Starship On-orbit refueling - Options and Discussion

[OTV Booster]

If there's no advantage to using a selective solar coating (which seems to be true), then we're back to MLI.

I wouldn't say no advantage. At a minimum, you could use the paint (not the tiles) on the parts of the ship that don't have to be jet black and that would reject 90% of the heat from sun exposure at very little cost. But it definitely doesn't buy you ZBO. Not with a standard Starship, anyway.

At this point, I'd like to talk a little bit about exactly how the fuel is going to be transported. (Apologies if this has already been discussed. Extra apologies if it's already been discarded with good reason!) :-)

I know Elon doesn't like depos, but I'm starting to think it makes more sense to have a special "fuel depot" (for want of a better word) that a standard-issue Starship would carry as cargo rather than have a special tanker version of Starship. That would mean that two Starships would never dock to each other directly--not for purposes of fueling, anyway.

My reasoning is as follows:

First, that means you don't have an extra version of Starship with extra-large tanks. Second, the fuel depot would have the hardware required to mate to the QD ports on a Starship and do the refueling, so you wouldn't need a version of Starship with that equipment on it somewhere. (Or, worse, add that hardware to every Starship.)

The basic fuel depot would just be fuel tanks, the docking attachment, and pumps. It wouldn't ever leave the cargo bay of the Starship that brought it up to orbit. It's the sort of thing you'd use to refuel the Lunar Starship, assuming that only needs one extra load of fuel. You'd launch the two Starships at about the same time so boiloff wouldn't be a big concern.

A more complex fuel depot would be larger and have a cryocooler, radiator, extra insulation, and QD ports identical to those on a Starship. It would be put in orbit by itself and get filled up on multiple visits by Starships carrying basic depots full of fuel and LOX. (Note that you wouldn't need to develop this until after the lunar missions.)

That sort of assembly probably would benefit from extensive use of Solar White tiles. Yes, it would still need a cryocooler, but the tiles ought to reduce the mass of the cooler, the radiator, and the solar panels needed to power it.

You would cope with the orbital precession problem by timing the initial orbit and the refueling orbits so the depot is in the right place at the right time to refuel the Starship dedicated to the mission. After the mission Starship departed, you could send up another Starship to retrieve the empty depot.

If you wanted to use the same depot for multiple missions, you could use something like the RAAN-Agnostic 3-Burn Departure Methadology for Deep Space Missions from LEO Depots that has been talked about earlier.

Again, apologies if this has already been debated/discarded.

There has been discussion on several of the ideas you've offered up.

Keeping all the transfer hardware kit as a cargo has good points and bad points. It all boils down to the classic "it depends."

First off, a disclaimer: I am opposed to something that stays inside a cargo bay. It could work if the ship carrying it is nosed up between the tanker and the receiving ship and just acts as a fuel pump but that has safety issues for a crewed receiver. If it has its own tankage so it can act as an accumulator and that tankage is separate from the ship it is mounted in, it's not enough. If it uses the ships tanks, it needs some one off plumbing, which moves it in the direction of a special build. Then at transfer time it will need to pump up hill just to get the props to the point of transfer.

An important point you addressed is GSE plate gender mating. The four solutions so far discussed were: 1) the tanker deploying a gender bender on top of its GSE plate; 2) a tanker having a second, gender bent GSE plate; 3) the tanker being already gender bent and a temporary adapter mounted on the ground GSE for launch; 4) A free flyer with all transfer kit mounted and two gender bent GSE fittings.

If memory serves I suggested #1, and I now formally renounce it as 3 & 4 seem better ideas. Number 2 calls for extra mass and a small bit of extra plumbing and valving. IMO, 3 is a better approach. Number 3 requires an accumulator to have specialized hardware which makes it a depot. IMO this is the best approach if a formal depot is decided upon. Number 4 is much like what you suggested and IMO is the way to go if a specialized depot is to be avoided.

The idea of snaking a couple of hoses is not going to work. At minimum there also needs to be ullage transfer to minimize pump size and power, with the benefit of minimizing wastage. Past that, looking long term, it's not unreasonable to expect a transfer of any other fluids needed on later builds, and juicing up batteries has a certain appeal if a transfer campaign get drawn out. That and to my knowledge, fluid transfer has never been done between two bodies not firmly connected. Might as well take advantage of the GSE hardware that's already there and not get too fancy.

A minimal free flyer would be a boxy structure with two interconnected gender bent GSE plates, pumps and a small amount of propellant tankage and thrusters to allow it to maneuver onto the GSE plate of a tanker acting as an accumulator. The tanker would then take over maneuvering to mate the free flyer with another tanker for transfer. We'd see two SS's dorsal to dorsal with the free flyer connecting them at the GSE plates.

The debate on how tightly the two ships would need to be bound was ongoing with no consensus that I remember. This issue applies to all transfer schemes, not just the free flyer.

A possible strong point for the free flyer is it's central position offers a good location for ullage settling thrust. That might imply a larger gimbaled engine than that needed for its own maneuvering. Maybe not. A classic "it depends".

Moving up from a minimal free flyer adds PV, a radiator and cryocooler. Maybe MLI. This would call for a structure of some sort, maybe a telescoping truss, that would extend parallel to the ship(s) it's attached to. This has the advantage of offering more attachment points if they're needed.

