Realistic, near-term, rotating Space Station

[mikelepage]

The angular momentum required to precess the axis 360 degrees will be 2 pi times the station's initial AM, not 4x. AM is a vector quantity and you want to add tiny bits of AM to make the vector 'walk' in a circle, so the 'walking distance' is just the perimeter of that circle.

My bad. Thanks for the correction, I misremembered in my post above. I vaguely recall we worked it out many pages ago on this thread. 

"Walking the vector around in a circle" is a good way to think of how big the additional energy cost is. Makes the question clearer - Is the collection area of solar PV that you can put on a sun-facing torus more than 2 pi bigger than what you can put on an ecliptic-perpendicular station?     

That, or have less overweight stations. 

If you have a ~3,000 tonne tomahawk station with a 3-to-1 mast, r = 300 m, Isp = 2500s, then it's only 11 N and 14 tonnes per year. Roughly 270 kW continuous, assuming no improvements beyond existing Starlink thrusters.

Maybe I'm imagining a tomahawk station as different to you. Is the thruster only firing once per rotation when it gets to the right position, or is the mast attached to a counter-rotating hub so it always sticks out at the correct orientation? Seems to me that either scenario has problems.

[Paul451]

You can put solar arrays on the edge facing the sun - but having that amount of area only getting illuminated once a rotation seems like a terribly inefficient use of solar PV.

OTOH, solar panels are cheap enough that the trade is probably going to be "waste solar panels" rather than, "spend extra on complex, heavy and expensive systems to have fewer solar panels".

I feel the same with reflected sunlight into the hab (and all the added issues of thermal control and solving day/night timing) vs having extra solar panels (which you use as a sun-shade) & LEDs.

Is it "waste" when it saves you so much complexity? Perhaps sometimes the best part is lots of the cheapest part.

I’ve let go of the idea I used to have that you need a flywheel with sufficient capacity to bring the whole station to a rotational stop (that needs to be done propulsively), but I am still including a much smaller CMG that can deliver sufficient torque to deliver the once yearly axis rotation on electric power alone.

Does this saturate the CMG? I can see where the energy is coming from, but can't see where the extra rotational momentum comes from unless the CMG continually increases speed.

[Paul451]

Re: Small thruster on a long stick

A rotating head, like a Canfield joint, would allow a single thruster to fire continuously through an endless 360 degree rotation (or anti-rotation, in this case.)

[Twark_Main]

The angular momentum required to precess the axis 360 degrees will be 2 pi times the station's initial AM, not 4x. AM is a vector quantity and you want to add tiny bits of AM to make the vector 'walk' in a circle, so the 'walking distance' is just the perimeter of that circle.

My bad. Thanks for the correction, I misremembered in my post above. I vaguely recall we worked it out many pages ago on this thread. 

"Walking the vector around in a circle" is a good way to think of how big the additional energy cost is. Makes the question clearer - Is the collection area of solar PV that you can put on a sun-facing torus more than 2 pi bigger than what you can put on an ecliptic-perpendicular station?     

Yep. I definitely identify with the old academicians' koan, "I've forgotten more than I know."   

I should say again it's the angular momentum cost (that's the vector going in a circle), not energy.  It's an important distinction because there are tricks for getting more AM out of a limited supply of energy / mass, with the big one being to maximize the effective thruster radius.

On an ecliptic-perpendicular station the cost of pointing is effectively zero (not one tau-th as much). There could still be non-zero costs for orbital stationkeeping or reboost, of course.

That, or have less overweight stations. 

If you have a ~3,000 tonne tomahawk station with a 3-to-1 mast, r = 300 m, Isp = 2500s, then it's only 11 N and 14 tonnes per year. Roughly 270 kW continuous, assuming no improvements beyond existing Starlink thrusters.

Maybe I'm imagining a tomahawk station as different to you. Is the thruster only firing once per rotation when it gets to the right position, or is the mast attached to a counter-rotating hub so it always sticks out at the correct orientation? Seems to me that either scenario has problems.

The first one. Technically it fires in an arc centered on the right position, like all real (non-impulsive) burns. This results in a small geometric loss, and you trade the extra propellant cost/mass vs the hardware required to increase the thrust.

For just doing precession torque you fire directly parallel to the spin axis, alternating North and South every 180 degrees. For mechanical cycle lifetime reasons you probably don't want to repoint twice per rotation, so I think you want (at least) two independently-gimballed thrusters or arrays, so one points North and one points South, with only trimming adjustments required.

Possibly it's a 'RAID array' with lots of small thrusters each with their own gimbal, for economies-of-scale and massive redundancy.

If you want to "mix in" some translation delta-v in the North / South direction, this is free: you just change the relative duty cycle between North and South. If you want translational delta-v perpendicular to the spin axis you just tilt the thrusters so they fire slightly off-axis in the same direction. This can potentially use very little extra propellant, due to the small-angle economics.

Similarly if you want to trim up or down the AM (to maintain the rotation rate as stuff moves, or to spin/despin), tilt so they fire off-axis in opposite directions, canceling out net translational movement but applying a net torque.

One architectural strategy is to roughly align the orbital altitude with this available "free" delta-v capability, so the annual precession fuel and/or energy does double duty for reboost.

[mikelepage]

I’ve let go of the idea I used to have that you need a flywheel with sufficient capacity to bring the whole station to a rotational stop (that needs to be done propulsively), but I am still including a much smaller CMG that can deliver sufficient torque to deliver the once yearly axis rotation on electric power alone.

Does this saturate the CMG? I can see where the energy is coming from, but can't see where the extra rotational momentum comes from unless the CMG continually increases speed.

Must admit I still struggle with torque/angular momentum problems, and personally think it is even harder than orbital mechanics, so perhaps I am using "CMG" incorrectly.

