You are not logged in.
Not knowing to put this, I put it here:
from AIAA "Daily Launch" email newsletter for 28 Oct 2026:
Aerospace Global News
SpaceX fires 33 super heavy engines ahead of historic dual-catch starship test
SpaceX completed a full-duration static fire of its Super Heavy booster at Starbase. Flight 15 aims to catch both the booster and the upper stage for the first time.
SpaceNews
Blue Origin still pushing for return to flight of New Glenn this year
Blue Origin expects to have its rebuilt New Glenn launch pad ready by the first half of December to support a return to flight of the rocket this year.
IEEE Spectrum
Will Fiber-optic Cables Finally Get Made in Orbit? Maybe
In a lab in Luxembourg in 2024, engineers at Flawless Photonics unpacked a spool of glass fiber that had just returned from the International Space Station (ISS). As they threaded it onto their rewinding machine, the machine jammed—as happens sometimes—and they braced for the usual outcome: The fiber would snap. This time, though, nothing broke.
from the AIAA "Daily Launch" email newsletter for 20 Oct 2026:
Ars Technica
SpaceX no longer needs NASA, and that’s a problem for America’s space agency
SpaceX went from needing the US government as an indispensable customer to becoming an indispensable contractor for the nation. The pendulum has swung so far now that SpaceX seems increasingly disinterested in the US government for most contracts outside of data services, such as Starshield for military communications and observations.
Defense One
The US space industry will suffer for Ukraine’s Starlink outage, experts say
Elon Musk’s decision to hobble satellite services to Ukraine’s military, and his shifting explanations for doing so, have undermined the U.S. space industry and the country’s diplomatic and intelligence partnerships, say lawmakers, foreign officials, and former senior members of the U.S. military.
GW
Bob:
The image with data on it that you posted above has takeoff thrust listed as 8240 mt-f in the vehicle summary section, contradicting what is in the performance block just above (which lists 10,000 mt-f).
8240 mt-f / 33 engines = ~ 250 mt-f per engine, which is more or less in line with what I thought Raptor-2 max thrust performance might prove to be (a couple of years ago), with the higher-capacity turbopump design than the Raptor-1, which fell woefully short. This contradicts what Musk claimed in his social media post. I do not know where you got the image, and social media is utterly notorious for lies. I think you are reasoning with disinformation and misinformation, much of it very likely deliberate.
And you keep ignoring my main point: for a vehicle still deep in developmental flight test, there can be NO guaranteed payload capacity! There are ONLY projections of what it might be! And the test outcomes will change them. Maybe a little, maybe a lot. We will see.
My own opinion is this: once they get the unreliability issues with the Raptor-3 corrected, the basic cargo vehicle they are flying in tests will likely prove to have a payload capacity fairly close to those projected. But if they cannot get the engines reliable, there is no point! No vehicle. No capacity to worry about! And THAT is where they are right now! And in 3 flights with the same relight troubles, they clearly have not figured out what is wrong, because they so evidently have not yet corrected it.
Bear in mind, there will be development flight tests required of every alternate configuration: the HLS, the tankers, the cargo ship with big clamshell doors, ALL OF THOSE! And every one of them will end up with a slightly-different payload capacity.
GW
The payload capacity of Starship/Superheavy v.3 CANNOT be known with certainty (!!!) until it finishes flight test development and is well into its final flight tests doing pretty-much-certification testing! (And there must be several of those, before it can be "man-rated".) There are ONLY projections of what that number might turn out to be.
No one in their right mind would try flying in flight test development at max payload, so saying that its capacity will only turn out to be what it has been carrying (near 50 tons, so far) while in flight test development (and it still is in that phase!!!) is NONSENSE! The projections (nearer 150 tons., maybe more) are "more accurate", as if anything can be.
The propellant capacity has grown to something nearer 1600 tons in v.3 according to the image you posted, although I do know know it to be factual in its data, even if you got it from SpaceX, and there is no way to ensure that it came from them in a world overrun with AI slop.
