The fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.
Quote from: Twark_Main on 10/22/2019 05:21 amI'd be pretty surprised if SpaceX made two duplicate systems for performing very similar functions on a highly mass-optimized vehicle. "Null[ing] out misalignments and torques" sounds exactly like the type of system you'd need to stack the stages with an unassisted or minimally-assisted crane op.As ZChris13 mentioned, SpaceX is transferring propellant through Super Heavy (no umbilical). So this is already a point of commonality with the SpaceX hardware roadmap.No. A crane is for all intents and purposes a four-degree-of-freedom system (z, zdot, zrot, zrotdot). Beyond that, you can have guides and manual tweaks for wind gusts, or whatever. By far the most important restricitions, though, are that gravity eliminates the xrot and yrot degrees of freedom, and simply knowing the position of the mount eliminates x and y.In contrast, an on-orbit docking system is a 12-degree-of-freedom system (x, y, z, xrot, yrot, zrot, and all dots of the previous). It needs much larger tolerances for soft capture, and then a way to convert that soft capture to an adequate hard capture. Think of what you're proposing as a super-duper hard capture and you won't be far off.Look, I think this idea is terrific. But it definitely would require a much heftier docking system, with a lot more complexity, to work. And it only gets heftier and more complex as you add thrust. Complexity, especially when you're talking about docking something to a crew that might not undock properly, is a big deal.
I'd be pretty surprised if SpaceX made two duplicate systems for performing very similar functions on a highly mass-optimized vehicle. "Null[ing] out misalignments and torques" sounds exactly like the type of system you'd need to stack the stages with an unassisted or minimally-assisted crane op.As ZChris13 mentioned, SpaceX is transferring propellant through Super Heavy (no umbilical). So this is already a point of commonality with the SpaceX hardware roadmap.
Which brings us to...QuoteMore than "slightly." At 626 kg/s (from livingjw's latest schematic), a 1500 tonne tank would take fully 40 minutes to empty. That's almost half an orbit, which has got to incur a substantial Oberth penalty. For comparison the Apollo TMI burn lasted less than 6 minutes, and Von Braun certainly did his homework.I'd be looking for 3-6 (vacuum) engines, in the ideal case. But see also Elon's latest comment about needing SL engines for engine-out.Here's a radical idea: perhaps have a set of 5 (yes, 5) vacuum engines, plus a single contingency central engine? In a nominal burn the central engine doesn't necessarily fire, but if a vacuum engine failed it would light and gimbal hard away from the failed engine to counteract the torque. Having an odd number of engines means the central engine has maximum gimbal range in the case with the highest torque asymmetry (ie 1 engine out).This configuration might even be able to land on Earth, given the upgraded methox hot gas thrusters, and that Musk said the dual-bell Raptor Vacs would work at SL. Possible Starship evolution path? I'll buy the 40 minute burn with one engine for a large-sized cargo (I used 135 t up to the lunar surface and 25 t back down to Earth). Maybe that needs 3 engines. However, two things:1) From a crew reliability standpoint, lower thrust is better on the docking system (see above). For the 25 t up and down case, your pusher burn is about 25 minutes, which is still hefty, but...
More than "slightly." At 626 kg/s (from livingjw's latest schematic), a 1500 tonne tank would take fully 40 minutes to empty. That's almost half an orbit, which has got to incur a substantial Oberth penalty. For comparison the Apollo TMI burn lasted less than 6 minutes, and Von Braun certainly did his homework.I'd be looking for 3-6 (vacuum) engines, in the ideal case. But see also Elon's latest comment about needing SL engines for engine-out.Here's a radical idea: perhaps have a set of 5 (yes, 5) vacuum engines, plus a single contingency central engine? In a nominal burn the central engine doesn't necessarily fire, but if a vacuum engine failed it would light and gimbal hard away from the failed engine to counteract the torque. Having an odd number of engines means the central engine has maximum gimbal range in the case with the highest torque asymmetry (ie 1 engine out).This configuration might even be able to land on Earth, given the upgraded methox hot gas thrusters, and that Musk said the dual-bell Raptor Vacs would work at SL. Possible Starship evolution path?
2) You're starting out in a circular orbit, so there's no "Oberth effect" per se. Instead, there are some efficiency losses from a non-impulsive burn.
