SpaceX, rapid reuse, and vertical rocket stacking (integration)

[meekGee]

To start off, this is based on engineering reasoning, not on any inside information or anything that SpaceX has made public.

I think that with rapid reuse, SpaceX will move to vertical rocket stacking.

The usual mantra is that "SpaceX does horizontal integration since it's cheaper".   The word that comes to mind for me, however, when I see horizontal integration is "agility" rather than "low cost".

There's no argument of course that a strong back, a railway, and a couple of ladders (to over-simplify things) is cheaper than a mobile service tower.

However, in any industrial process, there's the trend that as the tempo of operations increases, you invest in more expensive, more dedicated infrastructure that reduced the per-operation costs.  Of course you MUST have enough volume to amortize the cost of the infrastructure over.  Many companies failed because they invested too quickly in such infrastructure, and were then stuck with the debt of the capital costs, while sales volume didn't increase as fast as they thought it would.  A mobile service tower, when you're only launching once a month, is probably an overkill from a cost perspective.

Elon's a smart guy.  I reckon the one thing he doesn't want to do is repeat other people's mistakes.  If he has to make any, he wants to make his own.

So back to the infrastructure.  The best example I have in mind is the standard diamond jack in your car.  It's cheap, agile (can be used with pretty much any car) and does the job.   But if you go to a NASCAR race, you'll see a different type of jack:  Dedicated, expensive, lifts the entire car at once, and does so in 2 seconds.  (Further, it gets the operator away from the path of the wheel-changers and refueling guy, so operations can happen even faster)   In NASCAR, time is money, and this jack is worth every penny.

So back to rockets.

The strongback arrangement is agile.  SpaceX flew the F9 1.0 only what, 5 times over 2 years?  and I bet there were changes to it all the time, and then ta-da! F9 1.1 is some 20% longer.  So now what - add stories to the mobile service tower?  Clearly horizontal integration is the correct choice for their current mode of operation.

However, when you want to turn around a rocket in a matter of days, it's a completely different story.  The rocket is not changing any more.  You can build dedicated support structures for it, with safe gantries at all necessary levels.

A vertical rocket gives you access to everything at once.  All 9 engines, in vertical position, the bells can be centered and unloaded, all vertical walls are accessible at eye level simply by walking to them, all fluid tanks and plumbing only ever see one orientation, so draining and such is easier (trap-wise), etc.

A horizontal rocket only gives you sort-of-good access to the lower engine, and you have to rotate it.  So if you're servicing engine #3 and need to escalate something for deeper inspection, all the other teams (e.g working on the RCS thrusters) have to wait since the rocket can't roll.  (not to mention that access to the center engine is awkward)

A vertical first stage can be rolled, as-is, on its legs, into a fixed service tower, doors can close around it, and the rocket can processed by as many people as you wish.  It can then be rolled to the pad for attachment to the launch holds, and fueling.  Faster, and per operation, cheaper.

Stacking of the US and payload can occur at the service tower (need somewhat stronger legs), or at the pad. In my mind, for a rapidly reusable rocket, the connection between the first and second stage is basically only mechanical.  Each is an independent flying vehicle, and so basically they just need to be latched together. If they need to communicate, it's low bandwidth digital stuff, and a two piece opto-coupler will do the trick just fine.

Finally, there's no tilting.  In an expendable rocket, since it's transported horizontally, there will always be a tilt-up. Either before integration, or before launch.   But a reusable rocket lands vertically.  At this point horizontal integration becomes ADDED complexity.

How soon?  How would I know... Development wise, SpaceX is moving incredibly fast.  Will this happen with F9 1.2? (or whatever the naming scheme will be) - maybe.  It depends whether the current F9 was designed with this option in mind. I have no inside information, so am just waiting to see.

So, again, this is basically only my personal opinion.  What SpaceX plans to or will do is, well, obviously up to them.

[guckyfan]

The idea of a full service tower may well be good, if a fast launch rate justifies the expense.

However regarding speed of operation. The landing pad will be a few km away from the launchpad. Getting the stage horizontal and transporting it that way will be a LOT faster and safer than moving it vertically. Once you have it horizontal you can mate first stage and pre-loaded second stage very easily. The second stage with integrated payload waits in the hangar. The first stage enters through the back door, they are mated and move out the frontdoor to the pad. Like on a roll on roll of ferry. Service that can be done better horizontal in the hangar will be done there. Service that can be done better vertical in a launch tower will be done there.

The need for a complex launch structure at the pad seems to me way in the future. Servicing the engines does not need it as they demonstrated with the SES-8 scrubs and engine service.

