send e-mail to A Voyage To Arcturus (note: Spam Arrest is on, so a one-time-only confirmation will be requested)
[ 20040130 ]
OK, here's the second paradigm-shifty idea: replace the Hubble, not with one telescope, but with many. We shouldn't be thinking of "space telescope" in the singular any more than we should think that there can be only one "space transportation system," carrying a handful of humans into orbit three or four times a year.
But if that one telescope will cost $1.6 billion, how can we launch "many" -- and within the next three years -- without devouring the entire space science budget (see page 18; warning: 2.4 MB *.pdf) and then some? Recalling the triple constraint, if I'm suggesting holding the line on resources and accelerating schedule, then scope -- that is, requirements -- must be reduced.
The obvious requirement to lower is size. Telescope cost scales roughly as the cube of aperture, meaning that a 1-meter telescope identical to Hubble in all other respects would cost about 7% as much. Another requirement to either greatly relax or eliminate is cryogenically-cooled instrumentation; the Spitzer Space Telescope (formerly known as SIRTF) cost $740 million even though its aperture is only 0.85 m, largely because of a stringent requirement to detect faint sources with its infrared array camera and spectrograph and multi-band imaging photometer.
Even with these constraints, any one of our replacement space telescopes would outperform nearly every telescope on Earth; Hubble was the best telescope humans had ever built before the repair mission. So how expensive would they be?
Disclaimer: this is very much a ROM estimate, meaning that actual costs could be the number I come up with, multiplied or divided by (at least) 3.
Hubble's development costs were around $2 billion -- closer to $4 billion in today's dollars. I recall hearing an astronomy grad student at Chicago in the late '70s grouse about how for the cost of the space telescope, we could have built eight 4-meter ground-based telescopes. Applying the cube-of-aperture rule mentioned above, each 4-meter scope would ordinarily cost nearly 5 times what a 2.4-meter telescope would. The implied cost differential is a factor of nearly 40. Now, browsing this source, we find: "A 0.4 meter telescope produced by a manufacturer of amateur telescopes may cost 1/4 the cost of the same size professional telescope. This difference is partly due to the quantity of telescopes produced, but is largely due to the step up in performance."
I infer that a relatively simple space-based telescope should cost about 150 times as much as a ground-based amateur telescope of the same size. Grazing over to this page, I note -- after wiping the drool off my chin -- a price of $10,749 for a 25" 'scope. Once again applying the cube-of-aperture relationship, then multiplying by 150, I arrive at a figure of only $6.3 million for a 1-meter telescope in space.
Well, er, except for launch costs. And if we're going to figure those in, we'd better have some idea of the size and mass of one of these things. For comparison, the Hubble is 13.2 meters long, 4.2 m in diameter, and masses about 11,000 kg. Our hypothetical 1-meter telescope would be perhaps 5 meters long, 1.5 meters in diameter, and mass about 800 kg. What could launch it, and how much would it cost?
Reviewing various reports available via this page, and looking up payload capacities in Mark Wade's incomparable Encyclopedia Astronautica, I built a table of launchers by asking price, payload to low Earth orbit, and cost per kilogram of payload to LEO:
Dnepr | $8-11M | 4,500kg | $1,800-2,400 |
START | $9M | 632kg | $14,200 |
Rockot | $12-15M | 1,800kg | $6,700-8,300 |
Cosmos | $12M | 1,400 kg | $8,600 |
Pegasus | $14-18M | 443 kg | $32,000-41,000 |
Taurus | $20-30M | 1,380kg | $14,500-21,700 |
Soyuz | $30-50M | 6,220kg | $4,800-8,000 |
Delta 2 | $45-55M | 5,089 kg | $8,800-10,800 |
Proton | $60-85M | 21,000 kg | $2,900-4,000 |
Atlas2 | $65-75M | 7,280kg | $8,900-10,300 |
Atlas3 | $65-75M | 10,718kg | $6,100-7,000 |
Atlas5 | $65-75M | 12,500kg | $5,200-6,000 |
Zenit3SL | $65-85M | 13,740kg | $4,700-6,200 |
H2A | $70-100M | 11,730kg | $6,000-8,500 |
Ariane5 | $125-155M | 16,000kg | $7,800-9,700 |
Notwithstanding that the above are approximate figures -- the asking price is rarely obtained in the current depressed launcher market, booster configurations vary, and performance varies significantly by orbital altitude and the latitude of the launch site -- we may reasonably expect to pay no more than $12 million for the launch. I note that one of the least expensive vehicles, the Dnepr, could launch several such telescopes at once if they could somehow be fit inside its payload fairing.
I conclude that less than $20 million could put us well on the way to launching one or more space telescopes before Hubble ceases operation. Compare perhaps half a billion dollars for the cancelled Hubble-maintenance Shuttle mission.