A big disadvantage here is figuring out what to do with all the foldy stuff when two ships are involved. PV and radiators could probably be swing out of the way but MLI would be a challenge, but maybe not needed if the cryocooler is sized big enough, which leads us back to the current discussion on expected boiloff.

A free flyer could stay attached to one tanker for use over many refueling campaigns or be bought back into the cargo bay for return after one campaign. They might have to jettison the floppy stuff for a return. James Web was awesome but can you imagine the teams reaction if they got orders to design for refolding and redeploy?

I've had a dog in this fight but I've tried to keep this summary as dispassionate and neutral (with some identified opinion interjected) as possible. Summaries are useful in pulling together and clarifying discussions scattered through many pages. If I've missed, I apologize. Corrections welcomed.

[Greg Hullender]

Thanks! That was hugely helpful. For the first time, I think I actually understand a few of the terms I saw in discussions further upthread.

I am opposed to something that stays inside a cargo bay. It could work if the ship carrying it is nosed up between the tanker and the receiving ship and just acts as a fuel pump but that has safety issues for a crewed receiver.

What are the safety issues? Is there a real concern that a vehicle being refilled might explode?

An important point you addressed is GSE plate gender mating. The four solutions so far discussed were: 1) the tanker deploying a gender bender on top of its GSE plate; 2) a tanker having a second, gender bent GSE plate; 3) the tanker being already gender bent and a temporary adapter mounted on the ground GSE for launch; 4) A free flyer with all transfer kit mounted and two gender bent GSE fittings.

Hmm. I actually visualized a "free flyer" depot with two sets of fittings: one male and one female. My thought was that you'd fill it up through the female fittings and it would fill a Starship via the male fittings. A basic depot would get filled on the ground before loading into the cargo bay. An advanced depot would get repeated loads of fuel while in orbit and then (when full) would unload into a Starship. There would never be a time when two Starships were connected directly to each other. (Although the case where the depot is secured inside the cargo bay is, admittedly, not very different from that.) You could even chain multiple depots together, if there were any advantage to be had from that. (E.g. greater capacity but sharing the same power source/coolant or something.)

The idea of snaking a couple of hoses is not going to work. At minimum there also needs to be ullage transfer to minimize pump size and power, with the benefit of minimizing wastage. Past that, looking long term, it's not unreasonable to expect a transfer of any other fluids needed on later builds, and juicing up batteries has a certain appeal if a transfer campaign get drawn out. That and to my knowledge, fluid transfer has never been done between two bodies not firmly connected. Might as well take advantage of the GSE hardware that's already there and not get too fancy.

I guess I'm just concerned that the GSE hardware either won't be long enough to reach or else it'll protrude too much at launch time. Perhaps that's a silly thing to worry about, but I've had trouble visualizing how it would actually work. Likewise, the free flyer with two male GSE adaptors that bridges two Starships troubles me because I don't see how it ever gets into place without someone in a space suit moving it around. Of course, given the name, I guess it could have its own propulsion and guidance system, but, man, that seems really complicated. At that point, it really is a vehicle. (It's a beautifully symmetrical solution, though.)

Thanks again for your clear and helpful responses.

[Paul451]

Delete the tanks, use Starship's. Now you just two pumps and two hoses. That's a win.

But is that a win? Now you've got another kind of Starship--one that can't be used for any other purpose. Is that desirable?

Orbital refuelling seems to be a core part of Starship's standard operating procedure. Tankers are therefore going to get plenty of use. There's no savings in having a special payload-bay tank that goes in a regular cargo Starship. Even if there was an advantage to having separate tanks (simplified plumbing, say) then it would still make sense to have dedicated tankers that have that plumbing built in.

Personally, I like box-in-a-box. There's a reason shipping containers revolutionised transport. But it's against SpaceX's design philosophy, and dedicated tankers aren't going to be sitting around wasted, so there's no reason for it.

[Paul451]

I am opposed to something that stays inside a cargo bay. It could work if the ship carrying it is nosed up between the tanker and the receiving ship and just acts as a fuel pump but that has safety issues for a crewed receiver.

What are the safety issues? Is there a real concern that a vehicle being refilled might explode?

A loss of pressure during refuelling could cause a BLEVE.

BLEVE's are bad.

[Paul451]

What I would say is that, for a Starship in LEO, there is no completely passive ZBO solution. Not at the moment, anyway. There must be some cryocooler, with all the extra hardware that implies. The best that Solar White can do for you is reduce the mass of the cryocooler.

Or use boil-off. [...]

Maybe I'm missing something, but how is this different from just doing nothing?

I was reacting to "There must be some cryocooler". Not necessarily. Just because you can't passively cool to ZBO, doesn't mean you can't passively reflect enough sunlight/Earthlight to reach acceptable levels of boil-off.

The choice isn't "reflect/emit until ZBO" or else "you must use a cryocooler". There's a third choice, "propellant is cheap, reflect enough until you can tolerate the boil-off losses". There's also a continuum with the cryocooler, where you tolerate a certain amount of boil-off to reduce the demand on the cryocooler, or as part of the cooling loop. (For example, using methalox boil-off to dump heat from a LOx-only cryocooler. Effectively an open-cycle radiator.)

