LM Descent Stages Today?

[monophonic]

5 hours ago, sevenperforce said:

Dragon 2 (with kick stage in the trunk)

Could one make use of the eight SuperDracos on the D2? Pack extra fuel in the trunk and run some extra fuel lines? They are the LES after all, so they cannot be removed despite NASA wanting parachute splashdowns instead of powered landings.

[sevenperforce]

14 minutes ago, monophonic said:

Could one make use of the eight SuperDracos on the D2? Pack extra fuel in the trunk and run some extra fuel lines? They are the LES after all, so they cannot be removed despite NASA wanting parachute splashdowns instead of powered landings.

The SuperDracos are underexpanded even at sea level, to allow very good throttle response, so their ISP is very poor...on the order of 250 s in vacuum. Plus, they are canted 15 degrees off-axis, which means cosine losses. TWR is no problem, but you'd need an enormous amount of propellant -- like, twice as much as could fit in the whole trunk.

Plus, trying to crossfeed the SuperDracos from an auxiliary tank would require a complete system redesign (the SuperDracos are pressure-fed hypergolics, which means helium COPVs).

A better solution would be to simply cluster a few ordinary Dracos in the trunk with their own tanks. They are already vacuum-expanded (300 s) and while they're not super-thrusty, you don't really need a huge TWR for the lunar injection or return burns. The whole system could be attached to the payload adapter inside the trunk, so you have zero modifications to the Dragon 2 itself.

You need 1.3 km/s each for the lunar-orbit insertion burn and the Earth-return burn. That's a 1.4:1 propellant fraction, though if SpaceX could add an expansion nozzle to the Dracos to boost the isp up to around 320 s, it would drop to 1.2:1. The standard Dragon 2 is 7.8 tonnes with onboard propellant; allowing another tonne for crew and consumables bumps this up to 8.8 tonnes. Assuming that tankage, extra engines, and extra structure increase the dry mass slightly, you'd be looking at around a 13-tonne kick stage. The density of the hypergols used by the Dracos is 1.2 g/cc, so the kick stage would occupy a volume of 11 cubic meters, well within the 14-cubic-meter unpressurized volume of the Dragon 2's trunk. The whole stack would mass 21.8 tonnes, which is just under an expendable Falcon Heavy's TLI capacity. If the Falcon upper stage could be modified to allow for extended restarts, you'd save about 8 tonnes of mass on the kick stage, which might be enough margin to allow for side-booster recovery (the MVac's higher isp would make up for the additional dry mass).

All this assumes JLOR. If we want to try ELOR like Constellation, other possibilities open up, though the Dragon 2 is always going to need at least a 6-tonne hypergolic kick stage to get from low lunar orbit back to Earth.

[sevenperforce]

P.S. The reason the helium COPVs are a problem is because a pressure-fed engine system doesn't lend itself well to having alternate plumbing shunted into it.

If SpaceX did use Dracos to build a drop-in hypergolic kick stage for Dragon 2 with a good 2.6 km/s of dV, it could be tested in a free-return lunar orbit mission for the cost of an expendable Falcon 9 or a full-recovery FH. The LV puts it into a high elliptical orbit, and then the kick stage fires and bumps it up into the free-return.

[wumpus]

More thoughts:

Ideally, if the Merlins aren't enough to send the mass to lunar intercept, your first choice should be a RL-10 (hydrolox expander engine).  Storing enough hydrogen for the return flight might be iffy, but it is ideal for leaving Earth.

While pressure-fed hypergolics are nearly always the way to go for ascent/descent engines, Rocket Lab's new Rutherford engine are small and throttlable (and far too bleeding edge to trust with human life).  If the rocket scientists need any "add magic to reduce mass", they might be a way out of a tight spot.  A more kerbal design might include a powerful main engine that would suicide burn to some safe altitude (100m?) and use pressure fed engines for final landing (and similarly use the main engine for ascent and the pressure feds for docking).  If you are bringing hydrogen to the moon, an RL-10 based system would make an ideal "main engine" for this (I'm no fan of hydrogen past Earth orbit).

