Author Topic: Starship In-orbit refueling - Options and Discussion  (Read 229281 times)

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 4149
  • United Kingdom
  • Liked: 6025
  • Likes Given: 34
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1040 on: 07/05/2022 06:17 pm »
I'd just use the QD itself for that. It has alignment pins, it has driveable latches, and on the ship QD side it sits on the end of an 5-axis actuated (possibly 6-axis, not clear if the QD plate itself can 'roll' about its own short axis) arm. It seems rather redundant to add an additional arm with alignment and latching capability just to avoid using your arm with alignment and latching capability.

Assuming that the active QD is on the depot side, you can wind up pushing on a fairly light payload Starship (less than 200t).  How much force do you need to make it latch?  And what happens if you've given it a shove not anywhere near the CoM and it doesn't latch?  Seems like a fine way to spin the payload Starship's nose into the depot's tail.

At the very least, you need something to fend off the compressive spinning load in a latch failure.  Better still is something that can handle a tensile load, so you can jab at that sucker over and over if you need to.
The current latches appear to be purely rotary (i.e. no force imparted 'against' the plate). If a latching mechanism without active driving is required, the spherical tips of the alignment pins could pull double-duty by adding lateral detents such that fully seating the alignment pins adds a few newtons or tens or newtons of resistance to extraction.
Imparted rotary force does look like an issue in case of a 'bonk' by the QD extension mechanism. Maybe the dorsal side of the depot could inflate a few rows of concave bumpers during prox-ops so that any contact low velocity events are gentle bounces rather than hard impacts. Could even work for the stabilisation during settling thrust if the QD latches are strong enough to 'pull' the two vehicles together against the bumpers.
Quote
But then the arm needs some active articulation to find the grapple on the other side, QD or otherwise.  The nice thing about the booms is that you can have relatively sloppy prox ops and still wind up latched, because the cradling effect of four booms will steer the target toward the latches on the booms.  And even if you fail to latch, I don't think you can wind up with a rotational failure that can cause a collision.
The downside of the booms is that any misalignment rubs the effectors up and down the sides of the Starship, which are covered in tiles.
Some articulation may be needed for alignment, but it only needs to be fairly gross alignment (get the narrow tip of the alignment pin into the wide opening of the socket) which could be accomplished with a lockable complaint mechanism (rather than powered actuation).

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1041 on: 07/05/2022 07:10 pm »
This idea sure has moved ahead. No or minimal separation between ships. An extendable QD plate in the payload bay and the ships hugging in a particularly erotic way. Maybe an arm boom or two between the ships.

Unless both ships are capable of RCS there is zero chance of thrust passing exactly through the CoM except maybe momentarily.   Without some thrust mods somewhere this sounds like a way to move orbital mechanics from deterministic to probabilistic. With both ships pointing in the same direction this becomes a non problem with no further mods needed.

With no standoff between ships there is zero margin for screwup. In my experience screwups sometimes happen no matter how good the planning. The trick is to have enough margin to give a reasonable chance of recovery.

This design moves from a mostly static QD on a standoff blister, admittedly needing a one time gender bender install, to a QD designed to extend from a mostly removed payload bay.

A point making this catawompus setup a little more realistic is that the depot side extendable QD plate and any booms need not be stowed after deploy. They'll be needed again.

The only advantage I can see in catawompus mounting is avoiding the dreaded gender/layout problem when the QD plates face each other and/or the difficulty in deploying a gender bender. The attached drawing shows a simple way around this.

The gender bender would stow for launching on the hull directly above the QD plate. The gender bender could have some streamlining on top but if possible, it should remain open faced to facilitate deploy. The two arms would be aligned such that they would swing the gender bender exactly onto the QD plate and be locked into place using standard QD locking hardware. The gender bender would have to duplicate this hardware in order to ply its trade. The deployment would be powered by a couple springs concentric to pivot points on the arms. The gender bender can be built arbitrarily deep, consistent with launch aerodynamics, to act as a standoff.

The biggest downside to this that I can come up with is the deployment arms would hinder the little bit of QD slop needed for final alignment. Sloppy hinges on the deploy arms might do. Mass should be the same or less than the payload bay QD. Side thought: if the payload bay QD has virtues I've not identified and ends up making sense, just build it into the ships hull like the main QD, and put a pantograph behind it. Deploy would be easier.