Edit: purely my opinion of course.

[Jim]

  Each is an independent flying vehicle, and so basically they just need to be latched together. If they need to communicate, it's low bandwidth digital stuff, and a two piece opto-coupler will do the trick just fine.

Not they are not independent. There is high bandwidth requirements and many connections.   The upperstage provides guidance for the stack.  There are range safety harnesses.  There are telemetry lines v

[Jim]

1.  A vertical rocket gives you access to everything at once.  All 9 engines, in vertical position, the bells can be centered and unloaded, all vertical walls are accessible at eye level simply by walking to them, a

2.  A horizontal rocket only gives you sort-of-good access to the lower engine, and you have to rotate it.  So if you're servicing engine #3 and need to escalate something for deeper inspection, all the other teams (e.g working on the RCS thrusters) have to wait since the rocket can't roll.  (not to mention that access to the center engine is awkward)

No, there is a hole under the engines and there are vehicle holddowns and umbilicals in the way.  There is nothing around the aft section of the vehicle in the hangar.

And because the design of the vehicle is so great and the manpower so low, there is no need or capability for simultaneous access, the engines can get worked on and then the RCS.

[meekGee]

  Each is an independent flying vehicle, and so basically they just need to be latched together. If they need to communicate, it's low bandwidth digital stuff, and a two piece opto-coupler will do the trick just fine.

Not they are not independent. There is high bandwidth requirements and many connections.   The upperstage provides guidance for the stack.  There are range safety harnesses.  There are telemetry lines v

I think that's one of the things that will change.  There is absolutely no reason why the first stage can't control the ascent till separation.  It has all the necessary avionics and sensors, and they are redundant.

What you describe is the logical way to do it in an expendable rocket.

That's part of re-thinking the design when you build a reusable rocket.

The first stage is an independent "carrier plane" for the upper stage.  So you don't need the avionics at the U/S to sense and control the engines on the first stage.  It's just that the two avionics suites need to communicate at a high level, the the US can shadow the flight.

The first stage will have its own independent range safety system anyway, and its own downlink anyway, since it needs to talk to the ground when if comes back to land.

So again, no connections necessary.

[guckyfan]

1.  A vertical rocket gives you access to everything at once.  All 9 engines, in vertical position, the bells can be centered and unloaded, all vertical walls are accessible at eye level simply by walking to them, a

2.  A horizontal rocket only gives you sort-of-good access to the lower engine, and you have to rotate it.  So if you're servicing engine #3 and need to escalate something for deeper inspection, all the other teams (e.g working on the RCS thrusters) have to wait since the rocket can't roll.  (not to mention that access to the center engine is awkward)

No, there is a hole under the engines and there are vehicle holddowns and umbilicals in the way.  There is nothing around the aft section of the vehicle in the hangar.

And because the design of the vehicle is so great and the manpower so low, there is no need or capability for simultaneous access, the engines can get worked on and then the RCS.

Where do these quotes come from? They are not from me and I don't see them in this thread.

Edit: Fixed quotes

Second Edit: I see the quotes, they are in meekGees opening post, but not mine.

[Jim]

Stacking of the US and payload can occur at the service tower (need somewhat stronger legs), or at the pad.

Nip this one in bud.  That is two separate lifts.  There is no positives and only negatives.  Lifting a fueled spacecraft is a facility/complex clear.  Lifting the spacecraft integrated with the second stage on the first stage means there had to be an early lift of the spacecraft on to the second stage, which would be another hazardous lift with clears.  The mantra of keeping spacecraft and launch vehicles apart as long as possible fits into Spacex's matra.  Assemble the launch vehicle and test it and then put the spacecraft on as close to launch as possible.  This keep the two product lines independent and from interfering with each other until late as possible.

[meekGee]

1.  A vertical rocket gives you access to everything at once.  All 9 engines, in vertical position, the bells can be centered and unloaded, all vertical walls are accessible at eye level simply by walking to them, a

2.  A horizontal rocket only gives you sort-of-good access to the lower engine, and you have to rotate it.  So if you're servicing engine #3 and need to escalate something for deeper inspection, all the other teams (e.g working on the RCS thrusters) have to wait since the rocket can't roll.  (not to mention that access to the center engine is awkward)

No, there is a hole under the engines and there are vehicle holddowns and umbilicals in the way.  There is nothing around the aft section of the vehicle in the hangar.

What hole?  Why would the fixed service tower have a hole under the engines?  If anything, it has a nice jack there, so you the engines are easy to pull.