The mental model that I have learnt for dealing with this stuff was through right hand rules is as follows: 1) Angular momentum: The fingers curl in the direction of rotation, then your angular momentum vector points in the direction of the thumb. 2) Torque: The fingers curl in the direction of rotation that would result from the force being applied if there was no angular momentum, then your torque vector points in the direction of the thumb.
3) If you apply torque to a body that already has angular momentum, then the *actual* direction of of the spin axis will rotate along a path from the direction of of your thumb in (1), to the direction of your thumb in (2).

In answer to your question... I don't think so.

I had imagined the "CMG" as a gyro which nominally would be rotating with a rotation axis parallel to the rotation axis of the whole craft. Then - similar to what Twark_Main just described with the thruster firing in an arc centered on the right position with each rotation - this gyro's spin axis would be pivoted in the desired direction to apply torque once per craft rotation, centered on the right position. Not actually sure whether it makes a difference if this is out on a mast or near the center of mass/rotation. But yes I don't see why the spin speed of the gyro would need to change?

[mikelepage]

Re: Small thruster on a long stick

A rotating head, like a Canfield joint, would allow a single thruster to fire continuously through an endless 360 degree rotation (or anti-rotation, in this case.)

Given how hard it is to get "deployables" approved in the first place, I've been trying to avoid any space-exposed joints that would need to move continuously over the lifetime of the vehicle. Cold welding and all that. I'm trying to stick to things which deploy once and it's done. If you've got a duty cycle for thrusters turning on and off, that's not necessarily a bad thing either.

[mikelepage]

I should say again it's the angular momentum cost (that's the vector going in a circle), not energy.  It's an important distinction because there are tricks for getting more AM out of a limited supply of energy / mass, with the big one being to maximize the effective thruster radius.

I take your point. But those tricks for getting better bang-for-buck on AM are equally applicable to sun-facing versus ecliptic-perpendicular. The reason you'd choose to be sun-facing is because the energy acquisition advantage is real, I've just yet to see how it's worth it.

You can put solar arrays on the edge facing the sun - but having that amount of area only getting illuminated once a rotation seems like a terribly inefficient use of solar PV.

OTOH, solar panels are cheap enough that the trade is probably going to be "waste solar panels" rather than, "spend extra on complex, heavy and expensive systems to have fewer solar panels".

I feel the same with reflected sunlight into the hab (and all the added issues of thermal control and solving day/night timing) vs having extra solar panels (which you use as a sun-shade) & LEDs.

Is it "waste" when it saves you so much complexity? Perhaps sometimes the best part is lots of the cheapest part.

Good points. I think I'm back to designing ecliptic-perpendicular station  But a google search says LED lifetimes max out at (100k hours / ~11 years), so long-term, passive regulation still seems out of reach.

[Twark_Main]

I had imagined the "CMG" as a gyro which nominally would be rotating with a rotation axis parallel to the rotation axis of the whole craft. Then - similar to what Twark_Main just described with the thruster firing in an arc centered on the right position with each rotation - this gyro's spin axis would be pivoted in the desired direction to apply torque once per craft rotation, centered on the right position. Not actually sure whether it makes a difference if this is out on a mast or near the center of mass/rotation. But yes I don't see why the spin speed of the gyro would need to change?

Oh, how I wish it worked like that!

A CMG can apply a torque a few times, but if it keeps applying torque in the same direction (which it has to do here) then it will get saturated.

Think of a CMG as a tiny AM vector that has a fixed length (the AM capacity), but you can choose to point that vector in any direction and add it to your AM vector.  In order for the CMG to be able to make the AM "walk" in a circle... you need to store as much AM as the station spin, so it becomes the same as the case where you want to store the entire AM in the CMG.

Re: Small thruster on a long stick

A rotating head, like a Canfield joint, would allow a single thruster to fire continuously through an endless 360 degree rotation (or anti-rotation, in this case.)

Given how hard it is to get "deployables" approved in the first place, I've been trying to avoid any space-exposed joints that would need to move continuously over the lifetime of the vehicle. Cold welding and all that.

The Solar Alpha Rotary Joint is a counterexample, but nevertheless I'm still trying to avoid repointing every rotation. One cycle per orbit (similar to the SARJ) shouldn't be too bad.

I should say again it's the angular momentum cost (that's the vector going in a circle), not energy.  It's an important distinction because there are tricks for getting more AM out of a limited supply of energy / mass, with the big one being to maximize the effective thruster radius.

I take your point. But those tricks for getting better bang-for-buck on AM are equally applicable to sun-facing versus ecliptic-perpendicular.

The difference being you don't need to use the tricks, because you don't necessarily need to apply such big torques on a regular basis.  With a Sun-pointing station you don't have a choice.

You can put solar arrays on the edge facing the sun - but having that amount of area only getting illuminated once a rotation seems like a terribly inefficient use of solar PV.

OTOH, solar panels are cheap enough that the trade is probably going to be "waste solar panels" rather than, "spend extra on complex, heavy and expensive systems to have fewer solar panels".

I feel the same with reflected sunlight into the hab (and all the added issues of thermal control and solving day/night timing) vs having extra solar panels (which you use as a sun-shade) & LEDs.

Is it "waste" when it saves you so much complexity? Perhaps sometimes the best part is lots of the cheapest part.

Good points. I think I'm back to designing ecliptic-perpendicular station  But a google search says LED lifetimes max out at (100k hours / ~11 years), so long-term, passive regulation still seems out of reach.

You can boost both the lifetime (which is limited by thermal diffusion within the semiconductor) and efficiency of LEDs by underdriving them. See the "Dubai bulb" for an example. All it takes is more LED emitter semiconductors, which is a small fraction of the overall lamp mass.