Something to keep in mind is that while micro g is most probably adequate to keep propellants settled, something more would most probably be needed to get it initially settled in an acceptable time frame. This seems to imply symmetrical RCS, which in turn implies the ships pointing in the same direction or some new RCS mods.

The idea of a boom or booms makes sense. It seems like a minimalist approach to adding just a little bit more to give a lot of margin. One good RCS belch might make for a bad day otherwise. All SS's have a place to be grabbed by the chopsticks. Two brace booms stowed on either side of the depot, would do the job. Make them telescoping with ullage powered double acting pistons and it has both mobility in that dimension and shock absorber functionality. If the chopstick grab points calls for grossly oversized brace ends, cobble up that suggested universally installed grab point.

ISTM the most reasonable place to stow the braces would be inside the strakes (what are we calling them now?) that the nonexistent fins mount through. A one time latch on a long door lets each brace deploy with the base end already where it needs to be for a wide stance. Motors on the base give two further axis of freedom. The grabber end may or may not need further degrees of freedom. Sure is starting to sound like a low end mini Canadarm.


Edit: went back and saw the notional payload bay QD is mounted in the wall.
« Last Edit: 07/05/2022 07:39 pm by OTV Booster »
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1042 on: 07/05/2022 07:43 pm »
I'd just use the QD itself for that. It has alignment pins, it has driveable latches, and on the ship QD side it sits on the end of an 5-axis actuated (possibly 6-axis, not clear if the QD plate itself can 'roll' about its own short axis) arm. It seems rather redundant to add an additional arm with alignment and latching capability just to avoid using your arm with alignment and latching capability.

Assuming that the active QD is on the depot side, you can wind up pushing on a fairly light payload Starship (less than 200t).  How much force do you need to make it latch?  And what happens if you've given it a shove not anywhere near the CoM and it doesn't latch?  Seems like a fine way to spin the payload Starship's nose into the depot's tail.

At the very least, you need something to fend off the compressive spinning load in a latch failure.  Better still is something that can handle a tensile load, so you can jab at that sucker over and over if you need to.

Quote
If you really need an extra connection point (to limit QD loads to pure shear rather than shear + torsion) a single point could be added for nose-to-tail at the other end. Store a short supplementary arm of very limited capability (only enough articulation to pack away) inside the engine skirt of the depot ship, as in the 'cargo boxes' seen on old renders. Flip this arm out and latch to a hardpoint on the dorsal side of the nose - a lasso around the nose is attractive from mechanical simplicity but risks tile damage.
Less elegant would be to add the arm to the nose section of the receiving Starship (mass added to all vehicles rather than just the depot) and have it align with and latch to the otherwise unused QD plate (with unoccupied alignment pin sockets and latch bars) on the depot ship.

But then the arm needs some active articulation to find the grapple on the other side, QD or otherwise.  The nice thing about the booms is that you can have relatively sloppy prox ops and still wind up latched, because the cradling effect of four booms will steer the target toward the latches on the booms.  And even if you fail to latch, I don't think you can wind up with a rotational failure that can cause a collision.
I don't think it matters which side is active and which side is light. The force is what it is. The relative mass is what it is. Low mass moves a lot. High mass, not so much.
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1043 on: 07/05/2022 09:09 pm »
I'd just use the QD itself for that. It has alignment pins, it has driveable latches, and on the ship QD side it sits on the end of an 5-axis actuated (possibly 6-axis, not clear if the QD plate itself can 'roll' about its own short axis) arm. It seems rather redundant to add an additional arm with alignment and latching capability just to avoid using your arm with alignment and latching capability.

Assuming that the active QD is on the depot side, you can wind up pushing on a fairly light payload Starship (less than 200t).  How much force do you need to make it latch?  And what happens if you've given it a shove not anywhere near the CoM and it doesn't latch?  Seems like a fine way to spin the payload Starship's nose into the depot's tail.

