SpaceX Starship - IFT-3 - Starbase TX - Launch Thread and Post Launch Discussion

Launch Date: March 14
Launch Window: 7:00 am CDT (9:00 am EST, 13:00 UTC)
Launch site: LC-1 - Starbase, Boca Chica Beach, Texas
Core Booster Recovery: Expended in Gulf with a landing burn
Starship Recovery: A controlled reentry through the atmosphere to a terminal velocity splashdown in the Indian Ocean
Booster: Super Heavy Booster 10
Starship: Starship 28
Mass: No mass simulator mentioned
Orbit: LEO-ish
Yearly Launch Number: 26

A SpaceX Super Heavy and Starship launch vehicle will launch on its third not quite orbital integrated flight test designated IFT-3. The mission will attempt to place Starship into a nearly orbital trajectory that will attempt a controlled reentry through the atmosphere to a terminal velocity splashdown in the Indian Ocean . The Super Heavy booster will attempt a landing burn in the the Gulf of Mexico where it will likely be destroyed. This is a further test of Stage 0, the booster, full power ascent, Max-Q, stage separation using the new hot staging, a booster stage test of a hard turn and boostback, full burn boost of Starship to space and sub LEO, Starship will do one partial orbit, simulate a de-orbit burn, test tiles and heating from atmospheric reentry, until it has a splashdown in the Indian Ocean.

It has also been determined that for this test flight there will be a fuel transfer test done on Starship for NASA's Tipping Part contract. The Starship will also test its payload bay door in zero-G for a test of future Starlink 2.0 deployments.

 

[JB47394]

That's iffy currently. 1,500 tf vs 1,200 t of propellant, 150t of cargo, and 160 - 200t dry mass.
HLS landing/ ascent engines may be an option with a breakaway cargo section.

Could possibly ditch into the gulf and tip over.

Are you thinking of a cargo mission or a crewed mission? I'm not sure that a ditch in the Gulf would be so great for crew.

Aren't abort motors usually something that can ignite instantly, and not a few seconds later?

Yep. Hypergolics or solid motors. Like @ecarfan says, they'd have to leave the engines chilled and ready to go through the entire flight. I assume that would consume a certain amount of liquid oxygen, reducing maximum delta-V. But I have no idea how much.

Dragon's escape system kicks the capsule away at a peak of 3.3 g. We're not going to see that from a Starship, even with nine engines. Then there's the potential for those exposed engines to get wrecked in the event of a booster failure. The Apollo escape tower was on the nose of the stack, and Dragon's are integral to the capsule, on the sides. Perhaps the inert hot staging ring would provide sufficient protection.

 

[TunaBug]

, they'd have to leave the engines chilled and ready to go through the entire flight.

Wouldn't there still be 2 or 3 second delay to startup even if they were chilled and ready to go?

I won't be surprised to find out that SpaceX has an awesome answer for this. But right now it looks to me like the startship's abort story resembles the space shuttle's abort story.

 

[JB47394]

Wouldn't there still be 2 or 3 second delay to startup even if they were chilled and ready to go?

There's clearly a delay between pushing the button and generating thrust. I timed this video of an engine test and it appears to be 1.25 seconds. Not so great for a pad abort, but probably okay for an ascent abort.

I won't be surprised to find out that SpaceX has an awesome answer for this. But right now it looks to me like the startship's abort story resembles the space shuttle's abort story.

That's a good point. I suspect that with their many-engine approach, SpaceX should have much greater reliability with the Starship system than any prior system. Very few Merlin engines have failed, and I know of no missions failing due to a loss.

 

[TunaBug]

There's clearly a delay between pushing the button and generating thrust. I timed this video of an engine test and it appears to be 1.25 seconds. Not so great for a pad abort, but probably okay for an ascent abort.


That's a good point. I suspect that with their many-engine approach, SpaceX should have much greater reliability with the Starship system than any prior system. Very few Merlin engines have failed, and I know of no missions failing due to a loss.

Well you point out something I haven't considered: What are the scenarios to recover from? I wonder what NASA's requirements/goals for abort systems look like?

 

[JB47394]

Well you point out something I haven't considered: What are the scenarios to recover from? I wonder what NASA's requirements/goals for abort systems look like?

I found this PDF from NASA on man-rating. It's somebody's presentation and not a technical document, so be forewarned. Here's one informative slide.

NPR - NASA Procedural Requirements document
JSC - Johnson Space Center document

The whole thing suggests to me that crew survivability has to be extremely high, however the system achieves that. I'm not sure if the booster exploding 10 feet off the launch mount is something that SpaceX has to directly address. They may be able to quote the reliability of the booster technology and be done. I assume that's what they did with the Space Shuttle.

