But aren't there other ways to orient in space?
1. use pressure from the solar wind
2. have 3 electric motors on 3 axis. Wouldn't spinning those motors rotate the craft? Electric power to do it would come from solar panels, giving it plenty of fuel.
We're getting better all the time at building more reliable components (including reaction wheels and cryocoolers) though. Until a few years ago, the life of something like JWST would be limited by the amount of liquid helium on board to cool the components. Modern cryocooler technology (aka a space grade refrigerator) is good enough to cool it indefinitely. Solid state cryocoolers, previously unachievable, are now apparently available for some applications (important not only for reliability but also to reduce vibrations).
Reaction wheels can be used for attitude control but they still have to be unloaded by thrusters after maneuvering for a while. You're right that you could use a rudder (probably two rudders would be required for 3d attitude control) and have to have a balanced solar wind profile (JWST does actually have a solar wind balancing flap, but I don't think it's adjustable like a rudder). But solar wind won't act fast enough if you want to quickly change attitude for observations. And you can't use reaction wheels for stationkeeping. It very much matters where the telescope is, since if it drifts too far away from Earth it will be much harder to send high bandwidth data, and if it's too close to Earth, Moon etc. it will have no way to orient without heating up or blinding itself with the IR sunlight reflected by them.
And to calm down anyone afraid of JWST sharing the same fate - construction of reaction wheels have been changed some time ago to make them significantly more reliable. The source of issues on Hubble, Kepler, FUSE, Hayabusa, Dawn and TIMED was electrical arcing between metal parts of reaction wheels. Static charge was building up like when you rub a ballon against your head. That charge caused arcing that in turn caused metal pitting and increased friction leading to failures. That failure mode was understood only in late 2007, when Kepler was already fully build and ready for launch.
JWSt uses new generation ceramic bearing in its reaction wheels, they have been used in spacecrafts since 2010 with great performance.
I cannot find primary source for JWST using ceramic bearings right now, but I know they are manufactured by Rockwell Collins Deutchland and have good track record: https://jwst.nasa.gov/content/forScientists/faqScientists.ht...
TELDIX (today owned by Rockwell Collins Deutchland) invented use of ceramic bearing in reaction wheels for spacecraft use in 1978: https://patents.google.com/patent/DE3027209A1/en
so it would make sense that JWST uses ceramic bearings. Other manufacturers (ITHACO) have switched to ceramic too.
It should be noted that reaction wheels can saturate when the motor reaches its top speed. One then needs to spend fuel to provide a counter-force while the wheel to spins down.
So even with reaction wheels running off solar panels or similar you need fuel, though much less.
edit - apparently they do do something like this? https://news.ycombinator.com/item?id=29857031
BTW, I doubt there's enough experience with space hardware to accurately predict it's life, especially since each machine is a one-off.
https://www.jwst.nasa.gov/content/webbLaunch/deploymentExplo...
https://ntrs.nasa.gov/api/citations/20140007519/downloads/20...
https://space.stackexchange.com/questions/35399/how-will-jws...
Kepler used a similar strategy (though I don't know what its desaturation strategy was): it only ran out of fuel very quickly after its reaction wheels failed.
An internal fuel tank and a high-capacity tank externally
Once the fuel is getting low, launch a refuel. when the new supply gets closed, eject the previous tank, dock the new one.
Easy.
When they managed to du an in air refuel of a SR71 in 1970s . Surely we can dock a fuel tank in 2022?
"Unlike Hubble, Webb isn't designed to be fixed by astronauts. But it can be refueled robotically. Zurbuchen says that 'once this telescope is deployed, I'm going to put all the effort towards developing that technology, and so within the 10-year lifespan, we can go refuel it'"
https://nitter.42l.fr/marinakoren/status/1474367236244750345...
https://nitter.42l.fr/Dr_ThomasZ/status/1474398711505580032#...
https://hubblesite.org/mission-and-telescope/servicing-missi...
Nothing in space is easy.
The issue is that its position at the lagrange point L2 is an unstable equilibrium, which requires occasional adjustment using thrusters. In terms of gravitational potential energy, its position in space is a saddle point, not a local minimum.
Due to that JWST will always be on 'close side of L2' and technically in slow freefall back to Earth and boosted up periodically, but always a bit short of passing to the other side.
That's a fixable design problem. Or you can rotate the telescope.
> stray light reflected from Earth would limit its field of view
What? The earth and the moon are nearby anyway. What about that reflected light? And heck, what about the sun limiting its field of view?
You are right, it is fixable. It was fixed by adding active station keeping to the telescope.
> Or you can rotate the telescope.
There are limits on its rotation with respect to the Sun, dark side must be kept away from sunlight at all times. It can rotate 5 degrees "pitch down" toward the Sun and 45 degrees "pitch up". Gimballed antenna has enough authority so that it can communicate with Earth at whatever valid rotation telescope is so that science operations are not interrupted for transmitting data.
> The earth and the moon are nearby anyway. What about that reflected light? And heck, what about the sun limiting its field of view?
That light is reflected back by sunshield as all 3 bodies are behind it. Sun is limiting field of view but area of exclusion changes as telescope orbits and it can image every point in the sky at least every 6 months and 39% of the sky at any given moment.