"Two launches with lunar rendezvous" implies a lot of the stack aren't two stock Dragons, but designed for lunar operations.

[sevenperforce]

11 minutes ago, wumpus said:

"Two launches with lunar rendezvous" implies a lot of the stack aren't two stock Dragons, but designed for lunar operations.

Falcon Heavy already has the capacity to send an unmodified Dragon 2 on TLI, which will (purportedly) be used to do a free-return. With a Draco-based kick stage mated to the payload adapter in the Dragon 2's trunk, SpaceX would have the ability to send crew to lunar orbit and back. So that takes care of the easy part.

The hard part is the lander. We don't have a lander, or anything that could be easily modified to act as a lander. In theory, if we had a lander that was under 25 tonnes and less than 5 meters in diameter and could perform its own lunar-orbit insertion, then it could be dropped onto a Falcon Heavy and sent in parallel to the Dragon 2.

27 minutes ago, wumpus said:

Ideally, if the Merlins aren't enough to send the mass to lunar intercept, your first choice should be a RL-10 (hydrolox expander engine).  Storing enough hydrogen for the return flight might be iffy, but it is ideal for leaving Earth.

The fairing on FH is big, but I'm not sure if it is big enough to hold a lunar lander AND an Earth Departure stage. Hydrogen is not known for its surpassing density. 

42 minutes ago, wumpus said:

While pressure-fed hypergolics are nearly always the way to go for ascent/descent engines, Rocket Lab's new Rutherford engine are small and throttlable (and far too bleeding edge to trust with human life).  If the rocket scientists need any "add magic to reduce mass", they might be a way out of a tight spot.  A more kerbal design might include a powerful main engine that would suicide burn to some safe altitude (100m?) and use pressure fed engines for final landing (and similarly use the main engine for ascent and the pressure feds for docking).  If you are bringing hydrogen to the moon, an RL-10 based system would make an ideal "main engine" for this (I'm no fan of hydrogen past Earth orbit).

The one place you absolutely MUST have pressure-fed hypergolics is the lunar ascent engine. There's no alternative; trusting any ignition system other than hypergolics (or any kind of turbopump) with that one critical point is a non-starter.

Another very Kerbal solution (though not one without precedent; see the N-1 and the Nova) is to use a single restartable hydrolox stage for the lander's Earth departure burn, lunar injection burn, AND lunar landing burn, but drop it before the actual landing. You then use your ascent engine(s) for the actual touchdown and, later, the ascent. It's not the most rigorously efficient approach (the engine would be underpowered for the TLI burn and overpowered for the landing, and you're dragging along a bit of superfluous dry mass), but it's breathtakingly simple, and you can completely dispense with the whole mass and volume and added complexity of a formal descent stage.

Such a high-energy multipurpose stage would need between 3.2 and 5.9 km/s of dV, depending on how much of a periapsis kick the Falcon 9 upper is able to give toward TLI (the TLI costs 2.7 km/s). The RL-10 might be a good choice, but it's pricey as all hell; you could also choose a vacuum-expanded BE-3 though I don't know what the isp is like.

With hydrolox, you need a 50-60% propellant fraction to get 3.2 km/s or a 70-75% prop fraction to get 5.9 km/s. Falcon Heavy's 60 tonne+ expendable capacity to LEO would allow a loaded ascent stage mass of up to 14 tonnes (though the size of the LH2/LOX stage would be prohibitive), whereas if the Falcon upper stage performed the TLI, the hydrolox stage would only need to do lunar injection and descent, and so your loaded ascent stage budget would still be a healthy 8.4 tonnes.

For reference, the LM ascent stage was under 5 tonnes, wet. The much larger planned Altair ascent stage was around 10.8 tonnes.

 

[sevenperforce]

Of course, there's one other very, very interesting possibility.

Recall that one way of reducing the size of the Dragon 2 kick stage was to modify the Falcon 9 upper stage for delayed restart, and use it for 95% of the lunar orbit insertion burn.

