My more important threads about mostly rocket science listed here.
Disclaimer: keep the posting date in mind. Things might be outdated by now.
And: Keep criticising! I'm happy to learn!
A development thread: Given the recent discussion, I think I should explain the
@SpaceX
development style one more time. Unlike other rocket developments, SpaceX applies some form of Agile Development. This is an approach usually used in IT. 1/10
Ok, 2 things:
- The flames are scary! Is this the failing engine?
- And why did they not show the actual touch down and the falling over? That would be so much cooler!
Still great footage.
FFSC/Raptor thread: Recently, Musk reported that
@SpaceX
achieved 350 bar chamber pressure for Raptor 3. Impressive. Though, is there a limit to the maximum pressure? Or will this go up indefinitely? Does it make sense to further increase that value? 1/
Remember
@SpaceX
's ITS image? Impressive Isp, isn't it? Currently, a lot of us would be glad, if Raptor sea-level (SL) achieves at least 341 s and RVac 363 s. But what needs to change to achieve 361 s for Raptor SL and 382 s for RVac?
(1/9)
@waitbutwhy
Max payload of Starship V1 in expendable mode (like the other rockets) is ~200 tons.
V3 is expected to be ~200 tons with full reusability and ~400 tons expendable. Length will grow by 20 to 30 meters and thrust to ~10k tons.
Just to give you an idea what a fully filled Starship could deliver onto a transfer orbit to major celestial bodies in the Solar System.
This assumes 1,650 t prop, 120 t dry, 365 s Isp, 1% residuals.
Getting rid of flaps, shield and fairing would likely add >60 t!
Why did
@SpaceX
chose to stick with a 9 m Booster? This requires Raptor 2 to have nozzle exit pressure of 0.97 bar (ER 30, 1.3 m). Raptor 2 with 0.53 bar (ER 48, 1.63 m) would perform better. Needs 11.5 m Booster though. Just ~9 m/s more air drag loss but wider stance, too.
Catching might need more propellant than saved legs mass. 10s at 20m/s² = 100m/s actual dv + 100m/s gravity losses. Suiciding = 5s at 30m/s² = 100m/s actual dv + 50m/s gravity losses -> 50m/s delta at ~200t and 335s Isp = 3t propellant + hovering prop.
So complain as much as you want. But even if not all goals are reached, even some MVPs, a huge, cheap and quickly build LEO rocket, will fall out as a usable result. And will likely be better than 1.5 stage SLS, 2 stage Saturn V, Space Shuttle, Angara, you name it... 10/10
So, about
@stoke_space
's Stoker here.
@AndyLapsa
said, it's close to 100 ft. Crunching numbers with some assumptions, this thing might make sense. 2% payload share for fully reusable vehicle with ~200 t is impressive - if I'm right. Corrections? And how much prop for re-entry?
Quick summary of some aspects:
1. Only one damaged flap means the others did something right. That indicates a kinda working concept.
2. It did survive until the very end. Crucial for crew! Supports the choice of steel. And gives a sad hat tip to Columbia.
(1/4)
Despite loss of many tiles and a damaged flap, Starship made it all the way to a soft landing in the ocean!
Congratulations
@SpaceX
team on an epic achievement!!
So it's confirmed SpaceX is using pre-burner exhaust gas for re-pressurization and this resulted in water&CO2 ice clogging things on F3. Having sieves feels riskier than extra work on warming up pure gas. At 5 bar, the density ratio chart looks like this.
(1/3)
Rocket propellant production thread: Today, there's no doubt that going for methane or propane is key when considering propellant costs. But there might be a time when no fossil sourcing is allowed anymore. Then, electricity might be the main resource for propellant. 1/
Blue HLS thread: NASA decided recently on letting
@blueorigin
and team building and flying the 2nd HLS for some of the Artemis missions.
@DJSnM
summarized details quite well, here:
Yet, there are a few details, I want to discuss. 1/
@StarshipFairing
has quite ambitious plans for Starship. Something like an SSTO. Is that even possible? Maybe, when some additional improvements made for Raptor and some more mass is shaven of Ship itself. 100x200 km insertion + spare for circularization.
If Elon is right with >300 bar, a 9 engine Ship with 300 t more propellant could target 220 t to 200x200 km / 32°. Good thing: Booster would need less boost back propellant. Obviously, we gotta see if I'm right with my mass estimates. And I'm still not sure about SL Isp. Anyone?