At the very least, you need something to fend off the compressive spinning load in a latch failure.  Better still is something that can handle a tensile load, so you can jab at that sucker over and over if you need to.
The current latches appear to be purely rotary (i.e. no force imparted 'against' the plate). If a latching mechanism without active driving is required, the spherical tips of the alignment pins could pull double-duty by adding lateral detents such that fully seating the alignment pins adds a few newtons or tens or newtons of resistance to extraction.
Imparted rotary force does look like an issue in case of a 'bonk' by the QD extension mechanism. Maybe the dorsal side of the depot could inflate a few rows of concave bumpers during prox-ops so that any contact low velocity events are gentle bounces rather than hard impacts. Could even work for the stabilisation during settling thrust if the QD latches are strong enough to 'pull' the two vehicles together against the bumpers.
Quote
But then the arm needs some active articulation to find the grapple on the other side, QD or otherwise.  The nice thing about the booms is that you can have relatively sloppy prox ops and still wind up latched, because the cradling effect of four booms will steer the target toward the latches on the booms.  And even if you fail to latch, I don't think you can wind up with a rotational failure that can cause a collision.
The downside of the booms is that any misalignment rubs the effectors up and down the sides of the Starship, which are covered in tiles.
Some articulation may be needed for alignment, but it only needs to be fairly gross alignment (get the narrow tip of the alignment pin into the wide opening of the socket) which could be accomplished with a lockable complaint mechanism (rather than powered actuation).
I have to agree that QD latching isn't like sticking a chuck into a service station air line. Most probably rotating and actively powered as you say. A collet, actively clamped onto a simple shaft by a concentric motor would have a lot of tensile strength, and drawing the clamped collets in via an arm waving mechanism would allow an even distribution of force across the QD plate that would be independent of wear. Disconnect might be a problem as lathe collets often need the workpiece to be lightly tapped with a hammer. Maybe the arm waved draw mechanism can do double duty.


ISTM that using the QD plate to do anything more than float a bit to deal with that last bit of misalignment would be asking for trouble. It would be the fulcrum of a very long lever with cryo plumbing on the back side. The shuttle admittedly did this sorta kinda when it docked to the ISS but it didn't cycle the connection every 12-24 hours. How much time did the shuttle spend lining up with the docking mechanism and shedding torque? If it takes a day it won't work for this application.


Unless there is a separate deployment mechanism, any booms would have to be self deploying from their storage location. If there is danger of boom misalignment leading to ANY type of damage there is an argument for Active Boom Control (ABC, my first TLA). ABC doesn't have to have particularly powerful motors on its base or end effector but it will need a bit of power for extension and retraction. With two booms as describe up thread, once the booms have attached to the non-depot ship, and the QD is connected, just reel the booms in with all pivots free floating. Geometry will rule and as the geometry (trig really) hits the intended location, lock the joints and slightly pressurize the telescoping booms to act as final bracing.


Note: all of this is independent of the QD being fore or aft.

We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1044 on: 07/05/2022 10:13 pm »
I just realized that all we've been discussing on QD and physical hookup is dependent on the min and max power of the RCS system. Somebody with more skills than me might be able to rough in a max based on 6bar exiting through an aperture of some size then fudging for the vents that exit into the dryer vent covers. Are there other vents for attitude control and non longitudinal translation?


What about minimums? All this impacts everything from matching orbits, maneuvering for physical hookup and ullage settling.


The thrust impact between a butterfly valve at 89deg and 90deg is insignificant. The difference between 1deg and 1.25deg would be huge. I don't doubt that sub degree control at a high level of precision is possible, but at what cost?


One solution might be to do coarse correction at 6bar, allowing the pressure to bleed down for the finer work. That in turn might require pressure be built back up to facilitate propellant transfer. How many kg of methane for each bar? Of course the O2 should be untouched by all this.


Grumble, grumble. Hot gas thrusters for the depot might end up being more props stingy with the benefit tighter control at a lower cost.
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Online oldAtlas_Eguy

  • Senior Member
  • *****
  • Posts: 4970
  • Florida
  • Liked: 4474
  • Likes Given: 827
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1045 on: 07/05/2022 11:01 pm »
Most of the RCS designs use a binary position valve and not a variable position valve. Thus the impulse control min is how fast can the valve cycle from from closed to open and back to closed. There are current valves used on spacecraft that can operate in pulse mode for pulses as fast as 160ms. Thus for a 100kg thruster max (continuous on) to min (single pulse). The min from such a thruster is 16kgf divided into the mass of the vehicle. The nicety is that the regulation of how much thrust is given is easily controlled. Which is why for monopropellants such valves are commonly used. Even though the min on the example case would be 16kgf. The ability of the fine variants between that and full on is easily much much smaller increments since it is simple a matter of adjusting the on time by a few ms. Which can give a granularity/accuracy of control of >100 to 1 without much difficulty.

