The rocket has to be fully reusable with minimal turnaround time (hours, not days or weeks), so the new method is not designed to damage the first stage in any way. Whether it will live up to the design is currently unknown, but they cannot accept any damage at all to the booster. Not even ablative shielding.
I still have Jeb floating around somewhere in the Kerbin system. I think he’s been there for about five years. He seems fine.
Tom Mueller (ex propulsion lead at SpaceX, designer of Merlin engine) seems to think this is as much about avoiding ullage collapse as anything:
I don’t know precisely what he has in mind, but my read is that when the propellant shifts forward (since the staging event usually entails a slight deceleration), that possibly causes a huge increase in propellant surface area due to forming globules. If you can keep a nice, flat surface instead, there’s no risk of ullage collapse (or at least no more risk than usual).
Probably that’s just one of several things they gain. But possibly quite significant.
ETA: restructured my post a bit.
When the propellant (and oxidizer) sloshes forward, liquid supply to the stage is uncertain until / unless you can settle them back down. That’s bad. I know you know that, but some audience members may not.
IMO what Mueller is describing as “mass-inefficient” is designing enough baffling, tank pressurization, etc. to promptly suppress / correct the sloshing. Better to simply avoid sloshing in the first place by keeping a constant positive acceleration along the vehicle X axis.
See also
I’m a bit surprised that SpaceX is trying to drive staging by firing upper stage main engines. Maybe with extremely deep throttling or some other ducting whatsis they can achieve the effect of a vernier ullage motor as described in the wiki without incinerating the top of the soon-to-depart lower stage.
YMMV; IANA rocket engineer.
Agreed that there are other effects at work, like needing less baffling (scientists should be baffled; on rockets it’s dead mass).
But ullage collapse and “repress” has a specific meaning here related to the use of autogenous pressurization. The pressurization gas is the same as the propellant itself, and can condense back into the liquid. That will happen anyway to some extent, requiring a heat exchanger to re-gasify some proportion of the prop, but you want to minimize it. And the rate at which it happens is going to depend in part on the surface area. A nice, stable, flat surface minimizes this. Violent sloshing does not.
Hot staging isn’t a new invention; it’s been done before, both in the US and Russia. But hot staging plus reusability is certainly new.
A nice video of the Titan II doing hot staging:
They had some venting, but clearly it wasn’t quite enough, as the interstage has a somewhat energetic response to the second stage plume. But that could be fixed.
Musk says only the center Raptors will fire, and they will be throttled to 50%. Still, that’s a lot of energy in a small space, concentrated right at rhe top of a large cryo tank. it all fits with Musk’s “The best part is no part” philosophy, but it seems like they are adding a pretty big risk item. But maybe they have it all worked out.
I hope this is not the second time they misestimate the effects of constrained thrust on the surfaces below the engine nozzles.
It may just be a temporary design for testing with no expectation that the booster will survive.
Would seem a waste. But if the prime goal is testing the orbital part of the system with no plan for reuse of the booster in a future test, it would make sense.
I think the main goal is to get to orbit routinely and then tweak for re-use. Hot-staging can make use of a disposable interstage to intercept the damage to the first stage. The 2nd stage aft design will need to withstand some different temps and forces, but in the long run it doesn’t matter. Either this works or is a disaster and a change will come out of it either way. The Raptor engine is too efficient for this rocket not to be a success, at least as far as space flight is concerned.
Rocketry terminology is so much fun.
Grown-up kids playing with grown-up quantities of explosives are gonna be like that. 
They aren’t going to want a disposable interstage. This rocket has to be fully reusable with a fast turnaround, or the economics get iffy.
Remember, a single mission past Earth orbit needs at least seven flights. If you have to throw away seven interstages for each lunar or Mars mission, the economics completely change.
They made this change primarily to,gain a few percentage points of efficiency over their previous method. This only makes sense if the new method does not contribute significantly to cost or downtime.
Hiwever they are doing it, the hot staging cannot require any kind of refurbishment or expending of any expensive parts between flights, or they wouldn’t do it.