Now all this project needs is a name. ;)
Jay Manifold [4:04 PM]
So if the Hubble conks out in '07, we get a four-year hiatus before the James Webb Space Telescope comes on line, right? So we'd better save the Hubble, right?
Maybe not. I can think of a couple of better alternatives, and both of them involve a bit of paradigm shifting, which makes them that much more attractive.
Let's take the obvious one first: reading Who Says Astronomy Isn't Practical? (IV), we find that the technology for the JWST is borrowed from existing spy satellites -- that in fact, the JWST will be little more than a civilian version of a present-day spysat.
Fine. So let's point one of them up instead of down.
But wouldn't that show the Bad Guys™ too much about what we can do? -- Not unless we used the very newest one, and even then, it might not matter too much.
How long is a secret worth keeping? Via Virginia, we find the following astonishing recommendations in the Final Report of the Defense Science Board Task Force on Secrecy, co-authored by Edward Teller, no peacenik:
It is unlikely that classified information will remain secure for periods as long as five years, and it is more reasonable to assume that it will become known to others in periods as short as one year.
The Task Force noted that more might be gained than lost if our nation were to adopt -- unilaterally, if necessary -- a policy of complete openness in all areas of information ...
As a general guideline, one may set a period between one and five years for complete declassification. This time limit should be extended only if clear evidence is presented that changed circumstances make such an extension necessary.
Anybody reading this blog will be all too familiar with Moore's Law, and perhaps also with Ray Kurzweil's double exponential growth model of technological advance. The age is coming, and now is, when a lead of even a year or two is so large that we need not fear revealing our capabilities.
My second idea? See the next post, later today.
Jay Manifold [5:51 AM]
[ 20040129 ]
'Save the Hubble' campaign soars, says the BBC's Dr David Whitehouse, in an article that points to Save The Hubble, a Brazilian website. Meanwhile, Hubble, the Beloved is the amazing headline in the WaPo (registration required), where we read of well-meaning citizens endearingly but ineffectually trying to help:
One common suggestion from the public has been that NASA should take Hubble to the space station for the maintenance, but for a variety of reasons -- including the largely dissimilar orbits Hubble and the station occupy -- that would not be feasible, [Bruce] Margon [associate director for science at the Space Telescope Science Institute] said.
Hubble-lovers also have suggested sending an unmanned robotic mission to perform some of the tasks an astronaut normally would complete manually.
Margon said the institute will be setting up a Web site soon to accept public suggestions for the Hubble.
I draw the following lessons:
- For interested observers: the public, and the electorate, can be constructively fascinated with science, even one of the historical sciences whose central tenets many people refuse to accept -- astronomy. And affective gains precede cognitive gains (see items 33 and 34).
- For policymakers: aesthetics matter, and space science has aesthetics to burn if it's done right. Which leads to ...
- For NASA: require every future space probe to be capable of returning imagery. This need not be from the visual spectrum.
- For spacecraft designers: if it's unmanned, make it serviceable by a robot.
One more quote from the WaPo article:
"It's extremely difficult to think about alternate ways to accomplish these chores," Margon said. "But not impossible."
As another blogger would say: indeed. I'll share some ideas in my next post, so graze on back tomorrow.
Jay Manifold [5:54 PM]
[ 20040125 ]
Fred Kiesche of The Eternal Golden Braid saw the earlier post on this topic and promptly forwarded me a chunk of discussion from the "amastro" Yahoo! group, in which the invaluable Brian Skiff wrote:
Most folks simply declared that Hipparcos was wrong with the Pleiades distance, and although observations such as the one described in this news item are convincing, the real problem has been identifying where the error is in the Hipparcos data. That too has been done recently (see e.g. 2003AJ....126.2408M), and shown to be an engineering glitch in the reductions. I suspect (and hope) that a complete revised Hipparcos catalogue will be forthcoming that will get rid of all the lumps in the parallax error distribution.
