Comment by whoknowswhat11 | Hacker News Reader

Comment by whoknowswhat11 | Hacker News Reader

Parent Full thread

Exactly - try new things, experiment.

The existing funding streams are calcified, and we've gotten to a point with the monster projects that they basically turn into drag out career builders.

Things like James Webb, TMT etc. At some point we have to try new things.

If you look into TMT closely, the actual optics package etc - $3B is a ridiculous price. It's just totally crazy. If we pulled pricing apart we'd find out that an incredible amount is going into overheads of various sorts.

ASML makes machines with 100,000+ parts, the precision needed is absolutely mind blowing. 4 747's and 40 containers to ship it. $200 - $500M tops.

https://www.cnbc.com/2021/10/12/chip-machine-maker-asml-will...

I'm serious, the precision of mirrors in these machines absolutely rivals or crushes astro precision. "The mirrors are polished to a smoothness of less than one atom’s thickness. To put that in perspective, if the mirrors were the size of Germany, the tallest ‘mountain’ would be just 1 mm high."

Now also lookup how large these EUV mirrors are, how much light they actually reflect, and how fast you need to swap them out because they have been damaged.

The swap is generally for light collector, being exposed to a plasma light source! I would expect demands in astro to be much simpler. These machines are in the 10's to 100's of kw of power needed.

Size is 15-20 inches. Despite what you might think, "thirty" meter telescope does not have a 30M primary. The move is to segments.

EUV is one of the hardest to target for reflection. I'm not sure what astro is getting in terms of reflectivity if they even need EUV, my guess is Zeiss is crushing that in most cases.

You are assuming a great deal of ignorance among the people you are talking to or about (hardly anyone would be surprised that mirrors are segmented, this has been the case since forever). And you are at the same time sprinting past your own ignorance (look at how these mirrors are made or their reflective properties - a quick google would should you plenty of information about their extremely restrictive spectral properties). Or their stability or again, their size, and compare it to the requirements of large size space or terrestrial telescopes (and, as you might know, no, you can not use 10 small telescopes instead of one large one).

This is what is so interesting to me. I'm not saying EUV astro in particular. And frankly how does EUV even work terrestrially? From what I've seen it really doesn't work so well.

What I am pointing out is that something has gone wrong, badly, in astronomy. The TMT is the telescope I've been following most closely. The costing just doesn't make sense. The project delivery seems horrible.

Shall we use the TMT as our example? Its going to HAVE to have some sort of actuation on mirrors, because its sitting under atmosphere. Period. There is going to be noise from atmo in the imaging. So you are already starting behind.

Transmission losses for infra / UV.

My proposal is simple. Put a $2B RFP out for a 9 arc second telescope if that is what is needed, perhaps 10 $1M proposal awards, down select to 3 $20M awards for further work, then down select to 1 or 2 $1B-$2B awards depending on where the bidding comes in. Payment as delivery is met.

Find folks who instead of criticizing dirt cheap access to space can think of ways to take advantage of it.

Note that I say this having written sections of or supported very successful govt grant proposals, chased all the preference points, yadda yadda. An amazing waste of time if someone else had the "in" and you didn't realize it, and sometimes you crush everyone for no good reason other than your proposal checks every box very carefully.

If we set aside the EUV tech comparisons you were making, then sure, competitive awards make a ton of sense. The thing is, this is how many grant agencies usually work (unless they get some weird overrides from congress or the occasional failure of the competitive system (a.k.a. market failure?)).

> My proposal is simple. Put a $2B RFP out for a 9 arc second telescope if that is what is needed, perhaps 10 $1M proposal awards, down select to 3 $20M awards for further work, then down select to 1 or 2 $1B-$2B awards depending on where the bidding comes in. Payment as delivery is met.

This is actually how grant agencies like the NSF, NIH, DARPA, NASA, etc do it. Maybe not the exact parameters, money amounts, and specs you mentioned, but pretty much the exact overall process. Sure, "grant writing" expects a particular writing style depending on agency, and academic accolades are part of the (I would say reasonable) signaling of who is competent. But these proposals do indeed bubble up after competitions between many different teams, and nothing stops a startup or a hard-tech conglomerate in another field from putting in a proposal.

But it's not actually how they do it.

Payment is based on COST incurred, not delivery. Or expenditure to a limit regardless of outcome.

For Webb we have this:

"All told, NASA has spent about $9.7 billion on the project, which has been delayed by more than six years in the past decade. That's in stark contrast to estimates for the project when it was approved, when it was meant to cost $1 billion to $3.5 billion and launch sometime between 2007 and 2011, according to the GAO report."

Can I ask who is footing the bill if this was payment as delivery was met? Oh right, because it's not.

The only recent efforts towards fee for service were the HLS award and some of the commercial crew program work. Interestingly, the women heading that up (Kathy - very successful) appears to have been sidelined from directing new space development for someone from old space / old boys network.

We really do need a fresh start - the more I poke at this the worse it looks.

With these complaints and suggestions, I can mostly agree, and do support evolution in this direction with whatever voice I have.

Reply on news.ycombinator.com