Here are some interesting bits:

"Rev A [of NPR 8705.2] also led to the requirement that abort be accomplished without use of main propulsion and during loss of control scenarios."

"Rev B removed all references to escape and prescriptive failures on entry"

The latter was added for Constellation because it wasn't a Space Shuttle. Requiring escape from Constellation just wasn't practical.

Mostly, they seem to be making it up as they go, but using their institutional experience to know what to allow and what to reject. I could be wrong, and there may be very precise statements about what can and cannot be done, but I'm sure that the requirements will flex as new launch systems show up - like Starship.

Ultimately, NASA may say that Starship itself cannot be human-rated and that SpaceX has to come up with a new vehicle that is pretty much guaranteed to spare the lives of the crew in case of problems. I assume that SpaceX would not then be able to launch a Starship with crew on it (even from Starbase) because they'd never get the FAA license.

 

[Ben W]

If it were vertical and not turning, I'd agree. It's at a pretty good angle relative to the gravity vector (it looks like it's about 30 degrees off horizontal), and then it does that hard turn to get away from Starship. The propellant is certainly going to slosh, and I assume that the negative acceleration only makes it more pronounced.

Yes, the hard turn will certainly make the propellant slosh sideways. My point was that the "-.4g" doesn't mean the propellant is going to "fall toward the nose", which would undoubtedly create huge problems.

I wonder what they'd do with a fully loaded Starship after a pad abort. Hover the thing until it's light enough to be caught by the chopsticks?

Yeah, my assumption is that it would go up a mile or so and hover until it's light enough to be caught by a nearby catch tower. Probably no need to bellyflop like the early tests, just a continuous hover. It may even make sense to build dedicated Starship catch towers lower to the ground, so that in the case of a chopsticks "miss" it wouldn't have as far to fall? (Or risk damaging the high-value OLM?) A while back Elon had even mentioned the possibility of Starship being able to land on its skirt on solid ground in an emergency, which would cause damage to the engines but be survivable. Not sure if that's still being designed for. A survivable water landing mode for Starship is another possibility.

 

[Ben W]

Yes, the problem with a pad abort that blasted the ship up and away from the booster is that currently the ship has no legs. And since the abort could be triggered by a malfunctioning booster on the OLM that could damage the tower and the chopsticks system, a hover and chopsticks catch may well not be an option.

So…could four minimalist emergency landing legs be incorporated into the base of the booster so the ship could do a controlled propulsive descent onto a flat, firm surface? The legs would add weight and complexity but they would only be designed for a single use so do not have to be to robust.

Every launch site would presumably have multiple launch/catch towers, partially for this reason. And Elon at one point had mentioned that Starship would be capable of an emergency landing on its skirt on solid ground, which would cause engine damage but be survivable.

That said, can all 6 ship engines propel a fully loaded (props plus cargo) ship rapidly away from the booster starting from 0kph at sea level? And, how quickly can the engines ignite after a pad abort is triggered? The engines would have to be pre-chilled and ready for ignition at around T-0, and then presumably kept in that state up to staging. What problems would that present?

It seems to me there are a lot of potential obstacles to be overcome, but it’s an idea seriously worth considering.

Eventually Starship will have 9 engines, which should make such an abort easier. A stretched staging ring could increase the getaway thrust, and crewed Starships could have lighter payload and less fuel to provide extra safety margin for such an abort. I'm not sure what the expected behavior would be for lighting up nine Raptors without a pre-chill; would they necessarily fail, or would it just decrease their expected service life? Maybe the pre-chill could be done with minimal penalty, and also, the rapid-start ability may only be necessary for the first several seconds of flight. (Above say 1km, if the interstage provides reasonable physical protection from a booster failure, it might not have to detach immediately, so the Raptors would have more time to spin up.) Obviously a Starship RUD would be unsurvivable, which makes it different from a Falcon second-stage RUD. It will be very interesting to see what safety modifications are added to the crew variant.

 

[JB47394]

My point was that the "-.4g" doesn't mean the propellant is going to "fall toward the nose", which would undoubtedly create huge problems.

Even being angled at 30 degrees without any negative acceleration is going to expose the LOX inlets on the "top" engines with so little propellant in the tanks.

It may even make sense to build dedicated Starship catch towers lower to the ground, so that in the case of a chopsticks "miss" it wouldn't have as far to fall? (Or risk damaging the high-value OLM?)