I should also explain that the earlier post is the sort of thing that happens when I don't take melatonin and therefore am awake in the small hours of the morning. Not bad, but I missed one kinda obvious value-added thing, which is to compare the Pleiades to the Sun. So turning again to this page, which lists the brighter members of the cluster, we find that a simplified version of ranking by apparent luminosity would look like this:
Star | mv |
Alcyone | +2.86 |
Atlas | +3.62 |
Electra | +3.70 |
Maia | +3.86 |
Merope | +4.17 |
Taygeta | +4.29 |
Pleione | +5.09 |
Asterope | +5.31 |
Celaeno | +5.44 |
Now suppose that they are all really about 440 light-years away. Their absolute magnitude is that of their appearance at 10 parsecs, or 32.6 light-years. Since their actual distance is 13.5 times this, applying the inverse-square law, we find that they would look 182 times brighter. Each full unit on the stellar magnitude scale represents a brightness difference of 5√100 @ 2.512. The easy way to figure this on a scientific calculator is to take log10 182 = 2.26, and multiply it by 2.5 (since [5√100]2.5 = 10), yielding a stellar-magnitude difference of 5.65. So the absolute magnitudes of the Pleiades are the above values with 5.65 subtracted. Just for fun, I've created a table comparing them to the apparent (visual) magnitudes of the nine brightest stars in the night sky as seen from Earth:
Star | mv | Mv | Star | mv |
Alcyone | 2.86 | -2.79 | Sirius | -1.46 |
Atlas | 3.62 | -2.03 | Canopus | -0.72 |
Electra | 3.7 | -1.95 | a Centauri | -0.27 |
Maia | 3.86 | -1.79 | Arcturus | -0.04 |
Merope | 4.17 | -1.48 | Vega | +0.03 |
Taygeta | 4.29 | -1.36 | Capella | +0.08 |
Pleione | 5.09 | -0.56 | Rigel | +0.12 |
Asterope | 5.31 | -0.34 | Procyon | +0.38 |
Celaeno | 5.44 | -0.21 | Achernar | +0.46 |
As you can see, were the Pleiades to be relocated to the Solar neighborhood, they would comprise the five brightest stars in the night sky, and 8 of the brightest 10.
Now to compare them with the Sun, whose absolute magnitude is +4.83. Taking the difference between this and the values listed above and raising 5√100 to that power, we find:
Star | mv | Mv | LSun |
Alcyone | 2.86 | -2.79 | 1120 |
Atlas | 3.62 | -2.03 | 555 |
Electra | 3.7 | -1.95 | 515 |
Maia | 3.86 | -1.79 | 445 |
Merope | 4.17 | -1.48 | 334 |
Taygeta | 4.29 | -1.36 | 299 |
Pleione | 5.09 | -0.56 | 143 |
Asterope | 5.31 | -0.34 | 117 |
Celaeno | 5.44 | -0.21 | 104 |
From the above figures and the inverse-square law, we may infer that to receive the same amount of light that Earth does from the Sun, a planet revolving around one of these stars would have to be about Saturn's distance from Celaeno, and about Pluto's distance from Alcyone. But that light would have lots more UV in it, and stars as massive as these must be aren't going to stay on the main sequence for very long. We can figure their masses from the mass-luminosity relationship (page down), simply by raising the luminosity to the 1/3.5 power, yielding:
Star | mv | Mv | LSun | MSun |
Alcyone | 2.86 | -2.79 | 1120 | 7.4 |
Atlas | 3.62 | -2.03 | 555 | 6.1 |
Electra | 3.7 | -1.95 | 515 | 6.0 |
Maia | 3.86 | -1.79 | 445 | 5.7 |
Merope | 4.17 | -1.48 | 334 | 5.3 |
Taygeta | 4.29 | -1.36 | 299 | 5.1 |
Pleione | 5.09 | -0.56 | 143 | 4.1 |
Asterope | 5.31 | -0.34 | 117 | 3.9 |
Celaeno | 5.44 | -0.21 | 104 | 3.8 |
Now let's assume that these stars are using up their fusable isotopes at a rate proportional to their luminosity, that the amount of those isotopes is proportional to their mass, that they're 100 million years old, and that the total lifetime of the Sun (relative to whose luminosity and mass I have expressed the Pleiadeian values in the table) is 10 billion years. Then their remaining lifetimes are:
Star | mv | Mv | LSun | MSun | Years Left |
Alcyone | 2.86 | -2.79 | 1120 | 7.4 | (negative!) |
Atlas | 3.62 | -2.03 | 555 | 6.1 | 10 million |
Electra | 3.7 | -1.95 | 515 | 6.0 | 16 million |
Maia | 3.86 | -1.79 | 445 | 5.7 | 28 million |
Merope | 4.17 | -1.48 | 334 | 5.3 | 57 million |
Taygeta | 4.29 | -1.36 | 299 | 5.1 | 70 million |
Pleione | 5.09 | -0.56 | 143 | 4.1 | 190 million |
Asterope | 5.31 | -0.34 | 117 | 3.9 | 230 million |
Celaeno | 5.44 | -0.21 | 104 | 3.8 | 260 million |
Obviously this doesn't produce an entirely correct result in the case of Alcyone (then again, maybe it's already blown up and going to fry us when the g-rays get here, presumably by 2444 AD, heh). The actual lesson here is that stars that big don't stick around long enough for planets and life to appear. You can read lots more about the Pleiades at this page.
Jay Manifold [1:34 PM]
