How would you make "Sensors" for a starship?

[Junghalli]

You are not going to establish any of this until you start assigning some numbers. Put a pricetag on how much it costs to manufacturea a telescope of a given size, resolution, field of view, ect. and we can get a discussion going. Intelligence gathering and their countermeasures don't develop in a vacuum, and that's exactly what you're trying to do.

Determining how much telescopes will cost is going to depend on knowing a lot of stuff about the civilization's technology level and logistics that we don't know. The best thing I can do with what little we have is to try to estimate how much material sensor nets of different size would consume.

OK, I'll try a hypothetical test case. Let's use GMT's calculation based on modern telescopes being able to get an image of Pluto in a one second exposure. This gives you an effective sensor range of roughly 1 AU to detect a 4 GW light in a 1 second exposure. Going by my earlier calculations for decoys, this is sufficient to detect everything more energetic than HB-fusion rockets, but not low-power NERVA or relatively low thrust (1 m/s^2) chemical rockets.

GMT also quantified earlier that using a modern wide field telescope as a base we might calculate 400 telescopes to watch the entire celestial sphere. Obviously, for a decent sensor grid you will need several times this much.

The most efficient shape to give complete coverage of that 2 AU sphere is probably a rosette of 4 telescopes orbiting in the plane of the solar system, with two statites "hovering" over the north and south pole of the sun. That gives us 6X400 = 2400 telescopes. If we assume statites are two problematic and make the polar sattelites an orbital rosette that gives us 8X400 = 3200 telescopes.

As we go outward, spherical geometry suggests the size of our rosettes must octuple each time we double the radius of the surveyed area. To watch 4 AU we must have 25,600. To watch 8 AU we must have 204,800. To watch 16 AU we must have 1,638,400. And to watch 32 AU we must have 13 million.

Now, to determine how much material is required. Let's assume each individual telescope has a mass of 10 metric tons, a little less than Hubble. That means the entire array consumes 130 million tons of material.

Now, let's say around .1% of the material a typical asteroid is useful for building telescopes like this. This array would therefore require the deconstruction of a planetoid with a mass of 1.3 X 10^11 tons. Now, a 1 km diameter asteroid might have a mass of around 2 billion tons (ref). Applying spherical geometry, our array would require the deconstruction of a single spherical 4 km diameter asteroid.

Assuming the average warship masses 1000 tons, this represents an opportunity cost of 130,000 warships.

Realistically, a really effective sensor net would need to be at least several times denser, and therefore consume several times as much material. You could avoid being seen by this thing by using fraction of a second burns, for instance, or by using a number of low thrust, low Isp rockets.

On a strict materials basis, this network should be readily feasible. There are probably more than a million asteroids in the 1 km range (ref), so we have barely scratched the surface of the solar system's resources. Of course, this does not count the difficulty of turning that material into telescopes, which will vary enormously depending on the technology and industrial techniques available.

Conclusion: a mature spacefaring civilization which extensively exploits space resources and has an economy that dwarfs our own would probably find this worthwhile. One that's just starting out ... probably wouldn't.

---------

An alternate, probably more productive approach might be to try to station sensor platforms close to enemy bases to watch them, while maintaining only a bare minimum of sensor coverage elsewhere to track any burns enemy ships make in deep space. This, of course, will probably use up much less material, but will require a delicate balancing act. Your observation platforms must be able to get close enough to enemy bases to observe them in some detail, while being small and inconspicuous enough not to be detected by the enemy.

It's worth pointing out something here: you've been saying an observation platform may not be as stealthy as a warship because it will need to be a monster telescope. With this strategy you will not need a monster scope. You will need a scope capable of keeping an eye on an enemy base from relatively close range; a few hundred thousand to a few million km away.

Of course, this strategy has its own downsides. Like I said, it requires you to be able to sneak the platforms into range of enemy bases, which is going to be tricky, though it should still be considerably easier than trying to sneak around with warships. It also requires you to know where all the enemy bases are, knowledge you may not have depending on the situation. And there might well be some delay between the time an enemy base goes online and the time you can get a covert observation platform out there.

[Wyrm]

Determining how much telescopes will cost is going to depend on knowing a lot of stuff about the civilization's technology level and logistics that we don't know. The best thing I can do with what little we have is to try to estimate how much material sensor nets of different size would consume.

The logistical rules in effect in the universe obviously affect the balance of power between sensors and stealth.

OK, I'll try a hypothetical test case. Let's use GMT's calculation based on modern telescopes being able to get an image of Pluto in a one second exposure. This gives you an effective sensor range of roughly 1 AU to detect a 4 GW light in a 1 second exposure. Going by my earlier calculations for decoys, this is sufficient to detect everything more energetic than HB-fusion rockets, but not low-power NERVA or relatively low thrust (1 m/s^2) chemical rockets.