And every explosion as a result of limit testing will mark a potential point of failure and ask for fixing instead of hiding in margin and showing up once people flying with it and it's too late. 9/10
Aeon R thread: Remember the open cycle thread? Yeah, I'm still not a big fan of that cycle. But,
@thetimellis
once sad,
@relativityspace
will run Aeon R at a surprisingly high main chamber pressure. That got me thinking. What could that value be? 1/
Open cycle engine thread: Those engines are a bit tricky if you want to understand their efficiency a bit better. The problem is the gas generator. Usually you don't know the combustion temperature and even then, finding the associated oxidizer-to-fuel ratio (O/F) is hard. 1/x
Given the quickly and constantly changing requirements by the client who sometimes doesn't even know what she actually wants, a lengthy development period doesn't help in IT: Once something is done, after 2 years or so, it might not fit the requirements anymore. 2/10
Here’s the static fire video! This all-up test was really a hop mission simulation and included everything from flight avionics, power systems, computers, GNC, RCS, tank pressurization, and, of course, the engine and heat shield. The only thing we simulated was the position data,
Is this all an excuse for the very optimistic timelines and the obvious failure to achieve them? No. But it's an reminder that the messy testing campaign is not an indicator of general failure - it's just part of the plan to succeed. And there will be much more explosions. 8/10
Therefore, IT started to try to get as much feedback as possible during development, to change the direction early enough if needed. This requires the so called minimal viable product, the MVP. It's a product that at least partially already does what the client asked for. 3/10
We gotta talk about the fact that Booster separated at just ~1.5 km/s while it should've burned through 3.2 km/s. At this point, 1.7 km/s gravity losses is way too much. So it wasn't completely filled or much more left for boost back than expected?
With many small iterations and fast feedback loops, you can also test the limits of a product much easier. Because going to one limit doesn't ruin the whole 2 year plan but you can easily cycle back a few steps and adapt to the new reality. 5/10
The 3 flights show clear progress. And their quick idea to use hot-staging did work out technically already with the 3rd flight. Whether it will achieve its mass saving goals is up to further tests, I guess. 7/10
This can be very sketchy but it already allows the client to use and test it and adjust her vision about the final product while the developers see what works and what doesn't. That fail hard and fail fast is modern trial and error. 4/10
Did a bit of wiggling with given data and estimates. One major problem of
@relativityspace
's Terran R could be underpowered upper stage. Prop mass estimates came from estimating the tank size from Terran R picture. I'm even more disappointed. More analysis to come though.
This doesn't remove the need to plan but it allows to adjust plans and integrate lucky innovations much faster. Now, complaining about the 3 Starship stack flights when looking at the test campaign and the results of Falcon 9 might not be helpful. 6/10
Efficiency of a rocket engine is quite a flexible term. Propellant, TWR, ER, cycle - they all play a crucial part. Yet, even if everything is perfect, there's just a limit to the maximum Isp possible. Looking at propellant, there's the obvious limit of energy stored in it. 1/x
Open cycle engine thread: Those engines are a bit tricky if you want to understand their efficiency a bit better. The problem is the gas generator. Usually you don't know the combustion temperature and even then, finding the associated oxidizer-to-fuel ratio (O/F) is hard. 1/x
7. Not only did the steel structure survive the heat but also the actuators. Given the brutal environment, that's even more impressive.
8. As far as I can see,
@FAANews
doesn't have to investigate. So flight 5 soon?
(4/4)
@StarshipFairing
and I have discussed semi-reusable high payload/dv rockets on the Starship architecture. We touched the extremes a bit (330 bar Raptor). But even slightly more conservative approaches offer a great SLS replacement.
Last question first: Increasing chamber pressure (MCCp) is generally good and should be only limited to physical restrains. At some point, plumbing might get too heavy but we're faaar away from that point. Higher MCCp allows higher expansion ratio (ER) with same sized bell. 2/
I did expect it to fall apart any second seeing the tiles flying around. But it didn't. I'm in awe.
3. Even through the loss of tiles, landing propellant in the header tanks did not boil-off and supplied the still working engines to stand still.
(2/4)
4. 2 engines failing was no problem for Booster, it delivered within margin and sticked an impressive soft landing.