The weird thing is that the chopsticks can already get low enough to pick up a Starship on a transport stand, but they do so on the tank farm side. If they could get the chopsticks to lower on the far side, they could catch Starships without it hovering over the launch mount or being anywhere near the tank farm. Unfortunately, the ship quick disconnect is on that side. I thought it an odd arrangement.

Though even the idea of catching these things while in hover is odd.

A while back Elon had even mentioned the possibility of Starship being able to land on its skirt on solid ground in an emergency, which would cause damage to the engines but be survivable.

At an "unimproved airfield", I suspect they might try that no matter what. But when at Starbase or at the Cape, they could put in all sorts of infrastructure to aid in Starship landings (no additional launch mass). I wonder if they can put in a flame trench like the one at Massey's, but make the top of the mount move around by a couple meters and rotate, allowing it to steer under the landing ship. Then they could run the water deluge system during a landing.

Which make me wonder if they'll run the water deluge system for a booster catch.

 

[mongo]

Even being angled at 30 degrees without any negative acceleration is going to expose the LOX inlets on the "top" engines with so little propellant in the tanks.

Would they? Engine inlets are built into tank bottom which is convex (viewed from outside). If hovering, 30 degrees means a sin(30)*4.5=2.25 meter rise/drop from center level to wall (ignoring volumetric changes). Inner ring would see about half the level change or 1-1.5 meters.

Any ship thrust would reduce the fluid slope angle.

 

[JB47394]

Would they? Engine inlets are built into tank bottom which is convex (viewed from outside). If hovering, 30 degrees means a sin(30)*4.5=2.25 meter rise/drop from center level to wall (ignoring volumetric changes). Inner ring would see about half the level change or 1-1.5 meters.

Sure, when it's sitting still. What happens when there is both negative and positive acceleration plus rotation? That IFT-2 simulation showed a lot of sloshing. That said, sloshing may simply not be a problem for them. The plumbing may be designed to clear gas voids in the lines as soon as there's positive acceleration. I was just observing that being canted over at a sharp angle and applying that negative acceleration is going to bias towards exposing those inlets. Only a simulation would help to understand what happens, and even then it wouldn't be conclusive. We need SpaceX's footage from inside the tanks. It's terribly inconsiderate of them to not provide it.

As a bit more commentary, they got good relights on all the interior engines for the boostback. The shutdown order at the end of that burn was very odd, however, with the engines turning off from right to left in the middle ring. Not at all like the relight sequence, which was the usual pairings. Then the only engines that restarted (however briefly) during landing were on the left side. Something happened, but I don't have a smoking gun to suggest what it might have been. I wouldn't think that gas bubbles in the plumbing would allow a clean boostback burn. Which they got.

 

[Ben W]

Even being angled at 30 degrees without any negative acceleration is going to expose the LOX inlets on the "top" engines with so little propellant in the tanks.

It doesn't work that way. When Starship is accelerating at a 30-degree angle at stage separation, gravity makes the rocket "fall" ("gravity losses") in a way that exactly counterbalances the tilt, so the propellant stays aligned with the bottom of the tank. (Similarly, if the engines shut off, the fuel would float; it would not be pulled to the earthward-side of the tank, because there would be no force pushing the rocket "up" at that point. Or put another way, both the propellant and the tank would be pulled to the earthward-side of the tank, so it cancels out.)

The weird thing is that the chopsticks can already get low enough to pick up a Starship on a transport stand, but they do so on the tank farm side. If they could get the chopsticks to lower on the far side, they could catch Starships without it hovering over the launch mount or being anywhere near the tank farm. Unfortunately, the ship quick disconnect is on that side. I thought it an odd arrangement.

Good point that the chopsticks can be lowered far enough to pick up Starship! Still, I think though that until their catch rate reliability approaches 100% (after dozens if not hundreds of flights) I doubt they would risk the primary orbital launch mount for the catch. But who knows?

At an "unimproved airfield", I suspect they might try that no matter what. But when at Starbase or at the Cape, they could put in all sorts of infrastructure to aid in Starship landings (no additional launch mass). I wonder if they can put in a flame trench like the one at Massey's, but make the top of the mount move around by a couple meters and rotate, allowing it to steer under the landing ship. Then they could run the water deluge system during a landing.

Which make me wonder if they'll run the water deluge system for a booster catch.

It's hard to envision a situation in which the landing would be controlled enough to use such infrastructure but wouldn't at least attempt to use the chopsticks. There would have to be enough extra fuel so that if the chopsticks catch were attempted and failed (it would probably have to be due to chopstick failure, but maybe strong winds?), the ship would have enough remaining propellant to glide sideways to the emergency pad and soft-land. I would guess that the flat landing would be more likely in a mid-ascent abort scenario where Starship doesn't have enough delta-V to RTLS. (E.g. halfway into ascent it loses too many R-Vacs to achieve orbit, but can still reenter and soft-land with fewer engines.)