It will arrive in orbit at max payload, whatever that is, with enough propellant still aboard to deorbit and land, say 100 tons as a wild guess. It would be utterly insane to orbit one with no propellant aboard by which to deorbit and land. That means it might take 1500 tons to refill it fully for some demanding mission outside LEO. If the payload capacity measures deliverable propellant on orbit for a tanker flight (and it might well not), then a mere 50 ton payload would require 1500/50 = 30 flights for a full refill. See how ridiculous that is?
On the other hand, if payload capacity turns out to be 150 tons, that's 1500/150 = 10 tanker flights to refill. That's far better, but still awful high. If it turns out to be nearer the 200 tons in your image for v.3, that's 1500/200 = 7.5, call it 8 flights. Pretty close to what Musk says. To get the 6 in his 6-8 flight estimate, 1500/6 = 250 tons payload per flight. Which may be where the "200+" came from in your image for v.3. The 8 corresponds to 1500/8 = 187.5 tons per flight.
So that's SpaceX's projection range, at this stage of development flight testing: somewhere between 187 and 250 tons to low-inclination LEO. They do not yet know it any better than that! And there are still some serious engine reliability problems to solve with Raptor-3 before development flight testing is over and they can actually "certify" how much it might reliably carry with the final test flights. Even that might not be an actual demonstration flight at that max payload, although it might (or might not) be close.
But as I said, it is all speculation until the flight test development effort is over for v.3. And remember, development flight testing begins all over again, if and when they start working on v.4!
As for the credibility of the NASA Artemis moon landing schedule, there never was any credibility to that. They started projecting landing dates BEFORE they ever hired any vendors for the lander, based on the wildly-overoptimistic timelines put into proposals to win contracts. The NASA of the early 1960's knew better than that! You multiply those by factors of 2 to 4 before you ever commit to any planning dates! It would seem that the people who knew better inside NASA are long gone now, dead or retired.
You also have to consider the engine development history, as a thrust shortfall will reduce launch thrust/weight, perhaps doubling the gravity loss during 1st stage flight, and drastically lowering your payload capacity estimate. Raptor-1 indeed produced a severe thrust shortfall, never being able to reach the intended full chamber pressure. Not a structural thing, a turbopump delivery shortfall!
They "fixed" that with the Raptor-2 design, but not all at once. It took some flights before they could get up near the intended thrust levels. After a while, it was working both reliably, and at pretty near its intended max thrust. That's when they announced they finally got it working at the full intended chamber pressure, remember? There was a public announcement of that feat.
Now they went to a cleaned-up revised version called Raptor-3, and the changes have cost them the reliability they had in Raptor-2, although they were expecting a bit higher chamber pressure and thrust. They got the thrust, but not what they wanted for reliability. And they do not yet understand what causes the unreliability, or they would not have experienced the same basic relight problems on flights 12, 13, and 14, in succession! There is a vast difference between the environment in a flying engine bay, and the environment around a test stand on the ground. Plus, there are very likely multiple causes to root out, not just one!
They are still nowhere near the end of true development flight testing! Despite any hype in the news releases and Musk claims on social media. "And that's a fact, Jack"!
GW
By the way, all this stuff that I wrote in the previous postings is in the "orbits+" course materials, available for free download from a drop box, using the links in the postings of the "orbital mechanics class traditional" thread, here on the forums. Those materials include various self-teaching lessons, multiple spreadsheet tools, and the user manuals for them. It starts with 2-body ellipse mechanics, but goes beyond that into dV estimation, 3rd-body speed corrections, engine performance characterization (real engines deliver less Isp than those handbook values, and for a variety of reasons), vehicle sizing, and even entry, descent, and landing.
GW
For launch from Earth, if the vehicle is relatively slender and "clean", you can add about 5% of the required ideal launch dV to cover drag losses. Almost all of those go on the first stage of a two-stage vehicle, because staging occurs in very thin air about 35-45 km up. Slender mans L/D about 6 min to 12 max, and clean means no big step changes in cross section, only small ones that are well faired-in. If you don't meet those criteria, double the drag loss to 10%.