Depending on how cheap the interstage is, it might be a negligble disposable item, kind of like how windshield wipers are negligible in teh cost of operation of a car.
If it’s just a dumb plate of steel X meters in diameter surrounded by a ring of the same steel Y meters (for small values of Y) tall that may well be negligible in the overall cost of the launch. After all, the fuel is 100% expendable and that’s not no-cost either.
Agreed. SpaceX cares about two things here: cost and time. Which are sorta the same thing. In any case, if they come up with a system that is cheap and quickly replaceable (presumably after N uses, not necessarily every use), I think they’d be fine with it. Especially since they’ll be iterating on the system for a while. An ablative shield that costs $1M, lasts 10 uses, and can be replaced in a few hours isn’t a huge deal.
Rockets really change the scale at which you think about these things. I remember being amused at the old videos where George Goble lit a charcoal grill in a 3 seconds with a few gallons of liquid oxygen (Dave Barry wrote about this). It seemed pretty “energetic” at the time. Starship has almost a million gallons of liquid oxygen.
Yeah. A million gallons of LOX. The head swims.
As to the interstage, I’m envisioning something that resembles a gigantic steelpan calypso drum head*. Installed convex side up, it’s sort of a condom over the upper end of the lower stage. It’s just be a dumb hunk of metal; all the severable connections to it or through it would be installed on, and controlled by, the two stages.
When the upper stage fires as gently as possible (which is still none too gently) to climb away from the lower stage, the “condom” attached to the lower stage deflects the exhaust blast outwards. Once the stages are far enough apart either the condom is jettisoned, or is brought back home if the mass and mass distribution isn’t a problem for the landing.
I guess we’ll have to see, but I’m betting they’ll go without an interstage, at least at first.
An interstage is not that cheap, adds a fair bit of complexity, and reduces payload capcity. Because it’s not stiffened by fuel pressure, it needs lots of reinforcing to prevent buckling.
Again, the whole reason they are making this change, as I understand it, is because they want to avoid the performance loss of shutting everything down on the rocket for staging. So they want to hot stage, but if they have to add a disposable interstage to do it, it kind of defeats the point.
How much would a 9m diameter, 3m tall steel ring weigh? With reinforcements, attachment hardware, etc? I did some searching, and it appears that currently each ring on Starship (1.5m tall) weighs 1622 kg. An interstage would be at least twice that height, and would have to be insulated and strengthened. So, maybe 5,000 kg? That’s a lot of extra weight.
And Musk is shooting for a turnaround time of 1-2 hours per launch. Or at least a couple of launches per day. That’s hard to reconcile with having to install a new interstage for each flight, as opposed to just picking up Starship and placing it back on the booster.
Not so much for the booster. There’s typically a high ratio between first-stage mass and payload. Maybe 10:1 in the usual case, though probably a bit less for a return-to-launch-site booster (say 5:1).
The ullage gas itself has significant mass at MECO. Considering just the booster alone, there’s probably at least 2400 m^3 of volume for LOX, which at 6 bar and 500 K is ~11 tons. That could shoot up far higher if it got chilled due to slosh. Can’t be ignored.
Some speculation is in agreement with that idea:
Will the secondary dome have some extra protection? Heat tiles, or maybe just made from niobium? We shall see.
Can they even refuel that quickly? That kind of turnaround time just sounds absurd. Even if you want launches that frequent (and what mission could even justify that?), surely it’d be easier to accomplish with a pipeline of multiple vehicles (you launch A while B and C are still being prepped, and so on).
And is there even a need for such a turnaround in the foreseeable future?
I’m guessing SpaceX isn’t designing for reuse at this time, mainly getting the cost/kg and really the cost/Starlink capacity, minimized. After that the reuse trades start to make sense. I think it’s pretty cool an internal component of the business is the one putting pressure on Starship to be efficient as possible. At the same time I wonder if there is a decent market for such a cheap delivery service to orbit.