IF those modifications could be made to the MVac....

Falcon Heavy, expendable, can deliver 16.8 tonnes to Mars transfer, a cost of around 4.3 km/s. This is what it'll be doing with the Roadster (fingers crossed) on Tuesday. The Falcon 9 upper stage has a dry mass of 4 tonnes and carries 107.5 tonnes of propellant. A bit of math tells us that in such a configuration, the upper stage burns 48 tonnes of propellant at an isp of 345 s to deliver that 16.8 tonnes of payload (plus 4 tonnes of dry mass) from LEO to trans-Martian injection. Reversing the equation, this tells us that the other 59.5 tonnes of propellant provide only 2.2 km/s of the dV for orbit, which places expendable staging velocity for a 16.8-tonne payload at around 5.6 km/s.

If the core and boosters of an expendable Falcon Heavy can deliver 16.8 tonnes of payload, along with the upper stage, to 5.6 km/s, then they can surely deliver a smaller payload to the same staging velocity. Let us suppose that the upper stage provides the 2.2 km/s to get into orbit, plus the 2.7 km/s for TLI, and still has enough propellant left over for BOTH the lunar insertion AND for the almost-landing burn.

That's a total of 8.13 km/s, from Earth ascent staging to lunar landing burn termination. Our handy-dandy rocket equation tells us that for a single stage at 345 s isp to deliver 8.13 km/s, it needs to have a 90.97% prop fraction. Given the 107.5 tonnes of prop in the upper stage, this tells us that a Falcon Heavy flying expendable could deliver 6.67 tonnes of payload to a zero velocity just above the lunar surface, simply by modifying the upper stage for extended-delay restarts.

The Apollo LM ascent stage, again, was under 5 tonnes.

How much, exactly, would a Dragon V2 mass if you removed the aeroshell, heat shield, Superdracos, and parachutes?

[KSK]

2 hours ago, wumpus said:

More thoughts:

Ideally, if the Merlins aren't enough to send the mass to lunar intercept, your first choice should be a RL-10 (hydrolox expander engine).  Storing enough hydrogen for the return flight might be iffy, but it is ideal for leaving Earth.

While pressure-fed hypergolics are nearly always the way to go for ascent/descent engines, Rocket Lab's new Rutherford engine are small and throttlable (and far too bleeding edge to trust with human life).  If the rocket scientists need any "add magic to reduce mass", they might be a way out of a tight spot.  A more kerbal design might include a powerful main engine that would suicide burn to some safe altitude (100m?) and use pressure fed engines for final landing...

Nah - a truly kerbal design would use a solid motor for the suicide burn. Something of a one-way ticket though.

 

[sevenperforce]

The planned Altair lander ascent stage would have had a local TWR of 2.54. The Apollo ascent module had a local TWR of 2.1. Erring on the larger side, and assuming a 6.7-tonne ascent vehicle delivered to just above the surface by Falcon Heavy, that means you'd need 27.6 kN of thrust on ascent. You'd need almost 70 Dracos for that. Or, you know, half a SuperDraco. The SuperDraco would need a vacuum-expanded nozzle to kick up the isp, which would further increase the thrust. It can throttle deeply, but I don't know how deeply.

Anyone have a guess as to the mass ratio of pressure-fed hypergolic tanks?

[sevenperforce]

One problem in going with a single SuperDraco rather than clusters of smaller engines is that your vacuum nozzle ends up taking a substantial amount of the volume in the trunk, which is a problem if you want a drop-in kick/landing stage. It also exposes your engine to debris impingement on landing, which is bad if you want to use your landing engine for ascent.

If the Dracos got larger nozzles to bump up their isp and thrust, then you could get up to 320 s and 427 N, which would require 40 Dracos to get a local TWR of 1.6. Not the MOST efficient ascent, but closer to doable. That's only twice as many Dracos as a standard Dragon 1 carries.

