Hacker News Reader

I suggest the book "Failure is not an option" by Gene Kranz, an Apollo flight director (played by Ed Harris in Apollo 13). He describes how the primary work of flight controllers in all missions is risk management. You are constantly balancing mission needs, fuel needs, mass needs, temperature needs, and etc etc.

I don't think that a raw metric of the number of SPOF is the right way to measure the risk of this spacecraft. It's a fun term for PR purposes (and emphasizing the risk here) but the actual risk posture is more complex.

I imagine that in the course of developing this, they worked out a possible strategy without all of those SPOF - but doing so doesn't eliminate the risk, and the impact to mission is likely massive.

What I don't understand is for $11 billion - can't we do 3 $3B space telescopes that work reasonably well? Spread risk? More science?

Let's say a Falcon 9 launch is $90M. Falcon heavy let's say $200M.

So you take your 3x $3B. Put $200M/instrument into launch, have $2.8B per telescope leftover.

There just seems to be something wrong that it costs THIS much to build a telescope.

That said, the Thirty Meter Telescope is also a sort of "forever" job, the delays have stretched on and on.

I wonder if you did something like bid out and paid just on performance instead of this forever cost reimbursement thing. Right now if you can get onto one of these mega projects, and can stretch it out with delays, it basically can cover your career (ie, 20 year projects).

I'd bet that the vast majority of that cost isn't going to be materials -- it'll be staff time to design and optimise the telescope and make something that can work, including where necessary how to make new materials or processes to make that telescope. I'm not saying the actual hardware is cheap by any stretch of the imagination, but making three different telescopes isn't a linear function of that budget. They're literally pushing the envelope of what's possible here. If something terrible happens to the rocket, lots of forms will be written and people will be sad but fundamentally I think they'll build something a bit better on a few years and nail it. A bit like New Horizons (awesome Mars rover) vs Beagle 2 (awesome Mars rover that died on arrival to Mars).

The seemingly costly part of JWST - The Gold Mirrors are cheaper than one might guess, for eg. it’s a 0.1 micron coating that only used 45g of gold.

The material cost for a lot of "high technology" things can be very small, but the process is expensive. For example fancy chips are only some x grams of silicon, and expensive inverted mesa quartz crystals might cost a good bit while only being x microns thick piece of quartz..

JWST was (re)launched in 2005, the very idea of cheap spaceflight didn't exist back then, so you'd have ended up with three expensive flights. Also consider that the expected cost back then was $4.5 billion.

Perhaps that is what will happen from now on. But when JWST was initiated, the options looked different.

JWST is driven by the size of the mirror, the need for an IR instrument and the temperature you need to keep the instrument at to do the observations. Spamming a bunch of hubble-sized instruments up into orbit won't accomplish the same thing. And I don't think you can do space VLBI in the optical/IR or it would have been done already (but I don't know why?).

It is kind of like asking why 5 Ford Rangers can't replace one Lamborghini or something.

Spamming a bunch of hubble-sized instruments up into orbit won't accomplish the same thing.

It will, and then some, once we get optical interferometry nailed down. JWST is great, and needed to be done, and I'm glad it's finally getting deployed. But if I had an argument to make against it, it would be, "Let's wait until we know how to do this properly. We're not there yet."

That's a weak argument and should almost never be heeded, but it's also not wrong.

The reason we can't achieve the same resolution as JWST with interferometry has been mentioned in a recent startalk podcast. The distance between the telescopes would have to be coordinated to a precision somewhere on the order of nanometers.

The distance between the telescopes would have to be coordinated to a precision somewhere on the order of nanometers.

That'll happen. It's fundamentally a timing problem, and we can already build clocks that will tell you what floor of the building they're on.

Less hand-wavingly, it's fundamentally a data-acquisition and correlation problem of the sort that was solved long ago for microwave VLBI. Back in the day, the individual stations had their own maser clocks, but now I imagine they're all GNSS-based. It is by no means trivial to go from RF interferometry to optical, or to move the antenna elements from earth to space, but that's the basic approach that will ultimately be used.

> That's a weak argument and should almost never be heeded

Well that's an incredibly strong opinion with almost nothing to back it up.

Maybe it all goes boom tomorrow on launch or the deployment is ultimately completely FUBAR'd and you can feel satisfied with yourself, but I don't buy that your principle is what we should always follow.

Hopefully we start getting data from JWST in a few months and then I invite the scientific community to figure out how to spamcraft optical and IR instruments into LEO and achieve VLBI with them in the future, but we won't have to keep holding our breath for the technical breakthroughs there.