My math and tools aren't up to computing that but, intuitively, they ought to be pretty small. Essentially, you're bleeding off tangential velocity as your radial distance rises, which means that you add less energy than you would otherwise, as the burn progresses. Maybe somebody has a handy numerical sim for a non-impulsive burn of about 2 MN with about a 3210 t coupled spacecraft with Isp=365 s, departing from a 200x200 orbit?
snipThe fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.snip
Quote from: TheRadicalModerate on 10/22/2019 10:47 pmsnipThe fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.snipOne could imagine a scenario where the nose of the P/T is actually a fairing, which upon jettisoning reveals a non-aerodynamic structure for docking/thrust transfer. It could be modeled on the rear end of a SS, using the normal docking/refueling infrastructure otherwise required (but simply reversed), and also the thrust structures of SS, but (obviously) no raptors. Essentially, make the P/T have two tails, with a fairing covering the "nose-tail" during ascent.I see re-using the SS rear-end mainly as a benefit to not having to engineer some kind of one-off thrust structure - the drawback is that it's likely overkill, and will add significant dry mass to the P/T, reducing its overall utility - no idea to what extent, though.
Quote from: TheRadicalModerate on 10/18/2019 08:19 pmThe only way to avoid the extra two trips through the VAB and the long loiter at the edge of VAB #2 is to take the tanker all the way to TLI with the crewed Starship, then refuel in trans-lunar coast.There is one way I can think of, but it may be too "radically moderate" for your tastes. Have the Tanker push the Starship.All it needs is a second docking port on the front (refueling plumbing not necessary, just the structural hard points and pneumatic separation collets). During TLI the Tanker engines fire, followed by a more-or-less conventional stage separation (hence "moderate"), then Starship's engines fire to finish the TLI burn. Afterwards the Tanker does a small perigee lowering burn and aerobrakes back to LEO. If the Starship engines fail to light you can abort to LEO (or abort-to-reentry) the same way. For launch it might have a nose cone to reduce drag, either disposable or reusable (your choice depending on your assumption about flight rate). Reentry is probably out of the question, so such a vehicle would be reused on-orbit. Probably that means you'll want to throw some Hall effect thrusters on it too, for reboost and adjusting the orbit.Too radical? Or just right?
The only way to avoid the extra two trips through the VAB and the long loiter at the edge of VAB #2 is to take the tanker all the way to TLI with the crewed Starship, then refuel in trans-lunar coast.
Quote from: TheRadicalModerate on 10/22/2019 10:47 pmThe fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.I may have missed it in your previous discussions, but why wouldn’t the nose of the P/T be the same as the nose of SH. In fact, why not simply make the P/T a SH variant instead of a SS variant?edit - grammar
Quote from: gnari on 09/29/2019 09:17 pmQuote from: Norm38 on 09/29/2019 08:51 pmA tanker as depot makes the most sense to me. Launch one with special long term storage mods, fill it with regular tanker flights, then fill a Starship. If theres enough traffic, the storage tanker might just stay on orbit. Give it a permanent sun shield, solar array and radiator. Why land and relaunch that tanker?I am not understanding what advantage this is supposed to give, over just sending up tankers as needed. can someone explain?I am only seeing disadvantages, such as needing that sun shield, having that tanker unavailable for sending up fuel for lengths of time and also such a depot might be less useful for missions that could profit from being tanked in some other orbit than the depot happens to be in.You could fully fuel up the depot before launching the mars payload. That way you would only have 1 docking event for the main mission.
Quote from: Norm38 on 09/29/2019 08:51 pmA tanker as depot makes the most sense to me. Launch one with special long term storage mods, fill it with regular tanker flights, then fill a Starship. If theres enough traffic, the storage tanker might just stay on orbit. Give it a permanent sun shield, solar array and radiator. Why land and relaunch that tanker?I am not understanding what advantage this is supposed to give, over just sending up tankers as needed. can someone explain?I am only seeing disadvantages, such as needing that sun shield, having that tanker unavailable for sending up fuel for lengths of time and also such a depot might be less useful for missions that could profit from being tanked in some other orbit than the depot happens to be in.
A tanker as depot makes the most sense to me. Launch one with special long term storage mods, fill it with regular tanker flights, then fill a Starship. If theres enough traffic, the storage tanker might just stay on orbit. Give it a permanent sun shield, solar array and radiator. Why land and relaunch that tanker?
Quote from: SteveU on 10/23/2019 12:00 amQuote from: TheRadicalModerate on 10/22/2019 10:47 pmThe fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.I may have missed it in your previous discussions, but why wouldn’t the nose of the P/T be the same as the nose of SH. In fact, why not simply make the P/T a SH variant instead of a SS variant?edit - grammar Twark had mentioned something similar on the other thread. The big difference is going to be whether the interstage latch system can do double duty as a docking system. I'm high skeptical of that, but hopefully some of the interstage thrust transmission and plumbing can be the same.