[magnemoe]

6 hours ago, wumpus said:

More thoughts:

Ideally, if the Merlins aren't enough to send the mass to lunar intercept, your first choice should be a RL-10 (hydrolox expander engine).  Storing enough hydrogen for the return flight might be iffy, but it is ideal for leaving Earth.

While pressure-fed hypergolics are nearly always the way to go for ascent/descent engines, Rocket Lab's new Rutherford engine are small and throttlable (and far too bleeding edge to trust with human life).  If the rocket scientists need any "add magic to reduce mass", they might be a way out of a tight spot.  A more kerbal design might include a powerful main engine that would suicide burn to some safe altitude (100m?) and use pressure fed engines for final landing (and similarly use the main engine for ascent and the pressure feds for docking).  If you are bringing hydrogen to the moon, an RL-10 based system would make an ideal "main engine" for this (I'm no fan of hydrogen past Earth orbit).

"Two launches with lunar rendezvous" implies a lot of the stack aren't two stock Dragons, but designed for lunar operations.

They use electrical pumps rather than pressure feed, electrical pumps are probably less bleeding edge than radio
Large electrical car style battery packs on the other hand are bleeding tech and not space rated. Musk the FH roadster launch need an mission addition. 
 

[sevenperforce]

8 minutes ago, magnemoe said:

They use electrical pumps rather than pressure feed, electrical pumps are probably less bleeding edge than radio

Hmm. The vacuum-expanded Rutherford dials in at 343 s isp and 24 kN.

LOX boil-off would still be a problem, but if that could be solved, a Rutherford-based drop-in kick stage for both the command-module Dragon and the lander would be just about the right size. Kerolox doesn't have quite the bulk density of hypergolics, but it still might work. And the kick stage could be tested on any Dragon 1 mission.

[sevenperforce]

Another option, one with even more minimal modifications, would be to keep two of the standard eight SuperDracos on the modified Dragon 2 lander for the hover, landing, and initial takeoff, and use a lower-thrust Draco-based kick stage for the actual burn to orbit. This would require, in theory, zero engine modifications. Using the canted SuperDracos for landing and takeoff would prevent debris impingement, and their high thrust reduces the dry mass of the kick stage. The low isp of the SuperDracos isn't a problem because they are only delivering a few hundred m/s of dV, while the bulk of the impulse is provided by the much higher-efficiency Dracos.

You'd still need landing legs on the trunk, and modifications for egress and ingress, but everything else remains virtually untouched apart from removing the ballast sled, the heat shield, the aeroshell, and the chutes.

[wumpus]

8 hours ago, sevenperforce said:

The one place you absolutely MUST have pressure-fed hypergolics is the lunar ascent engine. There's no alternative; trusting any ignition system other than hypergolics (or any kind of turbopump) with that one critical point is a non-starter.

Trans-Earth injection is equally "burn or die", but also used pressure-fed hypergolics (although possibly chosen to avoid cryogenic bleed-off).

7 hours ago, KSK said:

Nah - a truly kerbal design would use a solid motor for the suicide burn. Something of a one-way ticket though.

A high-power solid motor (to supply the majority of the thrust) followed by pressure fed systems should be sufficiently simple, but the LM ascent stage had an Isp of 311s.  It would take some pretty extreme fear of evil chemicals to switch  N2O4 - Aerozine 50 tanks out for an SRB.  These were last used as shuttle maneuvering engines and scheduled for use in the Orion, so presumably they are somewhat available (you might have to scale them back down to LM size).

3 hours ago, magnemoe said:

They use electrical pumps rather than pressure feed, electrical pumps are probably less bleeding edge than radio
Large electrical car style battery packs on the other hand are bleeding tech and not space rated. Musk the FH roadster launch need an mission addition.

Electrical pumps might have looked reasonably similar from the Tesla-Edison battles to the 1990s, but since then they have changed radically (ok, inductive motors might resemble Tesla's designs, but the power supplies typically supply a custom-created AC instead of what the power plant provides).  Probably similar things could be said about radio transmission, but I suspect the changes started much earlier.  No matter how old the tech, nobody is trusting an entirely new engine design never designed for crewed flight with a single successful flight.