Hence you are trading "reflective coatings", "MLI", "cryocooler", "boil-off". The mass optimum would be some combination. The development-cost optimum for first-gen depots would favour non-MLI reflective coatings plus boil-off.

(Aside: In the boil-off scenario, venting wouldn't be passive, however. You want to actively pump boil-off gas into a low pressure expansion chamber that lets it boil-off at lower temperatures, to avoid having the entire propellant load hovering around the boiling point. This also allows you to cool the LOx via just methane boil-off. (Methane's latent heat content is greater than oxygen's, so it's more mass-efficient.) Thankfully, the energy calcs are the same.)

[OTV Booster]

Thanks! That was hugely helpful. For the first time, I think I actually understand a few of the terms I saw in discussions further upthread.

I am opposed to something that stays inside a cargo bay. It could work if the ship carrying it is nosed up between the tanker and the receiving ship and just acts as a fuel pump but that has safety issues for a crewed receiver.

What are the safety issues? Is there a real concern that a vehicle being refilled might explode?

An important point you addressed is GSE plate gender mating. The four solutions so far discussed were: 1) the tanker deploying a gender bender on top of its GSE plate; 2) a tanker having a second, gender bent GSE plate; 3) the tanker being already gender bent and a temporary adapter mounted on the ground GSE for launch; 4) A free flyer with all transfer kit mounted and two gender bent GSE fittings.

Hmm. I actually visualized a "free flyer" depot with two sets of fittings: one male and one female. My thought was that you'd fill it up through the female fittings and it would fill a Starship via the male fittings. A basic depot would get filled on the ground before loading into the cargo bay. An advanced depot would get repeated loads of fuel while in orbit and then (when full) would unload into a Starship. There would never be a time when two Starships were connected directly to each other. (Although the case where the depot is secured inside the cargo bay is, admittedly, not very different from that.) You could even chain multiple depots together, if there were any advantage to be had from that. (E.g. greater capacity but sharing the same power source/coolant or something.)

The idea of snaking a couple of hoses is not going to work. At minimum there also needs to be ullage transfer to minimize pump size and power, with the benefit of minimizing wastage. Past that, looking long term, it's not unreasonable to expect a transfer of any other fluids needed on later builds, and juicing up batteries has a certain appeal if a transfer campaign get drawn out. That and to my knowledge, fluid transfer has never been done between two bodies not firmly connected. Might as well take advantage of the GSE hardware that's already there and not get too fancy.

I guess I'm just concerned that the GSE hardware either won't be long enough to reach or else it'll protrude too much at launch time. Perhaps that's a silly thing to worry about, but I've had trouble visualizing how it would actually work. Likewise, the free flyer with two male GSE adaptors that bridges two Starships troubles me because I don't see how it ever gets into place without someone in a space suit moving it around. Of course, given the name, I guess it could have its own propulsion and guidance system, but, man, that seems really complicated. At that point, it really is a vehicle. (It's a beautifully symmetrical solution, though.)

Thanks again for your clear and helpful responses.

You're right about the safety issue with kit in nose design. I was having cranial flatulence. Still, it adds another GSE hookup and has to pump up hill. The plumbing has to reach down to the main tanks. I just don't see any advantage over the dedicated depot or the free flyer except for deploying, and enough downside to give it the hairy eyeball.


I wouldn't worry too much about a free flyer having guidance and control issues. It needn't move fast, so not much dV is needed. Aligning itself and attaching doesn't call for anything a Dragon doesn't already do routinely. The computers they use for engine controllers would probably be overkill. The optics could be GoPro.


Interesting factoid: one ton of propellant takes up 0.016m of tank height. A nominal 150t of propellant would need a 2.4m tank stretch. Just a matter of moving domes, wiring, plumbing and vents. This would make the tankers their own variant and save the weight and operational inconvenience of cargo bay tanks. Some see that as a problem. I don't. Opinion: Hardware rich won't end when operations start. Tanker use will be slow at first but my money says use will pick up as the utility proves itself. A 40t probe to the Oort cloud? No problemo.


The free flyer would need two of one type of GSE plate. It would exactly mimic the plate on the ground GSE, which would be exactly what the ship GSE plate expects to mate up with. The gender of a quick disconnect doesn't dictate flow direction. Gas station air line QD's line up the way they do for the convenience of having a barb on the air chuck or impact driver instead of a coupler.


There was discussion about how much of a stand-off would be needed between two ships. Opinions on the high end were a couple of meters. On the low end, zip. I'd say the minimum with the free flyer would be however much it would take to route all the internal crossovers, L to R and R to L, and pumps. Maybe around 1.5m? For a depot, maybe mount the gender bent QD on a pantograph with flex hose behind it for some minimal clearance. This would be too mushy for a firm connection between the ships so either another mechanism or some struts or something.


But then there were those who felt that a firm connection wasn't necessary...

[Twark_Main]

I already ran the numbers. That’s in my post. It’s just proportional to absolute temperature to the fourth power. 5.67e-8W/(m^2*Kelvin^4) is the Proportionality constant, the Stefan-Boltzmann constant.

I'm doing something different, actually. See upthread.