And even as a jobs program and Keynesian stimulus the JWST beats digging holes and filling them up again, even if it all goes boom. Better than building yet another weapons program as well. And the Senate just tacked on an extra $23B to the 2022 Defense Appropriations Act for one year of spending. The $10B that JWST cost over 25 years of the program doesn't really matter at all.

True, I don't disagree with any of that, but:

Maybe it all goes boom tomorrow on launch or the deployment is ultimately completely FUBAR'd and you can feel satisfied with yourself

A more charitable interpretation of my post would read something like this: "If it does go boom, we can either start building a new one now and try again in 20 years, or we can wait 15 years and build multiple better, cheaper ones in 5 more. Meanwhile, maybe a few hundred other researchers can get funded with the same money."

The reason why we are barely getting started into VLBI in the IR on the ground (and nothing that I have heard of yet in space) is that the different apertures need to be stable relative to each other with a precision better than a small fraction of a wavelength. (One tenth, one twelves and one twenties are often used, depending basically on which performance drop relative to the theoretical optimum you are willing to live with.)

For radio astronomy, where we do VLBI everyday, we have to handle waves of wavelength 1 cm and position antennas to a precision better than a millimeter. Not easy when the antennas are scattered across the country, but something we can pull off.

For IR astronomy we are talking wavelength in the range of 1000 nanometers to 30 microns. So at the easiest end of the spectrum you would have to position satellites to a precision better than 3 microns relative to each other, while flying on orbit and being pulled and pushed by tidal forces, gradients in the graviational fields and solar wind pressure (which contains turbulent fluctuations). For it to actually work in near IR you would have to get the positioning right to within 100nm.

For comparison: The mirrors of JWST itself are flat to within about 25nm. And in some sense we ARE doing IR VLBI with JWST since we have separate mirror segments that we all position correctly relatively to each other. But doing so we separate freeflying satellites is something we just aren't capable of yet.

PS: Yes, LISA Pathfinder has demonstrated measurements of spacecraft separation down to a few picometer, so we are slowly getting there.

That's kinda what I guessed. So on the ground you can measure the distance between two observatories that are a few hundred meters apart with incredibly high precision and track that over time as the ground under them slowly deforms and moves.

Geodesy is indeed one of the side products of VLBI observations on Earth.

There are many people in NASA and around that have made similar arguments.

I think the real reason that they never have a lot of traction, sadly, is that if you propose 3, Congress will give you 2. And then when 2 are over budget, it will get trimmed to one. Better to propose one big mission and get it to the point where it can’t be cut easily.

Yeah when you have a fickle funding source often the only way to get any long-term project done is to make it too big to fail, but then you see all the other terrible waste, corruption, etc. that accompanies that.

I think this is because of how the various constraints work against each other. We need the telescope to be as large as possible to work as well as possible. But the rockets can only be so big (i.e. we're constrained by the largest available booster). The only thing here under our control is how hard we work on fitting the biggest possible instrument in the available payload envelope, and that is exactly what happened.

With a machine this complex I think it's also not easy to crank out multiple copies since I'm imagining most of it is made by hand without the benefit of a production line.

To make an infrared telescope significantly better they need it in an orbit not facing the sun always. But other types of lower cost projects might be feasible.

> Is JWST as brittle as the news want to make us think? Or are there technical reasons or acceptable failure modes that gives confidence that those steps are not as critical as the news let us people know?

There is no room for redundancy in many aspects of the design, unlike, say a Boeing 777 or Airbus A350.

How can you have a redundant heat shield, or primary mirror (two parts of which swing)? I'm sure some computer systems have redundancy and perhaps comms.

But like with a helicopter: how can you have redundancy in the tail rotor?

So with the JWST: there's no way around many SPOFs.

Regarding that helicopter example:

https://verticalmag.com/news/bell-electrically-distributed-a...

The answer to your redundancy question is: Assembly line.

Instead of ramping up a project, and building 1 of something, you would plan to do more than one, and you could iterate over time as you learn. SpaceX is doing a good job of this.

If 1 Webb telescope is valuable then wouldn't 3 or 5 also be valuable?

We have a number of proven space designs at this point: Soyuz, Spirit/Opportunity rovers.

They calculated that. Building a second JWST would have added 10% to the budget, but the budget committee nixed that.

YOLO (you only launch once).

The JWST only has a planned mission duration of 10 years. If it’s as epic as everyone claims it could be, there must be a follow up mission planned. So even if it fails, we’ll still see a successor at some point.

There are quite a few more telescopes planned, I think the next major one is the Nancy Grace Roman telescope, which is another infrared scope, made from an old NRO telescope. The decade survey called for another massive space telescope to look at the optical and UV spectrums, so Id expect that to be the next truly big flagship telescope.

>there must be a follow up mission planned

https://www.universetoday.com/139461/what-comes-after-james-...