 

[scaesare]

Even being angled at 30 degrees without any negative acceleration is going to expose the LOX inlets on the "top" engines with so little propellant in the tanks.


The weird thing is that the chopsticks can already get low enough to pick up a Starship on a transport stand, but they do so on the tank farm side. If they could get the chopsticks to lower on the far side, they could catch Starships without it hovering over the launch mount or being anywhere near the tank farm. Unfortunately, the ship quick disconnect is on that side. I thought it an odd arrangement.

Though even the idea of catching these things while in hover is odd.



At an "unimproved airfield", I suspect they might try that no matter what. But when at Starbase or at the Cape, they could put in all sorts of infrastructure to aid in Starship landings (no additional launch mass). I wonder if they can put in a flame trench like the one at Massey's, but make the top of the mount move around by a couple meters and rotate, allowing it to steer under the landing ship. Then they could run the water deluge system during a landing.

Which make me wonder if they'll run the water deluge system for a booster catch.

I mean, this is where the Giant Bouncy House comes in to play, right?

 

[scaesare]

Every launch site would presumably have multiple launch/catch towers, partially for this reason. And Elon at one point had mentioned that Starship would be capable of an emergency landing on its skirt on solid ground, which would cause engine damage but be survivable.

Eventually Starship will have 9 engines, which should make such an abort easier. A stretched staging ring could increase the getaway thrust, and crewed Starships could have lighter payload and less fuel to provide extra safety margin for such an abort. I'm not sure what the expected behavior would be for lighting up nine Raptors without a pre-chill; would they necessarily fail, or would it just decrease their expected service life? Maybe the pre-chill could be done with minimal penalty, and also, the rapid-start ability may only be necessary for the first several seconds of flight. (Above say 1km, if the interstage provides reasonable physical protection from a booster failure, it might not have to detach immediately, so the Raptors would have more time to spin up.) Obviously a Starship RUD would be unsurvivable, which makes it different from a Falcon second-stage RUD. It will be very interesting to see what safety modifications are added to the crew variant.

I know Elon had eXclaimed about 3 more engines to get to 42:

https://twitter.com/x/status/1678205301349130240

But have we heard any concrete plans to actually do this?

 

[JB47394]

When Starship is accelerating at a 30-degree angle at stage separation, gravity makes the rocket "fall" ("gravity losses") in a way that exactly counterbalances the tilt, so the propellant stays aligned with the bottom of the tank. (Similarly, if the engines shut off, the fuel would float; it would not be pulled to the earthward-side of the tank, because there would be no force pushing the rocket "up" at that point. Or put another way, both the propellant and the tank would be pulled to the earthward-side of the tank, so it cancels out.)

Yep. Good points. But continue the scenario with a reduction in speed of the vehicle that produces a 0.4g negative acceleration. That's the point where it gets interesting to me. What happens to the propellant?

I doubt they would risk the primary orbital launch mount for the catch. But who knows?

Elon has stated that the goal is to catch the booster, put it right back on the launch mount, load it and launch it. He wants to launch something like hourly.

I can kinda see a quick catch-and-stack for stuff like tankers. They'd be very light when returning, relatively speaking, and wouldn't offer any more risk than a booster. Maybe having the booster sitting there might cramp things a bit, but the whole thing is bonkers anyway.

A stretched staging ring could increase the getaway thrust, and crewed Starships could have lighter payload and less fuel to provide extra safety margin for such an abort.

The Soviets used an open triangular grid to get spacing for their hot staging. I wonder if crewed Starships could use something similar and take it with them to orbit. Use it for emergency landings at unimproved spaceports. I'd justify the mass penalty for safety, of course, but if the Starship that puts crew in orbit is designed specifically for that, then it's just a Super Dragon anyway, and there's no need to pack as many people into it as possible. They'd all be transferred to stations, Mars-bound Starships or whatever.

But have we heard any concrete plans to actually do this?

C'mon. Look at those gaps. The gaps, man.

I assume that nine engines will be required for the stretched Starship. As @mongo always reminds me, you want to keep your gravity losses down, so more thrust is better. I would expect a stretched Starship to stage at a lower altitude, making that more of an issue with that Starship than the current one.

 

[mongo]

Yep. Good points. But continue the scenario with a reduction in speed of the vehicle that produces a 0.4g negative acceleration. That's the point where it gets interesting to me. What happens to the propellant?