Ok, so I ran the numbers and it's confirmed. At both LOX and CH4 boiling points, there's no advantage to using a selective surface assuming you're in LEO and your surface has the average Earth view factor.

Curves for LOX (111.2 K)

Curves for CH4 (138.4 K)

The upper curve is the combined solar + Earth IR incoming spectrum, modeled as blackbodies. The lower curve is the blackbody emission spectrum. Horizontal axis is wavelength in nanometers, vertical axis is arbitrarily scaled but area under the curves is proportional to power per square meter.

As you can see, the curves never cross. That means there's no upper wavelength cutoff beyond which a totally emissive blackbody would "beat" a totally reflective surface.



Btw SPHEREx is precisely the special case I described earlier: the cone always points up, so it never has a view of the Earth. It also needs to be located in a dusk/dawn polar orbit, such that it never has a view of the Sun. There are downsides to using a disk/dawn orbit for a depot: it's less efficient than launching eastward, and it has less flexibility for targeting a specific right ascension for departure.

[Greg Hullender]

As you can see, the curves never cross. That means there's no upper wavelength cutoff beyond which a totally emissive blackbody would "beat" a totally reflective surface.

Does such a surface exist, though?

[Greg Hullender]

You're right about the safety issue with kit in nose design. I was having cranial flatulence. Still, it adds another GSE hookup and has to pump up hill. The plumbing has to reach down to the main tanks. I just don't see any advantage over the dedicated depot or the free flyer except for deploying, and enough downside to give it the hairy eyeball.

I still don't think I'm clearly conveying what I'm visualizing. Let me give it another try. First, remember that I'm assuming no modifications to the cargo Starship whatsoever. It has a female GSE hookup down at the bottom of the tanks and that's all. There is no extra plumbing running up to the nose of it.

The basic depot has a male GSE hookup. It fits in the cargo hold of a standard starship and is firmly braced in there. It can move out a little bit (relative to the cargo hold), probably on rails or something, so it's always firmly attached. It does not have any external plumbing.

When cargo starship A refills starship B, A opens the cargo doors, lines up the depot GSE with B's GSE and connects them. Once the cargo doors are open, the depot rolls out, and the two GSE's can mate without any other contact between the two vehicles.

Both vehicles do a gentle ullage burn while the propellants are transferred. Because the docking is asymmetrical, the ullage burns will have to be calculated to generate net zero torque. Note, though, that even a symmetrical docking will have this problem because the masses of the two starships won't be the same and the center of mass will change dynamically during the refilling operation.

After the depot is depleted, they disconnect and the depot retracts. A goes home, and B goes on its mission. (Assuming it only needed one load.)

As a result, there's no plumbing that reaches down to the main tanks; they are refilled from the bottom, just as they are on the launch pad.

Is that clearer?

Interesting factoid: one ton of propellant takes up 0.016m of tank height. A nominal 150t of propellant would need a 2.4m tank stretch. Just a matter of moving domes, wiring, plumbing and vents. This would make the tankers their own variant and save the weight and operational inconvenience of cargo bay tanks. Some see that as a problem. I don't. Opinion: Hardware rich won't end when operations start. Tanker use will be slow at first but my money says use will pick up as the utility proves itself. A 40t probe to the Oort cloud? No problemo.

I just can't help thinking that if you're trying to save money by mass-producing these things, it will help a lot to have fewer versions. Right now I'm seeing a cargo version, a crew version, a tanker version, a lunar version, etc. and that just seems like a problem to me. Obviously it's easier if the lunar version is just the crew version minus the last steps (that attach tiles), and maybe the cargo version and the crew version share all the work up to the point where the upper part of the vehicle is configured, but the tanker version seems like a very different work flow. Also, there probably won't be very many lunar and crew versions built, so maybe it's okay that they're expensive to build; almost everything will be cargo starships, after all. But you'd likely need a lot of tankers.

Hence my thinking that it's a win if the basic depot is as simple as possible and just fits into the cargo hold. Again, I'm no expert here--just trying to think the whole thing through.

The free flyer would need two of one type of GSE plate. It would exactly mimic the plate on the ground GSE, which would be exactly what the ship GSE plate expects to mate up with. The gender of a quick disconnect doesn't dictate flow direction.

Yeah, I realize that. I had been thinking you'd want a female GSE on any depot just so you could fill it on the ground using the same hardware that fills a starship, but perhaps that's unnecessary. Also, it would let you chain multiple depots together--if that had any value. It also occurred to me that it might be easier to make pumps that only work in one direction, but that's sheer speculation on my part.

In that case, though, why does a depot need two GSE's? Why not just one male one that it uses to fill up from one starship (or on the ground) and then later pump the propellant into another one? It only seems you'd need two attachments if you wanted to be connected to two starships at the same time, but with a depot, I don't think you'd ever want to do that.

If a single refill (of 150 tons) is sufficient for the lunar missions, then a basic depot is all that would be required for a while. But I'm still fascinated by the idea of a "jumbo depot" that could hold much more propellant indefinitely.

Using your 0.016 m/mT figure, it would seem that an empty jumbo depot that filled the 18m cargo bay could hold 1125 tons of propellant. It would take 7.5 flights to fill such a thing, and three such depots could completely refill a starship (if there were a reason to do that--maybe if you wanted to take months hauling propellant up so you could quickly fill up a fleet of starships all heading to Mars in the same launch window).