The write up goes into detail on how missions are planned, and what is in line to follow JWST.

Since JWST was delayed ten years the successor will probably be delayed 20 years and launch in 2050.

> They calculated that. Building a second JWST would have added 10% to the budget, but the budget committee nixed that.

Do you have a citation for that?

It's really not - just by having more hardware available (at higher total expenses) doesn't make the pool of money available (public research funding) to book time on these things more. These things are one-off, you build a new one if you expect 10x improvement over the old one.

We don't need a fleet of X1 to break the sound barrier for the first time. We do need many Airbus/Boeings to fly people and stuff from A to B.

Note that that is the case with the unique research hardware you cite as well - we're not sending another Spirit/Opportunity, but have graduated to something else.

Soyuz is a different use case, as there is an economic demand to be filled - that's why a private company like SpaceX is in that sector with its Dragon. On the other hand, you don't see SpaceX cranking out Spirits or JWSTs or Washington Monuments.

Better helicopter example is the Jesus Nut.

"Because it it comes off that’s who you’re seeing next"

Does this have a heat shield? It won't re-enter right? It's far too far away for that.

It's actually a sunshield [1]

[1] https://webb.nasa.gov/content/observatory/sunshield.html

Not the kind of heat shield you would use to re-enter.

The telescope is designed to image very faint sources of infrared light. The problem is that everything (including the telescope itself) glows in infrared. The hotter things are the more infrared they emit. Because of this you want to keep the instrument as cold as possible. (You do this because you don’t want to drown the faint sources by the glow of the telescope itself.)

Now of course there are parts which has to be “hot”. At least relatively to the very cold instruments. The solar panels are heated by the sun, the transmission electronics and the processing turns electricity into heat. The positioning thrusters burn chemicals which makes them hot.

Because of this they designed the spacecraft with two sides, a cold one for the instrument and a hot one for everything else. They even choose the orbit cleverly so they can keep the sun and the earth and the moon always on the hot side of the vehicle.

And then you have this problem that you have to make sure that the hot side won’t warm up your cold side. This is where the heat shield comes into play. Sometimes it is also called a sun shield since the sun is the main source of heat for it to shield against of course, but it also shields the instrument from the heat of the hot side equipments.

Structurally it is a 5 layer lassagne. They just replaced the pasta with metalized kapton tape and the sauce with the vacuum of space. It is about the size of a tennis court, launches folded up and will un-fold in space. Hopefully. :)

Thanks for the detailed answer! I forgot it was mainly for infrared but I didn't think of this implication anyway. Very interesting! I hope we'll see this in the coming days.

Intermeshing rotors are interesting as an alternative to having a tail rotor. They are pretty slow & stable, should feather relatively safely as a failure mode.

I think all the SPOF talk is expectation management in case it fails. It's part space telescope mission, part engineering challenge. Even if the space telescope part fails, the engineering effort that's gone into it means something.

They must've calculated that the overall chance of success, and they have a target, and they met their target. Unfortunately, tests and theoretical modelling have a tendency to not exactly replicate a space environment (or any true production environment), nobody's perfect at anticipating everything, and management has ways of manipulating engineering estimates.

The Space Review [1] quotes NASA as saying there are 344 SPOF. They talk mainly about the sun shield, so that's probably the biggest risk, but consider all of them as about equal...

If each SPOF has a 0.1% chance of failure, net success rate is only 71%. Presumably most of the estimated failure probabilities are less than that, and the sun shield—which probably comprises many of the SPOFs—averages (far?) more than 0.1% per SPOF, because everyone seems to be particularly worried about that working.

I wonder what that figure is. Has it been published anywhere? Dear NASA and ESA, what do your engineers say about overall chance of failure?

[1] https://www.thespacereview.com/article/4303/1

I hope it's just badly thought through marketing. There is currently a AWS spot during nfl games, that shows a spectacular catch and then proclaims that the catch probability is only 3.6% or something. You are meant to be somehow impressed by the unlikelihood I believe, but that their model thought the catch is unlikely and the guy caught it implies a rejection of their model with p=3.6% < 5%.

The JWST marketing seems to work under a similar premise, they proudly proclaim that they couldn't mitigate hundreds of single points of failure, and you're supposed the be impressed by how difficult their task is. Hopefully the engineering did a reasonable job and the marketing is just playing up the wrong thing.

I don't think you are supposed to be impressed by the number of things that could go wrong. At least, that's not how I see it. They talk about the deployment phase with all these complex things that have to go exactly right as "the 30 days of terror." And that's a pretty accurate description of what many people with a stake in the JWST will be feeling. Not awe, but terror. This is a big part of space missions at the bleeding edge of science and technology, and I think it's great that many people are hearing about it!