5. No loss of communication during re-entry due to plasma is something new for me.
6. Booster is faster than hot staging ring. Surprise...^^
(3/4)
Or you could keep the ER and increase thrust. That's what SX is currently doing. Because there's limited space underneath Booster. Widening it would require changes for OLM and construction. Not sure about optimal compromise between thrust and Isp. 3/
Turbine blades only survive ~1000 K. In gas power plants, they increase that value with air film cooling for the blades (crazy engineering!). Doesn't seem to work for rocket engines - if at all only with a fuel rich stream, if oxygen gets too hot it'll eat the plumbing. 5/
Some time ago, I made this study. Showing the ridiculous result of increasing thrust but not enlarging the bell due to spacing issues. You leave a lot of valuable vacuum Isp on the road. But, due to construction tooling and OLM, widening Ship & Booster would be expensive.
(2/9)
Currently, they're likely working with a lot of film cooling. Decreasing MCCp and with it reducing film cooling would probably increase Isp higher up, without changing ER. BUT there's another strong limit. That's strongly depends on the heat load the turbopumps can bare. 4/
HLS is supposed
-to taxi between surface and NRHO,
-to do refilling,
-to deal with boil-off over months, -to house crew for weeks...
... but TLI (Trans Lunar Injection) and EOI (Earth Orbit Insertion) are the unsolvable problems?
I lowered the oxidizer stream temperature by 200 K, to have more realistic values. Yet, assuming the pressure drop estimates taken from RS-25 are right, the CEARUN data is valid, and I did not make any mistakes, Raptor still has room to increase MCCp far beyond 350 bar! 17/17
Hot take: rocket science isn't the hardest thing. It can be done by a lot of start-ups with relatively small investments.
Try developing a new chip architecture, cancer drugs, a new smart phone - or, apparently, Gen >III nuclear power...
The primary advantage of a superheavy lift RTLS architecture is a high right rate when compared to a ASDS only vehicle not capable of RTLS due to its hydrogen second stage like New Glenn.
As such, they propose that the flight rate of their competitor be capped:
The biggest update for me is in the numbers. A single BE-3U will have 1.1 MN of thrust. That's J-2 level and more than double than BE-3. Finally a proper thrust for the upper stage. Now, BE-4 needs the most likely possible thrust increase and New Glenn can do much more than 45 t.
I'll provide the data soon if someone is interested. We cut and smoothed it a little here and there. Will explain then. Thanks to
@schiffer_soft
for involving me here!
@SpaceX
you're watched every step on your interesting path! :)
So, with temperature for different O/Fs - and therefore possible work done by burning what's possible within the fluid streams, to enable Full Flow Staged Combustion - I tried to figure out the efficiency of the turbine. Here, F-1, Vulcain 2 and RS-25 data helped a bit. 8/
It seems, still a lot of people underestimate the potential of a high&cheap mass to orbit environment. While
@SpaceX
always claims to be heading for Mars, Mars isn't necessary. It could be just the carrot on the stick. (1/4)
if starship isn’t built to go to mars I wonder why they’re building a massive factory meant to scale to hundreds of ships and boosters per year, attempting to build like 5 pads and actively damaging some commercial prospects in order to accomplish that goal
So I'd be really happy to see them get there with future wider Starship versions. For now, the optimization lies somewhere else: A working heat shield!
But in maybe 5 - 10 years? Wider rocket, maybe?
(9/9)
Limits seem to be somewhere around 850 K for the oxidizer stream. Cap there, means a lot of oxidizer with just a bit of fuel which can't do much work. with CH4 & O2 it's an O/F of ~50 (stoichiometric = 4), which with H2 & O2 it's ~115 (stoichiometric = 8). 6/
The old ITS idea though was 13 m wide, not 9 m. So large bells would've been much less of a problem. But, at least for the SL engine, you don't want to go too large to avoid too low exit pressure, especially when throttling, which reduces SL Isp and might damage the bell.
(3/9)
@MrB3ard
@torybruno
There's a huge difference between shooting something below 350 km perigee or at ~440 km perigee. Anything below 350 km drops back to Earth in a few months. At 440 km it takes decades. And the stupid Chinese test was at ~850 km! That's even worse and was criticised!
Scoop – One of Blue Origin’s BE-4 rocket engines exploded during a test firing in Texas on June 30, according to CNBC sources.
The engine was to be delivered this month to ULA for Vulcan’s Cert-2 launch. More:
Fully reusable rocket architectures under development:
1. First stage return with main engine + side passive shield re-entry/main engine landing
2. First stage return with main engine + bottom active shield re-entry/main engine landing
That's all? No horizontals?