If we were going to construct such a "depot complex," I could imagine dividing the functionality between "storage modules" (just tanks and plumbing) and a single "command module." The command module would have solar panels, cryocooler, radiator, communication links, ullage engine etc. (but minimal fuel tanks) that would attach to one or more storage modules.

But that's starting to be pretty pie-in-the-sky. :-)

[Hog]

BLEVE=Boiling Liquid Expanding Vapour Explosion

[jimvela]

As you can see, the curves never cross. That means there's no upper wavelength cutoff beyond which a totally emissive blackbody would "beat" a totally reflective surface.

Does such a surface exist, though?

No, there's no such thing as a totally emissive blackbody nor a totally reflective surface.

I don't think the analysis in the thread has touched on all the available coatings. 
Just from my own experience, for example, there are things like BIRB which is used on cryogenic telescope missions.

I'm sure whomever does the work on the Starship in-space cryogenic systems will be aware of these kinds of options.

[Comga]

(snip)
Btw SPHEREx is precisely the special case I described earlier: the cone always points up, so it never has a view of the Earth. It also needs to be located in a dusk/dawn polar orbit, such that it never has a view of the Sun. There are downsides to using a disk/dawn orbit for a depot: it's less efficient than launching eastward, and it has less flexibility for targeting a specific right ascension for departure.

Jumping in late to this discussion, so apologies if this has been said before.
The dusk-dawn SSO orbit is a limiting case, one that is particularly well suited to all-sky surveys like SPHEREX and WISE.
(And particularly unsuited to a depot from which to start interplanetary trajectories, as you said.)
There, the angle from the local vertical to the Sun is always near 90 degrees.  (probably >90 minus 23.5 degree inclination of the Earth's rotation with respect to its orbit minus ~7 degrees from the inclination to the pole.)
However, the same goal of never pointing the radiator at the Sun can be achieved at lower inclinations for which the precession does not match the orbital rate of the Earth, i.e. with a variable beta angle.

Take the other extreme, an equatorial orbit.
A depot in equatorial orbit oriented to the Local Velocity, Local Horizon (LVLH) would need an Earth shield and a sunshield.
This could be cylindrical, with axis pointed at the orbit poles, perpendicular to velocity and to the local zenith.
A cryoradiator could point out either end.
However, it's Field Of View would be limited to 90 - 23.5 =~66 degrees from the axis, to prevent the Sun from shining into it regardless of season.
(The beta angle is zero at the equinoxes, when the pane of the equator passes through the Sun, and grows to ~22 deg at the solstices.)
Such radiators are commonly made with "Etendue converters", non-imaging optics, sometimes implemented with Winston cones.  (See some one dimensional implementations here.)

For higher inclination orbits, the angle is reduced, and by conservation of Etendue, the far end of the "cone" gets proportionally larger.
So it's a trade between making the inclination high enough to go through the departure points and restricting the size.

So it's not just the spectral properties that influence the solar and Earth heating and the thermal radiation. 
The geometry is also important.

[OTV Booster]

You're right about the safety issue with kit in nose design. I was having cranial flatulence. Still, it adds another GSE hookup and has to pump up hill. The plumbing has to reach down to the main tanks. I just don't see any advantage over the dedicated depot or the free flyer except for deploying, and enough downside to give it the hairy eyeball.

I still don't think I'm clearly conveying what I'm visualizing. Let me give it another try. First, remember that I'm assuming no modifications to the cargo Starship whatsoever. It has a female GSE hookup down at the bottom of the tanks and that's all. There is no extra plumbing running up to the nose of it.

The basic depot has a male GSE hookup. It fits in the cargo hold of a standard starship and is firmly braced in there. It can move out a little bit (relative to the cargo hold), probably on rails or something, so it's always firmly attached. It does not have any external plumbing.

When cargo starship A refills starship B, A opens the cargo doors, lines up the depot GSE with B's GSE and connects them. Once the cargo doors are open, the depot rolls out, and the two GSE's can mate without any other contact between the two vehicles.

Both vehicles do a gentle ullage burn while the propellants are transferred. Because the docking is asymmetrical, the ullage burns will have to be calculated to generate net zero torque. Note, though, that even a symmetrical docking will have this problem because the masses of the two starships won't be the same and the center of mass will change dynamically during the refilling operation.

After the depot is depleted, they disconnect and the depot retracts. A goes home, and B goes on its mission. (Assuming it only needed one load.)

As a result, there's no plumbing that reaches down to the main tanks; they are refilled from the bottom, just as they are on the launch pad.

Is that clearer?

Interesting factoid: one ton of propellant takes up 0.016m of tank height. A nominal 150t of propellant would need a 2.4m tank stretch. Just a matter of moving domes, wiring, plumbing and vents. This would make the tankers their own variant and save the weight and operational inconvenience of cargo bay tanks. Some see that as a problem. I don't. Opinion: Hardware rich won't end when operations start. Tanker use will be slow at first but my money says use will pick up as the utility proves itself. A 40t probe to the Oort cloud? No problemo.

