Replacing SLS/Orion using Starship HLS and Crew Dragon

[TheRadicalModerate]

Bat chart showing how to manage a D2 in lunar orbit, so two Starships never need to dock nose-to-nose.

List of RPODs:

1) One in LEO, after the D2 launch, to the ferry.  Ferry takes crew and D2 to lunar orbit, rendezvous with HLS.

2) Crew boards D2.  Undock from ferry and do RPOD (well, really more like POD) to HLS, to move crew from ferry to HLS.  Then D2 undocks and free-flies for the duration of the lunar surface mission.

3) POD back to HLS.  Crew boards D2.  Undock.

4) POD to ferry.  Crew boards ferry and ferry+D2 go back to LEO.  Crew re-boards D2, undocks, and does EDL.

So it's the same number of RPODs as in the case where the D2 stays in LEO.

[KWC]

Out of curiosity, is there a reason you can't just use one starship fitted for crew AND lunar landing AND Earth entry with tankers? Wouldn't that simplify everything a lot? Or am I missing something huge? You already need something similar for any Mars missions, right?

[Robotbeat]

Out of curiosity, is there a reason you can't just use one starship fitted for crew AND lunar landing AND Earth entry with tankers? Wouldn't that simplify everything a lot? Or am I missing something huge? You already need something similar for any Mars missions, right?

makes it a lot heavier to bring the TPS with you to the surface of the Moon. It’d require significantly more launches and probably also refueling locations.

[TheRadicalModerate]

Out of curiosity, is there a reason you can't just use one starship fitted for crew AND lunar landing AND Earth entry with tankers? Wouldn't that simplify everything a lot? Or am I missing something huge? You already need something similar for any Mars missions, right?

That requires:

1) Starship needs to be crew-certified for launch and EDL.

2) You have a Starship whose lunar landing thrusters don't impinge on the tiles and destroy them.  (This is also required for Mars, probably, but we're at least three, likely five, synods away from an attempt at a crewed Mars landing.  For uncrewed landings, you don't care if you destroy the tiles.)

3) The solar power system has some way of deploying over the tiles.  (The Sun moves around in a circle, so unless you're willing to land with a particular orientation, you need panels that get power from any azimuth.  See here for thoughts on what's going on.)

All of those things are possible, and likely desirable, in the long run.  But they're not going to be ready by Artemis 4.

[Robotbeat]

Starship basically has to do all those things for Mars anyway. Except for the first one, but the first one is largely a subset for starship HLS…

[TheRadicalModerate]

Starship basically has to do all those things for Mars anyway. Except for the first one, but the first one is largely a subset for starship HLS…

It's all in the timing.  The LSS + D2 architecture is pretty much ready to go the day that HLS is certified.  Earth launch / EDL, the 360º access solar array wings, and the landing thrusters making nice with the TPS are going to be significant work.  And none of them are required for Mars EDL tests and presupply.

[TheRadicalModerate]

I think Jon Goff proved that the RAAN problem isn't nearly as bad as we thought it was.  This dramatically simplifies the various conops we've been discussing, and eliminates any need to haul the D2 along for the ride on the HLS Starship.

[TheRadicalModerate]

If we want to use a Starship HLS as the LEO-LLO-LEO ferry and still have a worst-case 90º plane change for rotating the RAAN, we have to figure out how much extra delta-v we have beyond the basic LEO-LLO-LEO trip, then determine how high a 3-burn apogee would have to be.  The higher it is, of course, the longer it takes to get the crew home if we're doing an abort.

Block 2 LSS has about 350m/s¹ of extra delta-v if it's completely filled.  That would require a 300 x 285,000km orbit to get an apogee capable of doing the plane change, which would have a period of 6.6 days.

Block 3 LSS, completely full, would have 1180m/s of extra delta-v.  That would require a 300 x 78,000km orbit, which would have a period of about 27 hours.

Neither of these are great numbers.  They're worst case, but that's a lot of extra time added to an abort.  So we've made the RAAN problem much better, but it's still a bit of an impediment.

__________
¹Lots of assumptions about dry mass, Isp, boil-off, etc. baked into these.  I always assume 5t of extra up-mass and 1t of down-mass, with a 15t crew module.

[sdsds]

I don't question @jongoff's orbital mechanics, and still I wonder about the timeline for those maneuvers. If the trick uses what's essentially a GTO does that add ~24 hours to the contingency return schedule?

[TheRadicalModerate]

I don't question @jongoff's orbital mechanics, and still I wonder about the timeline for those maneuvers. If the trick uses what's essentially a GTO does that add ~24 hours to the contingency return schedule?

Yes, that's the issue I worry about.  BTW, GTO is only an 11-hour orbit, but some of the worst-case RAAN mismatches could require an orbit so high that they more than double the time back to LEO.  (Note that most TEIs are forced insertions into orbits that have apogees much higher than lunar distance, and therefore take less time to reach Earth.)

I still can't figure out the geometry of the rotation.  Imagine a football (the American kind) stuck through a pizza plate.  If it's stuck in at a 90º angle and you rotate the seams, they'll track along the outside of the pizza plate 1-for-1.  But if the football is stuck in at a 5º angle, then they'll race along the outside of the plate much quicker.  This is a piece of analytic geometry that's well beyond my capabilities, but it could be a big deal for the feasibility of matching RAANs, especially since translunar inclinations or generally in the 25-30º range.

[Twark_Main]

Anyone have a hard, downloadable copy of the spreadsheets?

Google has apparently decided to start putting up a hard loginwall when accessing any Google Docs file. Also it would be nice for our future-forum-readers if we had some version that won't vanish into the ether when Google decides to ban TRM's account and/or make breaking changes and/or discontinue yet another service. 

Yes, Google is failing pretty hard on the whole "don't be evil" mission statement.  But evil is always relative, which is why my last copy of Office is from 2016, and doesn't have the LET function in it, of which I've grown very fond.  So a naive export failed.

Spun up a VM just for this. You're welcome future historians.

[Greg Hullender]

Spun up a VM just for this. You're welcome future historians.

Thanks! But are there supposed to be a lot of #NAME errors in the Prop to VLEO column?

[Twark_Main]

Spun up a VM just for this. You're welcome future historians.

Thanks! But are there supposed to be a lot of #NAME errors in the Prop to VLEO column?

Probably not, but this is all Google will give us.  You can thank Evil.