[rcoppola]

  Each is an independent flying vehicle, and so basically they just need to be latched together. If they need to communicate, it's low bandwidth digital stuff, and a two piece opto-coupler will do the trick just fine.

Not they are not independent. There is high bandwidth requirements and many connections.   The upperstage provides guidance for the stack.  There are range safety harnesses.  There are telemetry lines v

I think that's one of the things that will change.  There is absolutely no reason why the first stage can't control the ascent till separation.  It has all the necessary avionics and sensors, and they are redundant.

What you describe is the logical way to do it in an expendable rocket.

That's part of re-thinking the design when you build a reusable rocket.

The first stage is an independent "carrier plane" for the upper stage.  So you don't need the avionics at the U/S to sense and control the engines on the first stage.  It's just that the two avionics suites need to communicate at a high level, the the US can shadow the flight.

The first stage will have its own independent range safety system anyway, and its own downlink anyway, since it needs to talk to the ground when if comes back to land.

So again, no connections necessary.

There will need to be connections from 1st to 2nd to Dragon when it's crewed for abort sequences.

Also, I think Payload integration will have more to do with how they introduce future vertical integration then reusability will.

[meekGee]

Stacking of the US and payload can occur at the service tower (need somewhat stronger legs), or at the pad.

Nip this one in bud.  That is two separate lifts.  There is no positives and only negatives.  Lifting a fueled spacecraft is a facility/complex clear.  Lifting the spacecraft integrated with the second stage on the first stage means there had to be an early lift of the spacecraft on to the second stage, which would be another hazardous lift with clears.  The mantra of keeping spacecraft and launch vehicles apart as long as possible fits into Spacex's matra.  Assemble the launch vehicle and test it and then put the spacecraft on as close to launch as possible.  This keep the two product lines independent and from interfering with each other until late as possible.

Maybe, but it's not an absolute.

I'd say the mantra (I hate mantras.  Every rule of thumb, every one-time statement becomes a mantra) preference would be to fuel the spacecraft at the pad - if the spacecraft allows it.

The reason I'm thinking about a combined lift (and yes, you're right, there would be a lift of the spacecraft onto the second stage first) is that Elon was talking about the first stage flying back within a day (or a few days), but the second stage will probably have a longer turn around time (longer time in space, maybe have to service the landing engines - we don't know much about those yet, and heat shield, etc)

So for steady state, you need more upper stages than you do first stages.   This doesn't directly mean you do the combined stack, but it means you can save time on your critical path by pre-integrating the US and the Spacecraft when possible.

The two options are not mutually exclusive btw.  If the Spacecraft absolutely has to be fueled in the service tower, then fine - do separate lifts, and slow everyone down - for a price of course.   (or fuel it after it's lifted, but still in the service tower)

Another thing you need to consider for rapid reuse is that payloads by definition have to be highly standardized. (As GS so eloquently put it - no "Giving birth")

So under this scenario, SpaceX will dictate the interface that the spacecraft must adhere to, and this will include on-pad refueling - so that the process can be streamlined.

[Jim]

I think that's one of the things that will change.  There is absolutely no reason why the first stage can't control the ascent till separation.  It has all the necessary avionics and sensors, and they are redundant.

What you describe is the logical way to do it in an expendable rocket.

That's part of re-thinking the design when you build a reusable rocket.

The first stage is an independent "carrier plane" for the upper stage.  So you don't need the avionics at the U/S to sense and control the engines on the first stage.  It's just that the two avionics suites need to communicate at a high level, the the US can shadow the flight.

The first stage will have its own independent range safety system anyway, and its own downlink anyway, since it needs to talk to the ground when if comes back to land.

So again, no connections necessary.

No, it is the wrong thinking based on the lack of knowledge in the field and using the excuse of "reusable" to make up for it.

The first stage is not "carrier plane".  It does not have excess fuel to loiter and reschedule the drop like carrier plane nor does it have to follow a precise trajectory.   

The upperstage needs to control the whole flight so that it can integrate the trajectory and make decisions on staging based on what is will be able to make up for first stage shortfalls or tell the first stage to keep burning and give up on recovery.

All launch vehicles have independent FTS systems on each stage but they are still tied together by data and sensing lines.  This is to ensure that received signal gets to all destruct packages and to shut down the engines and to sense inadvertent separation. 

[Jim]

this will include on-pad refueling - so that the process can be streamlined.

That will do the opposite and add many days to the flow.  Loading and servicing spacecraft on the pad is the exact opposite of what everybody and specifically Spacex want to do.   This really shows that you don't know what you are talking about. 