Engine force is less than 9.8m/s²*mass so vertical velocity is lost, but it's not -0.4g in the axial direction to unsettle the fluids.
Gravity is acting on structure and fluid equally, but engines act on structure which act on fluid keeping it in place. (I think, I'm overdue for lunch)

As @mongo always reminds me, you want to keep your gravity losses down, so more thrust is better.

Yeah, less is better, but it's a weird trajectory thing on how much gravity loss you incur...
If going vertical, more thrust is better, of going horizontal, thrust doesn't matter (unless you are at risk of hitting the ground).

 

[Ben W]

Yep. Good points. But continue the scenario with a reduction in speed of the vehicle that produces a 0.4g negative acceleration. That's the point where it gets interesting to me. What happens to the propellant?

There are no sideways forces on the rocket during hot-staging (apart from gravity, which acts on rocket and fuel equally so it cancels out), so the propellant doesn't move. In the frame of the rocket (in an equivalence-principle sense), the ship is still accelerating forward at +0.5g, even though in the Earth's frame it's -0.4g. (Slowing down, but not as fast as actual falling.) Of course, during the flip maneuver there is rotational movement so the propellant sloshes sideways, but that occurs after hot-staging and the Starship-caused deceleration.

Elon has stated that the goal is to catch the booster, put it right back on the launch mount, load it and launch it. He wants to launch something like hourly.

Yes, eventually. But I wouldn't expect this until the catch reliability is proven, with dozens or even hundreds of successful attempts in a row.

I can kinda see a quick catch-and-stack for stuff like tankers. They'd be very light when returning, relatively speaking, and wouldn't offer any more risk than a booster. Maybe having the booster sitting there might cramp things a bit, but the whole thing is bonkers anyway.

Having an empty booster nearby shouldn't be a problem. The ship and booster fuel at roughly the same rate, so there's no advantage to pre-filling the booster. Either way, I don't think they would catch Starship with the booster on the launch mount directly beneath it! Although this would be amazing to see.

The Soviets used an open triangular grid to get spacing for their hot staging. I wonder if crewed Starships could use something similar and take it with them to orbit. Use it for emergency landings at unimproved spaceports. I'd justify the mass penalty for safety, of course, but if the Starship that puts crew in orbit is designed specifically for that, then it's just a Super Dragon anyway, and there's no need to pack as many people into it as possible. They'd all be transferred to stations, Mars-bound Starships or whatever.

Interesting idea to have part of the interstage stay attached to Starship to use as an emergency landing skirt. Perhaps part of the interstage could be the literal emergency landing legs with crush core, slotted into the main interstage ring that stays with the booster.

C'mon. Look at those gaps. The gaps, man.

I assume that nine engines will be required for the stretched Starship. As @mongo always reminds me, you want to keep your gravity losses down, so more thrust is better. I would expect a stretched Starship to stage at a lower altitude, making that more of an issue with that Starship than the current one.

Nine engines will enable the stretched Starship, and a correspondingly stretched booster should enable staging at a similar altitude. Or, as discussed, nine engines on a non-stretched Starship might enable additional abort modes or margin for crewed flights. I wouldn't be surprised if a six-engine non-stretched Starship remains an option for smaller payloads, but I also wonder how the design might work out for a smaller fully reusable stack with Falcon 9 or Falcon Heavy equivalent payload capability. Could that be done with just shorter stages, or would it be better to scale all dimensions proportionally?

 

[JB47394]

Here's a link to Marcus House's latest information about IFT-3. He talks about the location and purpose of various vents, shows a new visualization of the way Starship reentered, and some other stuff. There's good information in the full video on a variety of topics.

 

[ecarfan]

What I dislike about Marcus is the non-stop hyperbole and breathless excited tone that he injects into every sentence. He’s over the top.

“SpaceX’s Frantic Push…”. SpaceX always operates at high speed and is always pushing, for the past 20+ years in fact.

“WOW THIS IS NUTS” is a silly title that imparts no information.

I have watched his videos in the past and came away feeling like I could have learned the same information in a fraction of the time if he could just edit himself.

In contrast, Scott Manley is focused, knowledgeable, informative, does his research, and sometimes includes a few low key jokes to lighten things up. The perfect combination.

 

[scaesare]

Another version of the 3-view render video, but with location and a data overlay.

The heating starts to be visible right at the 100Km Karman Line point... (about 13:50 in the video)

 

[scaesare]

Booster re-entry and RUD render:

https://twitter.com/x/status/1775174741986300298