a big, IR telescope was the big winner of the decadal survey.  The eventual loss of HST is going to be a real problem for UV astronomy.  You can't do real UV work from the ground (not at 100-300 nm wavelength, at least!).

I know that Johns Hopkins was pushing a HST-derived telescope, HOP, during the time right after O'Keefe cancelled the last SM, which of course ended up being uncancelled. HOP would have used the instruments meant for the final servicing mission in a non-serviceable telescope built around the basic HST design.

I always thought that was a pretty cool idea.

Building a telescope safe enough for servicing in space was very expensive.  Furthermore, maintaining the capability to mount servicing missions was also very expensive.  That's part of why HST cost so much.

Jeff

[Proponent]

And if memory serves, $1.5 billion is what it cost in the mid-1980s to build one Shuttle orbiter after all the development had been paid for and spares were available.  The cost of building one Shuttle orbiter from scratch now would be much, much greater.

I'm sure that's true as well, but this is a quibble over terms, right?  Or are you also saying that $8.7B is an ok number?

I'm just saying that the figure of $1.5 billion isn't a relevant comparison, and a comparison of the costs of JWST now and an additional Shuttle in the 1980s is not grounds for believing that NASA has deteriorated (it may have, but these numbers by themselves don't show it).  As for $8.7 billion, it's definitely not OK, because it's many times what the program was supposed to cost, and because at $8.7 billion (which is still just an estimate) it's a bad allocation of resources.

Of course, the $8.7-billion figure is now largely irrelevant; what matters is the additional cost of completing JWST; it is that cost which must be weighed in choosing the best course. 

Of

[JohnFornaro]

In case nobody guessed, there was a reason I asked the leading question.  Obviously, I'm trying to figure out is what do people think?  when is enough enough?

When do you give up? $8.7B is yet another estimate, and there is no apparent dictate to accomplish something.

I don't get it.

[Lee Jay]

In case nobody guessed, there was a reason I asked the leading question.  Obviously, I'm trying to figure out is what do people think?  when is enough enough?

When do you give up? $8.7B is yet another estimate, and there is no apparent dictate to accomplish something.

I don't get it.

HST cost $1.5B not including launch.  That's about $3B in today's dollars.  Then it cost probably another $1B to fix for a total of about $5B in today's dollars.  HST was based on an already-developed and deployed spy satellite.

JWST is bigger, much more complex and difficult to implement, and not based on anything in existence.  Is there some reason for us to believe it should be cheaper to build than HST was?

By the way, JWST's cost is comparable to the Large Hadron Collider.  Considering they are working on the same type of basic-physics problems, and may contribute in comparable ways to that endeavor, I'm personally fine with spending my tax money to do this.  There are plenty of things the government spends orders of magnitude more on that I think are ultimately much less useful.

[Proponent]

I lean toward the view that JWST should be canceled, for three reasons.  One is that $8.7 billion to completion is just another estimate; things could turn out worse than that.

Secondly, even at $8.7 billion, it's still going to take years to finish JWST.  Given the fiscal outlook, there are going to be periodic rounds of budget cutting in the next few years, and JWST will have to survive all of these to make it to the launch pad.  Particularly given its very troubled history, it's quite likely to be killed eventually.  Better to cut losses now.

Finally, maybe every once in a while we need an execution to send a message about cost control.

But I'm open to persuasion.

[as58]

The decision was made early in the program to choose a mirror coating that would optimize the IR properties of the telescope.  This meant that UV observations weren't feasible and even optical observations would be iffy.  I remember that in the early stages, there was hope of imaging down into the red optical bands, but am pretty sure that was lost in one of the first cost-overrun descoping rounds.

JWST is still planned to have imaging down to 0.6 um, but it's diffraction limited only down to 2 um. I think some testing at the shortest wavelengths has also been eliminated due to cost.

[robertross]

In case nobody guessed, there was a reason I asked the leading question.  Obviously, I'm trying to figure out is what do people think?  when is enough enough?

When do you give up? $8.7B is yet another estimate, and there is no apparent dictate to accomplish something.

I don't get it.

HST cost $1.5B not including launch.  That's about $3B in today's dollars.  Then it cost probably another $1B to fix for a total of about $5B in today's dollars.  HST was based on an already-developed and deployed spy satellite.

JWST is bigger, much more complex and difficult to implement, and not based on anything in existence.  Is there some reason for us to believe it should be cheaper to build than HST was?

By the way, JWST's cost is comparable to the Large Hadron Collider.  Considering they are working on the same type of basic-physics problems, and may contribute in comparable ways to that endeavor, I'm personally fine with spending my tax money to do this.  There are plenty of things the government spends orders of magnitude more on that I think are ultimately much less useful.

That would be in line with my type of response. However, I am concerned about the process that got it into this mess, and despite the change of management, these types of issues remain in place for other programs to fall into the same pitfalls.