I just can't help thinking that if you're trying to save money by mass-producing these things, it will help a lot to have fewer versions. Right now I'm seeing a cargo version, a crew version, a tanker version, a lunar version, etc. and that just seems like a problem to me. Obviously it's easier if the lunar version is just the crew version minus the last steps (that attach tiles), and maybe the cargo version and the crew version share all the work up to the point where the upper part of the vehicle is configured, but the tanker version seems like a very different work flow. Also, there probably won't be very many lunar and crew versions built, so maybe it's okay that they're expensive to build; almost everything will be cargo starships, after all. But you'd likely need a lot of tankers.

Hence my thinking that it's a win if the basic depot is as simple as possible and just fits into the cargo hold. Again, I'm no expert here--just trying to think the whole thing through.

The free flyer would need two of one type of GSE plate. It would exactly mimic the plate on the ground GSE, which would be exactly what the ship GSE plate expects to mate up with. The gender of a quick disconnect doesn't dictate flow direction.

Yeah, I realize that. I had been thinking you'd want a female GSE on any depot just so you could fill it on the ground using the same hardware that fills a starship, but perhaps that's unnecessary. Also, it would let you chain multiple depots together--if that had any value. It also occurred to me that it might be easier to make pumps that only work in one direction, but that's sheer speculation on my part.

In that case, though, why does a depot need two GSE's? Why not just one male one that it uses to fill up from one starship (or on the ground) and then later pump the propellant into another one? It only seems you'd need two attachments if you wanted to be connected to two starships at the same time, but with a depot, I don't think you'd ever want to do that.

If a single refill (of 150 tons) is sufficient for the lunar missions, then a basic depot is all that would be required for a while. But I'm still fascinated by the idea of a "jumbo depot" that could hold much more propellant indefinitely.

Using your 0.016 m/mT figure, it would seem that an empty jumbo depot that filled the 18m cargo bay could hold 1125 tons of propellant. It would take 7.5 flights to fill such a thing, and three such depots could completely refill a starship (if there were a reason to do that--maybe if you wanted to take months hauling propellant up so you could quickly fill up a fleet of starships all heading to Mars in the same launch window).

If we were going to construct such a "depot complex," I could imagine dividing the functionality between "storage modules" (just tanks and plumbing) and a single "command module." The command module would have solar panels, cryocooler, radiator, communication links, ullage engine etc. (but minimal fuel tanks) that would attach to one or more storage modules.

But that's starting to be pretty pie-in-the-sky. :-)

I'd like to do the embedded quotes but between fat fingers and a small phone... Bear with me.

Sorry to be so dense. IIUC, the cargo/tanker has GSE up in the nose along with propellant tanks. It noses up to the QD plate on the receiving ship, its QD plate extends, and it pumps over the propellant from the cargo bay tanks. Is this right?

The nominal 150t is more a placeholder based on last stated cargo capacity to LEO (100t?) than a firm number. If I've got your conops right, total propellant load would be the nominal 150t minus the mass of the internal tankage, pumps, QD plate and plate extension mechanism - for every cargo/tanker load. A bespoke tanker would have the same size and design as a cargo ship. The only difference would be the top and common dome slid up taking up some of the cargo space. That and longer wiring harnesses and plumbing.


The actual numbers will change as the ships capabilities evolve. Twark or Rad Mod or maybe Robo ran some numbers awhile back claiming that a tanker with a small stretch could load a total of 1600 or 1650t of propellant and deliver quite a bit more than 150t. A dedicated depot would most probably be a tanker variant with extra kit so this plays into supporting early Artemus missions where the extra propellant can be put to good use. Don't ask. It was a very detailed discussion.

As spaceship modifications go, moving the domes is no big thing. The number of rings per barrel would change but the assembly techniques stay the same and in the same order. Less difficulty than a pickup assembly line accommodating the occasional crew cab or dually rear end, short bed/long bed. The extra cost is offset to some degree by not needing a chomper or cargo hatch. The upside is less parasitic mass. The downside is dedicated use. Long term, this strikes me as a good tradeoff. An actual stretch would be a bit more difficult but not outrageously so.

The multi use cargo/tanker looks like a good proof of concept and maybe even an early operational architecture but not so good long term. Which points up an ongoing problem here at NSF. We often do not differentiate between short term and long term vision. Talking about SX makes it worse because they have a bad habit of changing plans in the blink of an eye and that pie in the sky comes on fast.


There is one point in your cargo/tanker idea where safety would be a concern. This may be where my brain fart happened. If a crewed ship needs refueling it should only happen once. Multiple small transfers each carry the same safety risks as one big gulp. This can be got around by waiting for the transfers to be finished before the crew launches and transfers over to the refueled ship. Another complexity, but maybe the best way.


You're correct in expecting the depot to have what you call female and what I call gender bent GSE. My terminology comes from electronics adapters (called gender benders) used to reverse a cables gender. This 'off gender' GSE connection complicates ground handling. Hey, that suckers does have to be fueled and launched. The solution that IMO, solves the problem best is to put an adapter (a gender bender) on the ground side GSE for depot launches only. Put it on. Launch. Take it off. The extra mass stays on the ground and most likely only a couple hours labor at each end. Depot launches will be rare so this wouldn't be much overall system impact.