So you're limited to ~0.6 bar exit pressure. Looking at the old study though, then, only 357 s Isp are possible. But I applied the overall engine efficiency of 94.5% here. That's a rough estimate based on the famous 327 s number provided by
@Erdayastronaut
.
(4/9)
By then, if assumed 96.3% overall engine efficiency, the 361 s for 0.6 bar exit pressure of a SL Raptor version and 382 s for 200 ER for RVac could be achieved. Would the larger bell diameters fit then?
Yes! Have look at my amazing Excel shape skills!
(8/9)
And here, I did a lot of fiddle work with CEARUN to generate lists of O/Fs, pressures and temperatures for kerolox, methalox & hydrolox. Feel free to ask. Data problems: kerolox seems way too hot for very low O/F and in some regions hydrolox temp doesn't change with pressure. 7/
My guts place CH4 between RP1 (94.2% for RD-180) and H2 (97.5% for RD-0120) at >96 %. Why? It's still quite a small molecule. It likely splits faster than the long and sluggish RP1. The reason for low efficiency, and
@elonmusk
said it himself, is the generous film cooling.
(5/9)
With a bit of RPA (based on CEARUN), I wiggled in efficiency values of ~8.4%. This means, I burn a specific amount of fuel to generate heat. This heat does work on the turbine, limited by low pressure ratio, turning ~17% of the heat energy into kinetic energy, ideally. 9/
Looking at the fuel side, with 1000 K, impressive pre-burner pressures are possible. ~1380 bar for methalox and ~860 bar for hydrolox. While H2 has great heat capacity to enable lots of energy release in a cool flame, it also needs huge pumps because of its low density. 11/
We cut the data slightly below the start of
@SpaceX
's sim due to the fact that the acceleration data doesn't make sense for the first second. Once caught, we smoothed the ragdolling. Here you'll find the data:
Does that work?
@dorfman_p
No, it won't. High radioactivity is a result of a short half-life. In the next several hundred years, it'll fall on a level that's easily comparable to a lot of highly toxic stuff that's found all over the planet.
This only works with ~70% efficiency. The kinetic output then is transformed, with another ~70% pump efficiency, into fluid pressure. With this, I have the maximum pressure I can put into a fluid stream at a given temperature. 10/
The specific densities therefore look like this. Yeah, that's the problem with hydrogen. Not only does it need to be crazily cold to be liquid (~20 K) which is already quite a challenge but it also is light which needs large tanks AND the pumps care only for volume. 12/
So between hydrolox and methalox, there's quite a difference on the pressure potential. With the oxidizer side, it looks different. Though, here we have the temperature problem more pressing. Nonetheless, at 850 K we still get impressive values around 900 bar! 13/
Remember this from the 2018 manual? Here, I concluded that the first stage might be capable of holding ~1.2 kt prop.
@_mgde_
's picture supports this. Total rocket mass at reasonable values then is ~1.56 kt -> At about 1.71 ktf thrust, that's a TWR of 1.1. I don't believe it...
***B R E A K I N G***
Potential New Glenn first stage flight hardware spotted outside at
@blueorigin
’s campus near the Kennedy Space Center.
This looks to be the LNG/LOx tank section of the first stage - now sporting a much anticipated livery.
📸 -
@NASASpaceflight
Assuming single shaft/geared turbines to distribute power optimally, and have the cooling done by both streams (which is possible,
@launcher
does that with E-2, so will
@blueorigin
with BE-7), then we have those maximum possible values. 16/
So I'm still in favor to widen
@spacex
's Starship (10 m) & Booster (11 m). Just more space for larger nozzles. And I tuned down MCC pressure to reduce film cooling (big assumptions on how this actually affect things - since
@elonmusk
STILL DOESN'T TALK ABOUT ISP! GRRR!!). 1/2
Trajectory modelers unite! ~203 t with the infos I have onto a 200x200ish km orbit (insertion at ~130 km with circ.), no re-entry burn. Lots of assumptions. Not considered e.g.: Thrust reduction to decrease film cooling and increase Isp, phase out etc. Thoughts, your results?