Why do you assume that the 4 GW is going to all be radiating at visual wavelengths for these craft, as per GMT's assumption? The waste heat calculation is the power radiated by the engine/plume integrated over all frequencies. The visual range is only a narrow window on that entire light curve, and for low temperatures, that is going to be rather small fraction of the total — and for an optical observation platform, this small fraction is all you're going to see.

GMT also quantified earlier that using a modern wide field telescope as a base we might calculate 400 telescopes to watch the entire celestial sphere. Obviously, for a decent sensor grid you will need several times this much.

With GMT's combination of 10 x 10 degree² and 95 megapixel detector gives a rough angluar resolution of 3.784734 seconds. At 1 AU, this gives you a square patch ~9.9 million kilometers on a side. An enemy at that distance will have plenty of room to maneuver before you're able to detect any proper motion out of him. Even if you see me, you will not know how I've moved thrusting no matter how many platforms you have.

[Junghalli]

Why do you assume that the 4 GW is going to all be radiating at visual wavelengths for these craft, as per GMT's assumption? The waste heat calculation is the power radiated by the engine/plume integrated over all frequencies. The visual range is only a narrow window on that entire light curve, and for low temperatures, that is going to be rather small fraction of the total — and for an optical observation platform, this small fraction is all you're going to see.

Ah. In that case you'd probably want a number of telescopes capable of searching different parts of the spectrum. This will increase the required size of the array, probably by at least an order of magnitude or two (guesstimate, please feel free to correct it).

With GMT's combination of 10 x 10 degree² and 95 megapixel detector gives a rough angluar resolution of 3.784734 seconds. At 1 AU, this gives you a square patch ~9.9 million kilometers on a side. An enemy at that distance will have plenty of room to maneuver before you're able to detect any proper motion out of him. Even if you see me, you will not know how I've moved thrusting no matter how many platforms you have.

Hmm, suggesting, again, the size of the array will have to be increased by several orders of magnitude?

Sounds like the idea of trying to sneak sensor platforms into close proximity to the enemy bases is likely to be a lot more practical using the tech assumptions I have been. Of course, it has its own downsides.

[Wyrm]

Ah. In that case you'd probably want a number of telescopes capable of searching different parts of the spectrum. This will increase the required size of the array, probably by at least an order of magnitude or two (guesstimate, please feel free to correct it).

The photons you receive from each sensor must be above the noise floor for each sensor separately. The noise floor means that you receive about the same spurious signal from all areas of the sky. That means that when you integrate over all the sensors, you get even more spurious signal from each pixel of the sky. To be detectable, the signal has to overcome the noise in all frequencies instead of just one. Simple math says that if the signal is below the noise floor of all the sensors, it will be below the noise floor of the integration. The only way around this is to increase the exposure time or the aperture size.

Hmm, suggesting, again, the size of the array will have to be increased by several orders of magnitude?

This is not an aperture size limitation. You can detect down to about 0.15 seconds of arc with a 1 m diameter aperture. This is a limit imposed by the fact you're using a 95 Mpx sensor to survey a 10 x 10 degree² patch of space. You need more pixels surveying the sky. That means complicated optics to split up the image into different patches at the very least. That being said, there are focal length issues.

[GrandMasterTerwynn]

Hmm, suggesting, again, the size of the array will have to be increased by several orders of magnitude?

This is not an aperture size limitation. You can detect down to about 0.15 seconds of arc with a 1 m diameter aperture. This is a limit imposed by the fact you're using a 95 Mpx sensor to survey a 10 x 10 degree² patch of space. You need more pixels surveying the sky. That means complicated optics to split up the image into different patches at the very least. That being said, there are focal length issues.

Indeed. Though, if we were to assume the builders of our theoretical telescope were to increase the density of their detectors by a factor of 10, we'd have a 1 gigapixel effective sensor watching our 10x10 degree patch of sky. This would cut the square down to a mere 826 kilometers. A warship accelerating at a modest 0.1 gravities would take over 21 minutes to cross this square. One doing a crushing 5 gravity emergency acceleration could do it in 3 minutes.

Not bad, one might think, except a denser detector generates more internal noise, and will require much more aggressive cooling and/or longer exposure times. One cannot, however, continue increasing the density of photon detectors ad-infinitum. Kepler's 95 megapixel-effective detector is an array of forty-two 2.2-megapixel sensors. One could envision replacing them with an identical array of twenty-two megapixel detectors, but to resolve down to the telescope's diffraction limit would require an array of 1.4-gigapixel sensors. Sensors likely made of unobtanium.

[Beowulf]

Unless I'm screwing up my optics, you don't necessarily need a physically denser sensor, but should be able to use a larger sensor. Given that Pan-STARRS does use a 1.4 Gigapixel sensor (actually 4 of them on 4 telescopes to compensate for the divisions between the sensors, and noise, etc), a 1.4 Gigapixel array of sensors doesn't seem too unlikely.

[Surlethe]

This is a highly informative thread. I'm going to move it to the Library.

If you want to continue informative discussion, feel free, but I get the feeling that the thread has mostly run its course.