A depot would have only the one QD plate. It's the free flyer that would have two of the same gender as the ground side GSE.  One side would mount to a tanker with standard QD plate that becomes a temporary accumulator. The other would mate to the tankers (also standard QD plate) arriving to fill the accumulator, and ultimately to the receiving ship (also with standard QD plate). Keep in mind, the free flyer and depot are not complimentary. The are competing concepts with the same goal.


What you call a command module is pretty much what the free flyer would be. It packs in whatever functionality a depot needs but allows plain Jane tankers to become storage tanks (accumulators). If you build the command module into a tanker you have - a depot. The free flyer concept that's been thrown around would be a new design with all that implies.


One free flyer variant gives it enough tankage for substantial dV so it can shift to where the next traffic is expected. Another has it staying on the tanker to hitch a ride to the next orbit. A third packs it back into a cargo bay for return and relaunch. They all have strengths and weaknesses. IMO it's way too early for hard decisions in depot vs free flyer, let alone the variants. Just gonna set back and enjoy the show.

[Greg Hullender]

Personally, I like box-in-a-box. There's a reason shipping containers revolutionised transport. But it's against SpaceX's design philosophy, and dedicated tankers aren't going to be sitting around wasted, so there's no reason for it.

Another reason to go with "box-in-a-box" is that it would let them refill vehicles that used different kinds of fuel. If they draw straight from the tanks of a special tanker starship, then the only thing they can ever refuel is something that uses CH4 and O2. But if the propellants are carried in separate tanks in the cargo hold, then they could fill up (in theory) a vehicle that used H2 and O2, or a nuclear-thermal vehicle that just used H2 or an ion drive that wanted xenon or krypton or whatever. Obviously the depots would be different, but the cargo starships would be the same.

Of course, that assumes there will ever be very many vehicles like that needing to be refueled.

[Greg Hullender]

IIUC, the cargo/tanker has GSE up in the nose along with propellant tanks. It noses up to the QD plate on the receiving ship, its QD plate extends, and it pumps over the propellant from the cargo bay tanks. Is this right?

Maybe I've been messing the terms up. I think QD is the "quick disconnect," which is the female socket that's at the bottom of a Starship. GSE is the "ground support equipment." This is the male plug that goes into the QD when a starship is filled on the ground. A "gender bender" is an adaptor that's male on one side and female on the other, with plumbing crossed over as needed to make it work. (Otherwise it'll be the mirror image of what you want on one side or the other.) Are we in agreement here? (Sorry if I've been mixing them up before.)

Given that, I've been visualizing the tanker starship as having GSE instead of a QD at the bottom. They attach a gender bender on the ground to fill it up, but they pull that off and the GSE needs to retract into the starship before it launches. In space, it extends the GSE and mates with the QD on a standard starship.

Are we still together?

If you want one tanker to refill another, you'll have to somehow attach a gender bender to one of them. This is what a "free flyer" is good for. (Among other things)

For what I was envisioning, the "depot" in the cargo hold would consist of two small tanks plus GSE. You'd fill these tanks the same way you would a tanker starship. In space, the cargo door would open, the GSE would extend, and it would mate with the QD on the target vehicle.

The actual numbers will change as the ships capabilities evolve. Twark or Rad Mod or maybe Robo ran some numbers awhile back claiming that a tanker with a small stretch could load a total of 1600 or 1650t of propellant and deliver quite a bit more than 150t. A dedicated depot would most probably be a tanker variant with extra kit so this plays into supporting early Artemus missions where the extra propellant can be put to good use. Don't ask. It was a very detailed discussion.

I guess the real question is "how much more propellant?" If it's twice as much, then, yeah, I can see how it might be worth it. But if it's just 20% more, maybe not.

Perhaps I should go back through the earlier posts on the thread. I suspect they'd make a lot more sense to me now.

There is one point in your cargo/tanker idea where safety would be a concern. This may be where my brain fart happened. If a crewed ship needs refueling it should only happen once. Multiple small transfers each carry the same safety risks as one big gulp. This can be got around by waiting for the transfers to be finished before the crew launches and transfers over to the refueled ship. Another complexity, but maybe the best way.

Yeah, I thought about this too. That's where a bespoke depot seems to make the most sense. You fill it up--no matter how long it takes--and when it's got enough propellant, then you send up the crewed ship and fill it just once. That's the case where special reflectors and/or cryocoolers make sense as well--given how long the propellant has to survive in LEO--and all that extra hardware really does seem like it should be on a depot--not a starship.

You're correct in expecting the depot to have what you call female and what I call gender bent GSE. My terminology comes from electronics adapters (called gender benders) used to reverse a cables gender.

Yep. I've soldered RS232 cables for this purpose myself, although not since Jimmy Carter was president. :-)

A depot would have only the one QD plate. It's the free flyer that would have two of the same gender as the ground side GSE.  One side would mount to a tanker with standard QD plate that becomes a temporary accumulator. The other would mate to the tankers (also standard QD plate) arriving to fill the accumulator, and ultimately to the receiving ship (also with standard QD plate). Keep in mind, the free flyer and depot are not complimentary. The are competing concepts with the same goal.