Docking was solved in 1966. Several dockings do not pose a problem.
The interstage latch system would not be used as the alignment system used for docking/latching. I think you're making this more difficult than it needs to be. We all agree that they will be able to achieve a mating of two SS to transfer fuel.In order to be able to have in-flight refueling, the two vessels will need to be docked - at some level of "hardness" - a "soft" dock won't cut it. They might be able to adjust for a few cm of misalignment along the x, y and z axis's but i cannot see how any movement can be tolerated while transferring fuel. The center of mass of the combined vessels will be constantly moving and any movement between the two ships at this time would be catastrophic (IMHO).What that leaves us with is just how well aligned will the vessels need to be in order to be able to engage the interstage latching mechanism? Since refueling will require cm accuracy, it isn't that difficult to imagine that the latching mechanism would be designed to accommodate a few cm of misalignment. Once the two vessels are "Docked" and all motion has ceased, engage the latches. Isn't this the same as would be done when stacking at the launch pad?What i'm envisioning for the docking mechanism - imagine a supersized Apollo Probe and Drogue. Using four of them on the SS located inboard of the +x and -x leg positions, probes at +x, drogues at -x you have a system of alignment that should be able to give sub-cm accuracy. If sturdy enough, they could be used to assist in alignment during stacking as well.PS - this brings us back to the discussions we had regarding popping the nose-cone off for cargo transfer!
Quote from: TheRadicalModerate on 10/23/2019 04:35 amQuote from: SteveU on 10/23/2019 12:00 amQuote from: TheRadicalModerate on 10/22/2019 10:47 pmThe fly in the ointment here is this thrust-bearing nose-to-tail docking scheme. The crew/cargo SS can obviously bear the load, because this is essentially how it gets launched, but the latch for an on-orbit docking that has to bear multiple meganewtons of thrust and gimbaling is a lot more complicated than one that just has to hold a pair of tail-to-tail tankers together under an ullage burn. Twark and I have been having a spirited discussion about this, but I don't think we've gotten to an answer. Any thoughts on the feasibility would be appreciated.I may have missed it in your previous discussions, but why wouldn’t the nose of the P/T be the same as the nose of SH. In fact, why not simply make the P/T a SH variant instead of a SS variant?edit - grammar Twark had mentioned something similar on the other thread. The big difference is going to be whether the interstage latch system can do double duty as a docking system. I'm high skeptical of that, but hopefully some of the interstage thrust transmission and plumbing can be the same.The interstage latch system would not be used as the alignment system used for docking/latching. I think you're making this more difficult than it needs to be. We all agree that they will be able to achieve a mating of two SS to transfer fuel.In order to be able to have in-flight refueling, the two vessels will need to be docked - at some level of "hardness" - a "soft" dock won't cut it. They might be able to adjust for a few cm of misalignment along the x, y and z axis's but i cannot see how any movement can be tolerated while transferring fuel. The center of mass of the combined vessels will be constantly moving and any movement between the two ships at this time would be catastrophic (IMHO).
What that leaves us with is just how well aligned will the vessels need to be in order to be able to engage the interstage latching mechanism? Since refueling will require cm accuracy, it isn't that difficult to imagine that the latching mechanism would be designed to accommodate a few cm of misalignment. Once the two vessels are "Docked" and all motion has ceased, engage the latches.
Isn't this the same as would be done when stacking at the launch pad?
What i'm envisioning for the docking mechanism - imagine a supersized Apollo Probe and Drogue. Using four of them on the SS located inboard of the +x and -x leg positions, probes at +x, drogues at -x you have a system of alignment that should be able to give sub-cm accuracy. If sturdy enough, they could be used to assist in alignment during stacking as well.
PS - this brings us back to the discussions we had regarding popping the nose-cone off for cargo transfer!
I don't believe a crane will constrain the movement of the Starship for on-pad assembly as much as you think it will, Twark_Main. Even a light wind against the massive surface area of Starship is going to move even the best rigged system. Steel lines have quite a bit of stretch in them at that scale. There's really no telling if the pad assembly equipment or a close facsimile will work and I find it unlikely that we'll ever find out.
Quote from: Oersted on 10/23/2019 09:03 amDocking was solved in 1966. Several dockings do not pose a problem.Reading this thread, and this is what stands out to me. All of these ideas to minimize rendezvous and docking maneuvers are cool, but it's additional hardware development in order to reduce something that SpaceX are planning to do thousands and more of.