[meekGee]

The idea of a full service tower may well be good, if a fast launch rate justifies the expense.

However regarding speed of operation. The landing pad will be a few km away from the launchpad. Getting the stage horizontal and transporting it that way will be a LOT faster and safer than moving it vertically. Once you have it horizontal you can mate first stage and pre-loaded second stage very easily. The second stage with integrated payload waits in the hangar. The first stage enters through the back door, they are mated and move out the frontdoor to the pad. Like on a roll on roll of ferry. Service that can be done better horizontal in the hangar will be done there. Service that can be done better vertical in a launch tower will be done there.

The need for a complex launch structure at the pad seems to me way in the future. Servicing the engines does not need it as they demonstrated with the SES-8 scrubs and engine service.

Edit: purely my opinion of course.

Maybe.  It would be an interesting race.   

1)  Grab/Cradle, tilt, drive-like-the-wind, untilt.
2)  Tow slowly.

How many km to you envision between launch and landing pad?

If there's a dirt mount between them, it can be a very short distance, just enough to guarantee separation under the divert maneuver strategy.

[guckyfan]

Engine service for SES-8 was done vertical and Elon Musk stated they left the Falcon on the pad because it is easier to access the engines there.

I hope I don't have to dig out that quote.

[guckyfan]

Maybe.  It would be an interesting race.   

1)  Grab/Cradle, tilt, drive-like-the-wind, untilt.
2)  Tow slowly.

How many km to you envision between launch and landing pad?

If there's a dirt mount between them, it can be a very short distance, just enough to guarantee separation under the divert maneuver strategy.

I was thinking of distances on Cape Canaveral which would be several km. But even if they are closer they would have to drive around the dirt mound which adds distance. I really don't think they will ever like to land within less than a km from the launch pad.

[Jim]

No, it is the wrong thinking based on the lack of knowledge in the field and using the excuse of "reusable" to make up for it.

You were doing good till this.   Not responding otherwise.

Its your standard mantra, "processes for ELV's don't apply for reusable vehicles" without understanding what the process is

[meekGee]

There will need to be connections from 1st to 2nd to Dragon when it's crewed for abort sequences.

Also, I think Payload integration will have more to do with how they introduce future vertical integration then reusability will.

True, but that's high-level, low-bandwidth - the opto-coupler or similar will suffice.

And payload integration is really a side discussion.  The logic in the OP holds even if you eventually load a pre-fueled spacecraft to the top of the stack.  I mean, that's how it'd done today with vertical integration, and I was proposing a move to vertical integration, so no conflict really.

[meekGee]

Maybe.  It would be an interesting race.   

1)  Grab/Cradle, tilt, drive-like-the-wind, untilt.
2)  Tow slowly.

How many km to you envision between launch and landing pad?

If there's a dirt mount between them, it can be a very short distance, just enough to guarantee separation under the divert maneuver strategy.

I was thinking of distances on Cape Canaveral which would be several km. But even if they are closer they would have to drive around the dirt mound which adds distance. I really don't think they will ever like to land within less than a km from the launch pad.

how long does it take to cradle, tilt, and untilt IYO?

[guckyfan]

how long does it take to cradle, tilt, and untilt IYO?

They have to cradle it somehow for moving, no matter moving horizontal or vertical.

Their declared aim is to go from hangar to launch in one hour. But that seems hard as it includes tanking and the whole launch sequence. But tilting horizontal and back should be less than that. I don't see them moving the vertical stage in that time by even only a few hundred meters.

[meekGee]

how long does it take to cradle, tilt, and untilt IYO?

They have to cradle it somehow for moving, no matter moving horizontal or vertical.

Their declared aim is to go from hangar to launch in one hour. But that seems hard as it includes tanking and the whole launch sequence. But tilting horizontal and back should be less than that. I don't see them moving the vertical stage in that time by even only a few hundred meters.

I think if it can land on its legs, it can be towed on its legs.  You need to attach a dolly to each one, just like they do to skidded helicopters, but that's much simpler than attaching a cradle to an empty unpressurized tank.

It's not a fragile process, there's no alignment necessary etc.   Just one dolly at a time, it can be done pretty much manually by the tow-tractor driver.

The tractor can then move at least as fast as the one pushing your plane back at the airport, which is a good walking speed, and so 3-5 km/h.   If you're 1 km away, you're done in 10-20 minutes.  (and honestly I think you can drive faster...)

If you want to get fancier, you can invest in motorized, individually controlled dollies, and no tractor.