[sevenperforce]

37 minutes ago, wumpus said:

Trans-Earth injection is equally "burn or die", but also used pressure-fed hypergolics (although possibly chosen to avoid cryogenic bleed-off).

Yeah, I was already assuming a hypergolic Draco-derived kick stage for the homeward journey.

[sevenperforce]

One big advantage of using the Falcon US as a crasher for the lander is that you can replace the trunk-borne propulsion unit with payload for one-way cargo missions. Crasher stage to the surface, use the SuperDracos to land, and jettison your cargo. Could be consumables, a rover, a BEAM unit, whatever.

The idea of using a drop-in propulsion unit attached to the payload adapter in the trunk has been suggested before -- it's been called a JPAP, or jettisonable propulsion assist pallet. One challenge to using a Draco-based system is that pressure-fed tanks tend to have poor mass ratios. An electrical turbopump might work, but then your dev costs skyrocket.

[sevenperforce]

I ran some more numbers, and just for reference...

Falcon Heavy, expendable, can deliver:

Reference Payloads, Falcon Heavy (expendable)

Destination	Delta-V (beyond LEO)	Payload (tonnes)
LEO	0.0 km/s	63.8
GTO	2.27 km/s	26.7
TLI	2.73 km/s	25.2
LLO*	4.04 km/s	16.7
Lunar Surface*†	5.93 km/s	9.0
TMI	4.3 km/s	16.8

*denotes requirement for extended-delay MVac restart; boil-off possible.

^†not suitable for landing; crasher stage only

Edited February 5, 2018 by sevenperforce

[magnemoe]

37 minutes ago, sevenperforce said:

I ran some more numbers, and just for reference...

Falcon Heavy, expendable, can deliver:

Reference Payloads, Falcon Heavy (expendable)

Destination	Delta-V (beyond LEO)	Payload (tonnes)
LEO	0.0 km/s	63.8
GTO	2.27 km/s	26.7
TLI	2.73 km/s	25.2
LLO*	4.04 km/s	16.7
Lunar Surface*†	5.93 km/s	9.0
TMI	4.3 km/s	16.7

*denotes requirement for extended-delay MVac restart; boil-off possible.

^†not suitable for landing; crasher stage only

Interesting as the Apollo lunar lander weight 16 ton. 

https://en.wikipedia.org/wiki/Apollo_Lunar_Module
Says that the accent module was only a bit above 2150 kg dry, 4700 kg gross who I assume is fueled and with samples and astronauts. 
Now using FH upper as an crashing stage adds an danger, upper stage has limited stay and any longer delay in the crew launch will spoil it. 

As Lunar orbit burn is just 680 m/s it might be beneficial to use LEM engines or crasher stage / drop tanks for this. 
An Lunar surface run would be most relevant for an Moon base but at this time BFR or other large rockets will be available. 

[sevenperforce]

4 minutes ago, magnemoe said:

Interesting as the Apollo lunar lander weight 16 ton. 

https://en.wikipedia.org/wiki/Apollo_Lunar_Module
Says that the accent module was only a bit above 2150 kg dry, 4700 kg gross who I assume is fueled and with samples and astronauts. 
Now using FH upper as an crashing stage adds an danger, upper stage has limited stay and any longer delay in the crew launch will spoil it. 

As Lunar orbit burn is just 680 m/s it might be beneficial to use LEM engines or crasher stage / drop tanks for this. 
An Lunar surface run would be most relevant for an Moon base but at this time BFR or other large rockets will be available. 

FUS failure still has an abort mode. The MVac suicide burn is mostly horizontal; if it fails, you can jettison the stage and burn hard-radial-out with the SuperDracos to get headed back up to orbit. Not pretty, but it gets the job done.

Abort for LOI burn failure of the crewed capsule is simple, too; the TLI burn out of LEO puts you on a free-return, so if the MVac doesn't ignite at lunar periapsis, you simply jettison it and head home.