In general your primarily haggling over already solved problems for large spacecraft. Note Shuttle was a 100t+ vehicle that had extremely fine RCS control and did not need most of the things being discussed.
« Last Edit: 07/05/2022 11:06 pm by oldAtlas_Eguy »

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 4149
  • United Kingdom
  • Liked: 6025
  • Likes Given: 34
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1046 on: 07/06/2022 12:13 am »
In general your primarily haggling over already solved problems for large spacecraft. Note Shuttle was a 100t+ vehicle that had extremely fine RCS control and did not need most of the things being discussed.
This is something I've harped on about for a while: moving a large heavy spacecraft into close proximity with another large heavy object and then gently touching the two together is a solved problem, not a new problem to solve.
The APAS ring also had a lot lower mass than the QD plate attached to Starship, and still tolerated reboost thrust (~0.2 milli-gee) with the orbiter joined to the station.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 2781
  • Tampa, FL
  • Liked: 2083
  • Likes Given: 402
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1047 on: 07/06/2022 04:42 am »
The downside of the booms is that any misalignment rubs the effectors up and down the sides of the Starship, which are covered in tiles.
Some articulation may be needed for alignment, but it only needs to be fairly gross alignment (get the narrow tip of the alignment pin into the wide opening of the socket) which could be accomplished with a lockable complaint mechanism (rather than powered actuation).

I think I had the geometry figured out so that the booms would only rub along the dorsal side, and then probably only within a meter or so of the hard point.  The idea was for the hard point to guide the effector/latch to it as the boom scraped along.  Whether that idea actually works... that's where the "half-baked" part came in.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 2781
  • Tampa, FL
  • Liked: 2083
  • Likes Given: 402
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1048 on: 07/06/2022 05:07 am »
In general your primarily haggling over already solved problems for large spacecraft. Note Shuttle was a 100t+ vehicle that had extremely fine RCS control and did not need most of the things being discussed.
This is something I've harped on about for a while: moving a large heavy spacecraft into close proximity with another large heavy object and then gently touching the two together is a solved problem, not a new problem to solve.
The APAS ring also had a lot lower mass than the QD plate attached to Starship, and still tolerated reboost thrust (~0.2 milli-gee) with the orbiter joined to the station.

There are some decent-sized caveats to that "solved problem" assertion:

1) Assuming you fill it full on the pad, any Starship is going to have a gross mass of >250t in LEO.  (The division between prop and payload may be different for different missions, but the gross mass will be the same.)  That's more than 2x the gross mass of an Orbiter.

2) You're dealing with some combination of cold gas, hot gas, methox, and/or methalox, not hypergolics.  Minimum impulse bit for anything with an igniter is going to be a lot bigger than the corresponding hypergolic system.  Similarly, minimum impulse bit for a cold gas thruster is going to be a lot lower.

3) There are a lot more roles for Starship's thrusters to play than there were on the Orbiter.  Orbiter used OMS for maneuvers with tens of m/s of delta-v, and it used RCS for attitude control, prox ops, and backup deorbit maneuvers.  Starship has to do all of those things plus ullage burns that go on for minutes to hours, and the LSS has landing thrusters (maybe).

4) You're dealing with a lot more slosh than the Orbiter had.

I'm sure that there's a solution somewhere in the trade space, but I'm less sure that the solution doesn't impact some aspects of docking or refueling.

Offline Paul451

  • Senior Member
  • *****
  • Posts: 3019
  • Australia
  • Liked: 2071
  • Likes Given: 1722
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1049 on: 07/06/2022 11:20 am »
I have to agree that QD latching isn't like sticking a chuck into a service station air line. Most probably rotating and actively powered as you say. A collet, actively clamped onto a simple shaft by a concentric motor would have

It's probably worth reminding people again what the QD passive and active plates actually look like. (See below.)

And I suggest that if you aren't referring to the actual QD system (or a very, very slight mod), you use other terminology for your variant (rather than QD) to avoid talking at cross purposes when someone is specifically referring to the existing mechanism and you're talking about something new.

The active side on the arm (male) contains guide pins, in addition it has rolling latches that lock into lugs on the female plate, clamping the plates together and apparently handling the final, fine alignment.