Another issue is leading edge technology development. It has the same applicability to the issues surrounding the future 'plan' for NASA & destinations: what are the technology requirements. At least in this case, the requirements were set and off they went to develop what was required. That cost $xB. If the alternative is to have the technology developed, then look for an application to use it on, it STILL cost you $xB (or around that amount, maybe more, maybe less) but under a different line item. In fact you have to wonder if it would ever have been developed at all based on the cost.

What we should be doing is looking at what have we gained in technology development because of this project, where can it be applied, so that it may help offset (at least symbolically) the cost of the program.

Again, thinking of all the military spy sats or that 'need' in the Mercury/Gemini/Apollo programs to develop faster and better technology got us to where we are today in a hurry. Was that money a sunk cost considering how it helped develop key computer technology? We owe a lot to that investment all those years ago. Our problem is we lost sight of that.

In the end, putting key technology development in the critical path of hard & fast milestones was the failure, which started all the way back at the proposal stage of this program.

[oldAtlas_Eguy]

The usual case is that end user missions is perferend over infrastructure development. Without a good infrastructure the end user missions are much more expensive than they have to be. Somewhere there is a optimal balance point between the two. Historically we have been more inclined to fund missions over infrastructure. STS was an aberation in that it was mostly a infrastructure investment that enabled a wide range of LEO missions, it was not a mission specific implementation but a generic Manned and cargo LV. Its intent was correct its execution of the intent was not.

[Lee Jay]

... I am concerned about the process that got it into this mess, and despite the change of management, these types of issues remain in place for other programs to fall into the same pitfalls.

The process that got us in this mess is, in my opinion, that this sort of thing is very expensive.  If you tell the truth about that up front, you don't get funded because it's too expensive.  If you go over too far, you get canceled.  So you have to pick just the right amount to lie to everyone or these sorts of things don't happen.  To fix that, the holders of the purse strings would have to recognize the expense and value of big, difficult science experiments and they just don't, probably because most of them are lawyers and business types, with few having any real scientific training or experience.  And I don't see a practical way to fix that.

[joek]

I am concerned about the process that got it into this mess, and despite the change of management, these types of issues remain in place for other programs to fall into the same pitfalls.

Not to put too fine a point on it, but the early JWST program management appear to have taken out their shovels and picks and happily dug those pits.

To fix that, the holders of the purse strings would have to recognize the expense and value of big, difficult science experiments and they just don't, probably because most of them are lawyers and business types, with few having any real scientific training or experience.  And I don't see a practical way to fix that.

There's purse strings at the macro level and then there's purse strings at the NASA level.  In this case the Science Directorate and astrophysics made a choice to divvy up the funds and give $X to JWST, and $Y to other programs.  Most all the people at those levels have significant scientific training and experience.

Yes, these programs can be difficult and the key technology development can involve significant risks. Virtually all of these types of programs have them, and JWST is singularly distinguished by their demonstrated mismanagement of those risks.  Even if imperfect, NASA has methods for dealing with such issues, including risk prioritization, budget and schedule reserves.

JWST management did not follow those methods.  The schedule and budget used for the confirmation to proceed was flawed.  JWST management then compounded the problem by managing to annual and incorrectly projected funding caps, which pushed work, schedule, and cost to the right.

JWST managed to fly under the radar for a while, but eventually it caught up with them.  (That they were able to do so is another management failure.)  The last estimate¹ put the cost to get the program back on track with a minimum cost-to-launch at ~$5.9B with a life-cycle cost of ~$6.5B.²

That estimate assumed a significant bump in funding (esp. FY11-FY12) to make up for past sins and with a 2015 launch.  Each year's slip was estimated to add $300M to the program cost. That bump in funding was not and does not appear to be forthcoming.

In short, the major reason for the ballooning budget and schedule has been due to mismanagement.  We can only speculate as to what might have been with proper management.  JWST may never have been confirmed with a more realistic budget and schedule.  Maybe it would have been canceled sooner.  Maybe we would still be looking forward to a 2014 launch and a lower cost comparable to Hubble.<sup>3


1</sup> The "Casani Report", Mar 2010 (previous post).  As far as I know, it is the last publicly available estimate.
² Cost-to-launch includes ~$50M for initial six-month commissioning.  Life-cycle costs include ~$600M or ~$125M/yr for phase E science operations over 5 years.
³ The JWST life-cycle baseline FY09 $4.9B; FY10 $5.1B.  Hubble cost-to-launch was $4.7B and cost through mirror fix was $5.8B (constant FY10 dollars).

edit: add link; clarify lower/Hubble cost.

[RocketEconomist327]

Nice Work Joe!

VR
RE327

[spectre9]

I don't think it's worth having any unserviceable telescope at L2 worth billions and billions.

Just a disaster waiting to happen.

We're just not in the day and age where we can build something so complex and yet so autonomous.

[RocketEconomist327]

People do not realize just how bad this is.  Poor planetary science.  No it isn't getting canceled but its flagship is on hold.

Bow down to that which is JWST.

VR
RE327