It has to fit in the fareing, survive launch stress and vibration, fly further than the moon and it has to weigh 6000 kgs +including fuel. If it works it will be one of the greatest engineering feats of history.

I'm guessing it's a case where all the various extreme requirements simply do not allow for redundancy in the places the engineers would prefer to have it. The options are likely (1) okay performance with a lower chance of failure or (2) extreme performance with a higher chance of failure.

> However, I'm always baffled how everyone seems proud that the telescope has something like 200 SPOF during deployment, and if even one of them fails the whole mission could fail.

I'm not sure I've seen anyone who is _proud_ of it, lots of people are just setting expectations. Probably due to the similarities with Hubble (although JWST can self-align it's mirrors!).

It also might be posturing to show how well the thing is built. Space is hard, like really hard, and these agencies keep knocking it out of the park.

Every kilo launched costs a proverbial and probably nearly literal tonne of fuel so things are not as simple as that.

It's also has to fit on the rocket hence the once off folding mechanism. And after deploying it has to be perfectly aligned (remember the Hubble with its slightly off mirror)

I think having redundancy for everything would just not make for a launchable spacecraft.

Unlike say, the Apollo program, which had a guaranteed successful outcome?

The thing to realize is that these are some of the hardest things humanity has tried their hand on and if it all works that's a great thing for all of us, if it fails we will learn something and we'll go back to the drawing board (but we won't have a JWST and that's a significant loss, besides the obvious future calls of 'look at what happened to JWST' which will no doubt have negative impact on finding funding for future space missions).

Also, I think you're mistaken about people being 'proud about the 200 SPOFs', if they could have made it one less they certainly would have because everybody involved wants this to succeed. Think of these as the ones that they simply could not get rid of no matter how hard they tried.

Perhaps it's a case of "we're proud that we got it all the way down to only 200 SPOFs in the final design, the earlier designs were way worse".

There is a fair chance of that, actually. If you start enumerating the ones that are obvious even to lay people (one rocket, one satellite, one set of mirrors and so on) then this is probably an extremely impressive low number.

I've been binging JWST content and definitely recall John Mayer saying in at least one interview that 344 single points of failure was as low as they could get it. And I can believe it. If you watch some of the more detailed interviews, he always stresses how every design choice was heavily labored over and reviewed.

It's an absolutely amazing effort, from the design through all of the problems they had to deal with along the way to seeing it sit there on top of that booster. A few months from now we could be in a completely new era of astronomy.

> Unlike say, the Apollo program, which had a guaranteed successful outcome?

Apollo was wildly dangerous. Apollo 1 killed the whole crew. The contemporary calculated failure odds for a Saturn V launch were 1/8. Compare that to the current Dragon 2 projected LOC risk of 1/276 [1].

[1] https://en.wikipedia.org/wiki/Crew_Dragon_Demo-2#:~:text=NAS....

I think GP was being sarcastic.

Absolutely. Those guys were laying their lives on the line with every launch.

I think this is your earnest attempt at it but its a good place to apply the story from yesterday 'be curious, not judgemental'.

That's literally what he is doing, isn't it? He thought something was off, and instead of spewing his opinion all over the place he requested information from folks more versed in the domain than him. He is literally being curious, not judgemental.

> I'm always baffled how everyone seems proud that the telescope has something like 200 SPOF during deployment, and if even one of them fails the whole mission could fail

I see this as judgmental, am I wrong in that?

It expresses an opinion, but I don’t think that has to be judgmental. The OP has to explain their position in order to ask for additional information.

I'd like to apologize if this sounded judgmental. English is not my main language, and perhaps my phrasing sounded like I assumed my domain specific opinion is better than the actual people working on the the project.

Reading the other comments, and maybe to contextualize to my question better, I'm more surprised by how the project is presented as marvelous to the public, rather than thinking that any technical part were overlooked.

While I'm sure that engineering teams at NASA and ESA have countless contingency plans, procedures and failure models. Medias around the project seem to focus on how fragile the deployment procedure is. Great engineering is an act of finding the best balance between opposing constrains, by building technically sound systems but also more importantly designing robust human or automated procedures.

In this story, in my opinion, the media presents a skewed explanation of why the project is incredible by highlighting that it would be incredible that such a brittle deployment procedure would even work.

While a farm tractor might have a factor of safety of 10, spacecraft are usually closer to 1.3, due to mass and efficiency constraints.

Space is hard.

> However, I'm always baffled how everyone seems proud that the telescope has something like 200 SPOF during deployment, and if even one of them fails the whole mission could fail.

Genuinely curious: how would you have achieved the mission goals with fewer SPOF?