Images start with a labelled version of the female (ship-side) plate.
Then the male plate.
Then a couple of varying quality images of the male plate closer up, to show the guide-pins as well as the rolling locking latches.
VENTING IS NORMAL

Offline Paul451

  • Senior Member
  • *****
  • Posts: 3019
  • Australia
  • Liked: 2071
  • Likes Given: 1722
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1050 on: 07/06/2022 11:37 am »
While we're at it, I think everyone should clarify what they mean when they say "QD arm".

The tower arm with the QD assembly is called the "QD arm" (as distinct from the two "chopsticks") but is much, much more than just the QD plate on its specific assembly sub-arm. There's two "claws" that clamp around the SH to stabilise it, there's a bumper system which extends past the SH and I assume is used to bump the aft of the SS against when aligning for stacking. Then finally, on top of the arm holding those features, there's the relatively small QD assembly.

If you say "QD arm" are you talking about just the QD assembly sub-arm, or an entire grabber arm thing in addition to the specific QD part? Because those are very, very different situations.

See images below. Some are from during assembly, not everything's in place. And I've included a bunch of angles because focal foreshortening hides black-on-black trusses against each other, making depth perception difficult.

[edit: Ignore the green circles on the second image, that was already there.]
« Last Edit: 07/06/2022 11:39 am by Paul451 »
VENTING IS NORMAL

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 4149
  • United Kingdom
  • Liked: 6025
  • Likes Given: 34
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1051 on: 07/06/2022 12:11 pm »
If you say "QD arm" are you talking about just the QD assembly sub-arm, or an entire grabber arm thing in addition to the specific QD part? Because those are very, very different situations.
QD arm = the arm the QD plate sits on. It can extend and retract, move up-down-left-right relative to the QD plate, and the plate itself can pitch and yaw on the end oft he arm. It may be able to roll too, but that has not been observed.
Clip of the arm itself in motion: https://twitter.com/CSI_Starbase/status/1530366511784898562

The larger swing-out arm with the claw is the stabiliser arm. The QD arm sits on top of this.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1052 on: 07/06/2022 04:40 pm »
I have to agree that QD latching isn't like sticking a chuck into a service station air line. Most probably rotating and actively powered as you say. A collet, actively clamped onto a simple shaft by a concentric motor would have

It's probably worth reminding people again what the QD passive and active plates actually look like. (See below.)

And I suggest that if you aren't referring to the actual QD system (or a very, very slight mod), you use other terminology for your variant (rather than QD) to avoid talking at cross purposes when someone is specifically referring to the existing mechanism and you're talking about something new.

The active side on the arm (male) contains guide pins, in addition it has rolling latches that lock into lugs on the female plate, clamping the plates together and apparently handling the final, fine alignment.

Images start with a labelled version of the female (ship-side) plate.
Then the male plate.
Then a couple of varying quality images of the male plate closer up, to show the guide-pins as well as the rolling locking latches.
Can't speak for anybody else but I've been using QD plate much like you do. The gender bender would have an 'active' QD plate on each face to connect to the passive QD plate on the depot on one side and the passive QD plate of the second ship on the other side.

In discussion of latching, I've been speaking of the structural QD latching hooks. Do the structural latches hold the fluid connections together by brute force or is there another mechanism?


Edit: added the underlined words.
« Last Edit: 07/06/2022 04:43 pm by OTV Booster »
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 2781
  • Tampa, FL
  • Liked: 2083
  • Likes Given: 402
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1053 on: 07/06/2022 05:33 pm »
The question we need to focus on isn't about the quick disconnect, it's about the presence or absence of a quick connect

If edzieba is right that the QD can be latched reliably with very little force applied, then that dramatically simplifies the "docking" process.  I'm using scare quotes because, while we all agree that this isn't docking as we usually envision in it, it seems as if there are a wide variety of opinions on how rigid and accurate the docking needs to be.

At one extreme, there's no docking at all, except for the QD.  At the other, we need one or more structurally rigid connections capable of handling worst-case torques produced by the most heavily loaded depot-tanker, depot-payload, or tanker-tanker system.  In between, we've seen suggestions for articulated grapples, fairly dumb booms, with various mechanisms to make them sticky or latched, and plain ol' vanilla IDSS or APAS-like docks that don't happen to need pressurized tunnels.