And that's only the main fluid stream. The one's feeding into the pre-burner chamber need even more. But to simplify things, I'm not going into detail here. Applying those RS-25 values, this gives us the minimum MCCp possible. Here, the weaker stream limits things. 15/
Those are the pre-burner pressures. But there's a drop by going throw the pipes, even partially passing the chamber and nozzle cooling system, before entering the main combustion chamber. For RS-25 with only the fuel cooling, it's factor 2 on the fuel side and 1.4 for O2. 14/
On Friday, I posted some thoughts on the pressure potential for Raptor (). For the analysis, I used CEARUN data (). The combustion temp spectrum seen here goes from 1 to 1,200 bar. I'm not confident that CEARUN is correct. 1/4
FFSC/Raptor thread: Recently, Musk reported that
@SpaceX
achieved 350 bar chamber pressure for Raptor 3. Impressive. Though, is there a limit to the maximum pressure? Or will this go up indefinitely? Does it make sense to further increase that value? 1/
Trajectory modelers unite! ~203 t with the infos I have onto a 200x200ish km orbit (insertion at ~130 km with circ.), no re-entry burn. Lots of assumptions. Not considered e.g.: Thrust reduction to decrease film cooling and increase Isp, phase out etc. Thoughts, your results?
That's my current view on the Raptor versions. I'm baffled by the numbers and likely wrong. So any info regarding the current sizing (nozzle exit diameter) would be helpful. Efficiency and equilibrium numbers seem through the roof.
Well, if Boca Chica grain silo, with the enlarged 9-engine upper stage, doesn't care for max q restrictions (including not piercing through it but hitting it with significant AoA), even 229 t might be possible... Dude... Still given, my SL Isp and dry mass assumptions are right.
Btw. Though New Glenn officially can carry 45 t to LEO, I consider this a strong understatement. Given the size of that rocket + reducing the VERY generous thrust margin on BE-4, 50% more are likely possible, so having Blue HLS fully filled in LEO should be easily possible. 10/10
Not much new besides that they gotta check with the team if there are any show-stoppers for Booster catch and that Ship landed only 6 km away from planned.
But Elon, I said widen the BOOSTER not widen yourself!
NEWS: Elon Musk gives a thorough update after the 4th launch of Starship.
Will we see a catch for Flight 5?
How will Starlink be even better on the next flight?
These are some of the questions Elon answers for us.
I hope this interview informs you about the incredible work SpaceX
Less pressure in the chamber would mean less dense hot gas and a larger chamber (when keeping the overall thrust), so the surface to volume ratio would improve. Film cooling could be reduced. And even the above mentioned engines have some.
(6/9)
When SpaceX achieves 10 ktf thrust for Booster, the supplied O2 alone would burn ~6 t/s CH4. That's 330 GW UHV heating. Burned in CCGT plants, that's comparable to 38% of the total US gas power plant capacity. The 6.8 t/s exhausted CH4 equals 26% of US natural gas production.
The year 2063: With Bhutan, the 37th country put people on Mars. The University of Applied Science Reutlingen installed its own space station around Moon with the help of Easy Space GbR. Boeing found an issue with Starliner's plastic covered RCS nozzles: Pushes test date to 2064.
Just to remind you: It's utterly embarrassing that in over 50 years (!) nothing has been achieved regarding efficiency and cost (especially given SLS cost numbers are optimistic)!
What's wrong with rocketry? They had slide rulers. Today there's CAD + high performance simulations.
But you just need it for 1% of the prop that you vaporize for re-pressurization. So that energy is basically nothing and you take it away anyway, regardless of pre-burner or pure gas. So it all comes down to avoid an extra heat exchanger and piping? I don't get it.
(3/3)
True, fully reusable, Starship (coming versions) might only get ~8 t to GTO (~30 when expending a standard Ship, ~90 for a stripped one).
It gets ~200 t to LEO, though. With an expendable 3rd stage then, capabilities are looking roughly like this (bodies = transfer only): (1/2)
What you see on that image is BE-7, the (apparently) Dual Expander Cycle engine, Blue is developing for Moon.
@JeffBezos
said, it'll have 453 s Isp. Realistic? Yes. Applying a measure, throat vs nozzle exit indicate ER at ~120. 453 s at 5.8 O/F indicates a 96.3% efficiency. 2/
Yes, yes, I discussed it enough that widening the stack is harder than other changes, still, here, you get an idea, how much Isp Raptor leaves behind because there's not enough space for a larger bell. ER is at about half what it should be. So more than 10 s wasted.
If
@SpaceX
achieves 150 t to LEO, 16 flights needed to get 150 t to Lunar surface and return a standard Starship to Earth. That's reasonable. And doesn't need mods to Starship beside cooling. ~8 for expendable. so once 8 flights are cheaper than an expendable lander...