I don't believe a crane will constrain the movement of the Starship for on-pad assembly as much as you think it will, Twark_Main. Even a light wind against the massive surface area of Starship is going to move even the best rigged system. Steel lines have quite a bit of stretch in them at that scale. There's really no telling if the pad assembly equipment or a close facsimile will work and I find it unlikely that we'll ever find out.Quote from: Oersted on 10/23/2019 09:03 amDocking was solved in 1966. Several dockings do not pose a problem.Reading this thread, and this is what stands out to me. All of these ideas to minimize rendezvous and docking maneuvers are cool, but it's additional hardware development in order to reduce something that SpaceX are planning to do thousands and more of.
...However hard doing the crane-based alignment is, the point was that it's vastly easier than a free-flying hard dock that can tolerate high injection loads and still reliability undock later.
Quote from: TheRadicalModerate on 10/23/2019 05:40 pm...However hard doing the crane-based alignment is, the point was that it's vastly easier than a free-flying hard dock that can tolerate high injection loads and still reliability undock later.I really disagree. When the two ships are in refueling position they will need to be as close to a hard dock as possible. In this state, there is no reason that the same interstage latching mechanism between SH/SS could not be used between P/T & SS. I do not believe that there can be movement between the two ships while transferring tens of thousands pounds of fuel. If you are concerned about a docking mechanism failing due to the masses involved, just how will fueling connectors hold up?The intersatge latching mechanisms are designed to tolerate the dynamic loads of launch and disengage at stage separation. When stacking the ships on the pad there will have to be a way for the latches to "self cinch" to obtain a tight grip on SS - unless you believe that this will need to be a manual operation. If the latching/cinching is automatic, I can't think of a reason why it wouldn't work in LEO. And the forces involved will be much less than during launch.
Quote from: SteveU on 10/23/2019 06:18 pmQuote from: TheRadicalModerate on 10/23/2019 05:40 pm...However hard doing the crane-based alignment is, the point was that it's vastly easier than a free-flying hard dock that can tolerate high injection loads and still reliability undock later.I really disagree. When the two ships are in refueling position they will need to be as close to a hard dock as possible. In this state, there is no reason that the same interstage latching mechanism between SH/SS could not be used between P/T & SS. I do not believe that there can be movement between the two ships while transferring tens of thousands pounds of fuel. If you are concerned about a docking mechanism failing due to the masses involved, just how will fueling connectors hold up?The intersatge latching mechanisms are designed to tolerate the dynamic loads of launch and disengage at stage separation. When stacking the ships on the pad there will have to be a way for the latches to "self cinch" to obtain a tight grip on SS - unless you believe that this will need to be a manual operation. If the latching/cinching is automatic, I can't think of a reason why it wouldn't work in LEO. And the forces involved will be much less than during launch.Let's break this down. I'll obviously agree that if you can completely replicate the conditions of stacking on-orbit, then we're done. So: what's different? Three things:1) Positional and momentum accuracy. You can stack the SS onto the SH in a hurricane and you'll still have fewer degrees of freedom than you will on-orbit, and better guidance.2) Pre-loading force before latching. Set an empty Starship down on an SH and there will be about 1.2 MN of force loading the latch system before they engage. I can't imagine a hard-dock retraction pre-loading a soft capture as evenly and reliably as good ol' gravity will.3) Recovery from latching errors. If things go screwy on the ground, a tech can inspect and, if necessary, reset the latch. That's not happening on-orbit. Even in the worst case of a jammed latch that can't be recovered during stacking, you can drill the silly thing out, pop the others, unstack, and repair. Again, none of those are options on-orbit, and the lack of a recovery mode means that your crew will have to be rescued from a 100 m long, not-very-structurally-sound vehicle with now-limited ability to manage its own attitude (because the stack is AFU).I completely agree that a hard dock of some quality is needed to do refueling ops. However, it's that "of some quality" modifier that's in dispute. The forces and torques incurred during refueling simply aren't anywhere near the forces and torques that would be required to allow a P/T to efficiently inject an SS into an orbit that's high enough energy for it to continue on to TLI--or TMI, if you want to do much-faster-than-Hohmann transfers.I believe that this is a solvable problem. However, I'm not sure it's solvable by reusing the latching system used for stacking, however desirable that might be. I'm also not sure that it's easily solvable if it isn't planned-for ahead of time. That might make a scheme like P/T non-viable later on--which would be a shame, because it solves some pretty important issues with crew safety and even cargo ops complexity quite elegantly.