The tower arm with the QD assembly is called the "QD arm" (as distinct from the two "chopsticks") but is much, much more than just the QD plate on its specific assembly sub-arm. There's two "claws" that clamp around the SH to stabilise it, there's a bumper system which extends past the SH and I assume is used to bump the aft of the SS against when aligning for stacking. Then finally, on top of the arm holding those features, there's the relatively small QD assembly.

If you say "QD arm" are you talking about just the QD assembly sub-arm, or an entire grabber arm thing in addition to the specific QD part? Because those are very, very different situations.

I've been assuming that it's fairly easy to take the cam and the bellows that are part of the GSE arm and adapt that so that the bellows is mostly looped in the remnants of the payload bay (with most of the cylinder consumed by moving the dome and bulkhead forward to contain more prop), with the cam able to extend the active plate through a door.  It wouldn't have to extend more than a meter or two.

But if you need the grabbers and bumpers and the entire stack stuck to the ground by gravity for the connection to be made reliably, then I doubt that you can make a grabber that extends out through a payload door, nor would it be sufficient to stabilize the pair of Starships even if it could.  Hence, the need for "docking".
« Last Edit: 07/06/2022 09:28 pm by TheRadicalModerate »

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1054 on: 07/07/2022 01:33 am »
The question we need to focus on isn't about the quick disconnect, it's about the presence or absence of a quick connect

If edzieba is right that the QD can be latched reliably with very little force applied, then that dramatically simplifies the "docking" process.  I'm using scare quotes because, while we all agree that this isn't docking as we usually envision in it, it seems as if there are a wide variety of opinions on how rigid and accurate the docking needs to be.

At one extreme, there's no docking at all, except for the QD.  At the other, we need one or more structurally rigid connections capable of handling worst-case torques produced by the most heavily loaded depot-tanker, depot-payload, or tanker-tanker system.  In between, we've seen suggestions for articulated grapples, fairly dumb booms, with various mechanisms to make them sticky or latched, and plain ol' vanilla IDSS or APAS-like docks that don't happen to need pressurized tunnels.

The tower arm with the QD assembly is called the "QD arm" (as distinct from the two "chopsticks") but is much, much more than just the QD plate on its specific assembly sub-arm. There's two "claws" that clamp around the SH to stabilise it, there's a bumper system which extends past the SH and I assume is used to bump the aft of the SS against when aligning for stacking. Then finally, on top of the arm holding those features, there's the relatively small QD assembly.

If you say "QD arm" are you talking about just the QD assembly sub-arm, or an entire grabber arm thing in addition to the specific QD part? Because those are very, very different situations.

I've been assuming that it's fairly easy to take the cam and the bellows that are part of the GSE arm and adapt that so that the bellows is mostly looped in the remnants of the payload bay (with most of the cylinder consumed by moving the dome and bulkhead forward to contain more prop), with the cam able to extend the active plate through a door.  It wouldn't have to extend more than a meter or two.

But if you need the grabbers and bumpers and the entire stack stuck to the ground by gravity for the connection to be made reliably, then I doubt that you can make a grabber that extends out through a payload door, nor would it be sufficient to stabilize the pair of Starships even if it could.  Hence, the need for "docking".
A decent summation. What are the chances that SX engineers have worn the same groves? I'd guess high. The only thing missing is data.


Another guess is that fiddling with StarLink will be high priority on the next flight followed closely by some maneuvering that would explore refueling ops.
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline colbourne

  • Full Member
  • ****
  • Posts: 424
  • Liked: 70
  • Likes Given: 25
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1055 on: 07/07/2022 11:44 am »
I think we will end up transferring full fuel tanks (possibly with engines attached) rather than pumping methane and oxygen across. This might require a redesign of the Mars transit craft but should solve many problems.

Online chopsticks

  • Full Member
  • ****
  • Posts: 436
  • Québec, Canada
  • Liked: 521
  • Likes Given: 95
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1056 on: 07/07/2022 12:23 pm »
I think we will end up transferring full fuel tanks (possibly with engines attached) rather than pumping methane and oxygen across. This might require a redesign of the Mars transit craft but should solve many problems.
Full fuel tanks with engines attached is basically what a rocket is. There's no "separate" tank that can be transfered - the fuel tank is the rocket body.