What you call a command module is pretty much what the free flyer would be. It packs in whatever functionality a depot needs but allows plain Jane tankers to become storage tanks (accumulators). If you build the command module into a tanker you have - a depot. The free flyer concept that's been thrown around would be a new design with all that implies.

Ah, that makes sense, although the gender combinations are a little challenging. E.g. if you want one tanker to fill another, you need a twisted female-female free flyer. But if you want a tanker to fill a regular starship, you need a straight-through male-female one. And if you know you'll always have a free flyer, then you want to make all starships female and all free flyers twisted male-male.

IMO it's way too early for hard decisions in depot vs free flyer, let alone the variants. Just gonna set back and enjoy the show.

Sure, but it's a lot easier to enjoy the show if you start off knowing what language they're speaking! :-)

[Paul451]

The basic depot [...] fits in the cargo hold of a standard starship and is firmly braced in there.
[...]
When cargo starship A refills starship B

Just to clarify, the terminology of the thread (and the NASA HLS contract) is:

Tanker: Starship that launches fuel, either to an orbiting depot, or directly to another Starship to be refuelled.

Depot: An accumulator in orbit that is refuelled by a number of tanker-flights until it is full (or full enough). Then-and-only-then the mission vehicle launches and docks with the depot to take on a full (or full enough) propellant load for its mission.

What you've described as a "depot" is what we are calling a "tanker". And your model of operations lacks a depot and requires the mission vehicle to act as the accumulator on orbit. The model of operations described by the HLS contract has an accumulator (depot) in orbit being filled by a series of tankers.

This doesn't change your argument for a modular tanks and plumbing in a cargo-vehicle to replace dedicated tankers. But it isn't the "depot", it's a "tanker".

I just can't help thinking that if you're trying to save money by mass-producing these things, it will help a lot to have fewer versions. Right now I'm seeing a cargo version, a crew version, a tanker version, a lunar version, etc. and that just seems like a problem to me.

Meh, it's pretty standard with aircraft. The hard-part is the basic design. Variations are trivial by comparison.

But you'd likely need a lot of tankers.

That's why it makes sense to specialise.

You only want a box-in-a-box if the particular box rarely flies, because you don't want a vehicle that spends most of its time not earning money, unless it's earning NASA-money. That isn't the case for tankers, because refuelling is the standard-operating-procedure for Starship.

[OTV Booster]

Personally, I like box-in-a-box. There's a reason shipping containers revolutionised transport. But it's against SpaceX's design philosophy, and dedicated tankers aren't going to be sitting around wasted, so there's no reason for it.

Another reason to go with "box-in-a-box" is that it would let them refill vehicles that used different kinds of fuel. If they draw straight from the tanks of a special tanker starship, then the only thing they can ever refuel is something that uses CH4 and O2. But if the propellants are carried in separate tanks in the cargo hold, then they could fill up (in theory) a vehicle that used H2 and O2, or a nuclear-thermal vehicle that just used H2 or an ion drive that wanted xenon or krypton or whatever. Obviously the depots would be different, but the cargo starships would be the same.

Of course, that assumes there will ever be very many vehicles like that needing to be refueled.

Of course, that assumes there will ever be very many vehicles like that needing to be refueled.

LOL. Get in your way back machine to around 1915 and try to explain a megaplex truck stop with 800 parking spots, a movie theater, an 8 bay repair facility, a hotel, a full service restaurant, half a dozen fast food joints and three C stores. You'll be locked up as a lunatic.


Back to spaceships. Right now SX is concerned with refueling their own ship. We're talking 1200+ tons of propellant. This is the cake. Customers would be the icing.


Customers in the foreseeable future will need only a few tons. This is where tanks and a transfer rig in the cargo bay make sense. SS to orbit. Open the pod bay Hal. A self propelled mini depot emerges and lights off to service a customer. Think spy sats.


Refueling works best if the target is designed for it. If someone starts designing for it today it won't show up for three years minimum. In the real world, probably 3-4 years after SX shows it to work.

[AC in NC]

Of course, that assumes there will ever be very many vehicles like that needing to be refueled.

LOL. Get in your way back machine to around 1915 and try to explain a megaplex truck stop with 800 parking spots, a movie theater, an 8 bay repair facility, a hotel, a full service restaurant, half a dozen fast food joints and three C stores. You'll be locked up as a lunatic.

And of course:

I think there is a world market for maybe five computers.
-- Thomas Watson, IBM Founder, 1943

[Twark_Main]

As you can see, the curves never cross. That means there's no upper wavelength cutoff beyond which a totally emissive blackbody would "beat" a totally reflective surface.

Does such a surface exist, though?

Obviously not, but you're missing the point entirely.

First you find the optimum. Then you try to design a material that gets as close as possible to that optimum. But if you don't know the "ideal" first, you don't know what spectral properties you're aiming for.


Additionally, if even that "ideal" material isn't able to achieve your primary goal (for instance, passive ZBO), then you know it's impossible for any real material. You can therefore avoid investing millions of dollars in material science research looking for something that provably can't exist.

"Success being one of the possible outcomes is very important."

[Twark_Main]

I don't think the analysis in the thread has touched on all the available coatings.

That's the beautiful part about it. You don't have to exhaustively search through all available coatings.