I don't see the prop transfer as being super hard thing to figure out, there are lots of good ideas of ways it could be done further up in this thread.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 2781
  • Tampa, FL
  • Liked: 2083
  • Likes Given: 402
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1057 on: 07/07/2022 04:40 pm »
Another guess is that fiddling with StarLink will be high priority on the next flight followed closely by some maneuvering that would explore refueling ops.

It should be noted that, if they really wind up with a dorsal-to-dorsal, front-to-back refueling configuration, getting this to play nice with the Starlink Pez dispenser isn't a slam-dunk.  This might not be essential; after all, Starlink missions don't require refueling.  But if you want to use Starships that have finished deploying their Starlinks  as test targets (or even test tankers) for refueling, then having them compatible might be nice.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1058 on: 07/08/2022 03:32 pm »
Another guess is that fiddling with StarLink will be high priority on the next flight followed closely by some maneuvering that would explore refueling ops.

It should be noted that, if they really wind up with a dorsal-to-dorsal, front-to-back refueling configuration, getting this to play nice with the Starlink Pez dispenser isn't a slam-dunk.  This might not be essential; after all, Starlink missions don't require refueling.  But if you want to use Starships that have finished deploying their Starlinks  as test targets (or even test tankers) for refueling, then having them compatible might be nice.
I was thinking just the solo (sub)orbital. Work the dispenser and then jinx around to get a feel for the handling. Two ships later on.


Two ships could be a short term problem. The first ship deploys sats then has to loiter until the second ship launches, however long that may be. It needs power and reaction mass. The power load would be computers, comms, RCS control and maybe some heaters. I've no idea of what the total load would be. The longer the loiter, the more batteries needed. PV is an option. At some length of loiter PV may be the only option.


This early in the game, depending on a booster turnaround looks high risk. With only one launch mount ISTM the best shot at a fast turnaround would be two full stacks, both having gone through a full pre launch test suite. Ship A would go back to the production facility after ground tests. Ship B would then test, then launch.


A big unknown is turnaround time for the table and tower. Another bottleneck could be propellant tankers. It's going to take a LOT of semi's. The upshot of this is, we (and SX) have no idea how much loiter time a ship will need while waiting for a second ship to show up.


If somebody has booster propellant capacity I'll work up the number of tankers needed to load a stack.



We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Offline OTV Booster

  • Senior Member
  • *****
  • Posts: 4181
  • Terra is my nation; currently Kansas
  • Liked: 2856
  • Likes Given: 4725
Re: Starship In-orbit refueling - Options and Discussion
« Reply #1059 on: 07/08/2022 04:23 pm »
Another guess is that fiddling with StarLink will be high priority on the next flight followed closely by some maneuvering that would explore refueling ops.

It should be noted that, if they really wind up with a dorsal-to-dorsal, front-to-back refueling configuration, getting this to play nice with the Starlink Pez dispenser isn't a slam-dunk.  This might not be essential; after all, Starlink missions don't require refueling.  But if you want to use Starships that have finished deploying their Starlinks  as test targets (or even test tankers) for refueling, then having them compatible might be nice.
I'm really not a fan of the catawompus alignment but if it works out to be best, ISTM it can be made to work.


First off, most of the cargo bay on a depot would be filled by O2 tankage and as you (or somebody) pointed out earlier, this would force the QD up into the ogive. Same, same for the Pez dispenser.


Instead of stowing the QD internally, build a dormer with the QD on its face instead of stowing it internally and deploying through a door. If it needs to extend this minimizes the 'tail' it has to drag around.


How all this would interact with aero launch forces could be a problem. If the QD plate gets a bit toasty during maxQ, extend the dormer overhang. Assuming the dormer can be made stable enough to make it through maxQ the problem would be drag. Cut he Pez dispenser down to more closely represent the depot top dome location and sacrifice a few starlinks, then move the QD dormer down a matching amount. This would both drop mass to allow powering through maxQ and reduce the dormers parasitic drag.
We are on the cusp of revolutionary access to space. One hallmark of a revolution is that there is a disjuncture through which projections do not work. The thread must be picked up anew and the tapestry of history woven with a fresh pattern.

Tags:
 

Advertisement NovaTech
Advertisement SkyTale Software GmbH
Advertisement Northrop Grumman
Advertisement
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0