If the vehicle has adequate launch thrust/weight ratio, so that it does not "dawdle around" getting to max q at about Mach 1, 20,000 feet, then the gravity loss you must cover is around 5% of the ideal launch speed dV. That is a thrust to weight ratio of at least 1.5, with 2 preferred. If you are closer to only 1.2-1.3, then double the gravity loss to about 10%. About 90% of that goes against the 1st stg of a 2-stage vehicle, because it it flying almost but not quite level, at staging.
The ideal launch dV is best obtained from a surfacing-grazing transfer ellipse using 2-body orbital mechanics. Put the apogee at the final desired circular orbit altitude. The ideal perigee velocity of that orbit is your ideal launch dV, to which you need to add the circularization burn dV, which for LEO is around 100 m/s. The factored-up ideal launch dV plus the circularization dV is the mass ratio-effective dV you must design into your launch vehicle.
GW
The drag equation used from incompressible to hypersonic is D = CD q A, where A is the reference area used for CD. It is q that beyond about Mach 0.5 to 0.8 can no longer be measured accurately with a pitot-static tube. It is simply calculated as 0.5 density velocity-squared, even hypersonic. The drag coefficient has a constant subsonic value up to around Mach 0.8, then there is a transonic drag rise to a value about twice the subsonic value, peaking in the Mach 1 to 1.1 range. On the supersonic side there is a decrease to values comparable to (and sometimes lower than) the subsonic value, by around Mach 3 or 4. From about Mach 5 on up, the supersonic/hypersonic drag coefficient value is almost constant (the shape of the shock field and flow pattern around the object no longer changes significantly with speed). For really blunt objects, that constancy of hypersonic patterns and drag coefficients starts nearer Mach 3 than Mach 5.
The lift force from a blunt object like a capsule is quite real, but not large at all compared to the drag, quite unlike airplanes. The capsule flies at an angle of attack not large enough to cause its lateral wall to become parallel to the flow on the one side, because of flow reattachment and much higher heating rates on that one side. During entry with most capsule shapes, lift/drag is only near 0.1 at most. You roll the angled capsule about the oncoming velocity vector axis to point that lift vector in the direction you want, to fine-control your entry path. Once past peak gees, you are slow enough that dynamic pressure is low, so that drag and lift are low, and this type of control no longer works. The trajectory inherently droops downward from there, and really droops once you are only supersonic again.
You have to be very careful with extended surfaces of any kind in hypersonic flight through the atmosphere, even up high in the thin air. There are all sorts of different geometries, and those that shed a bow shock wave which impinges upon an adjacent portion, like two parallel nacelles, one upon the other, will incur something called "shock-impingement heating". When that occurs, the normal supersonic convection heating rates get multiplied by a factor in the neighborhood of 7 or higher, locally, right in the impingement zone. Not the driving temperatures, just the heating rates. The material so affected will come to full thermal equilibrium in a second or so, to something pretty near the flow stagnation temperature, which above Mach 5 to 6 is usually higher than any known materials can survive, especially the metals.
That last occurred on the X-15 flight that reached 4520 mph at about 100,000 feet, the Mach 6.67 flight Pete Knight made in the X-15A-2. That flight had a scramjet test article mounted on the ventral fin stub, instead of the usual ventral fin. The scramjet spike shock impinged upon the ventral fin stub and on the underside of the tail section, nearly destroying the stub, and it came close to cutting the tail section off the airplane. This effect is quite real, and quite dangerous! And it is exactly why no hypersonic vehicle can ever have a shape resembling the SR-71. The shocks off the engine inlet spikes would cut the wings off.
Entry trajectories are at a shallow angle to the local horizontal, in order to be survivable at all. But once the trajectory starts to droop after peak gees, it rapidly converts to a steep downward angle. By the time the object is "only" supersonic, it is pretty steeply downward, and once subsonic, very nearly vertical. As a rule of thumb for crude estimates, I use about 45 degrees average, once the hypersonics are over at about Mach 3 for blunt objects. End of hypersonics at Earth typically occurs about 35-45 km altitude, and is relatively insensitive to ballistic coefficient. There is plenty of time to use chutes or something, to get very subsonic (typically 15-20 mph) by the time you near the surface.
Mars is different because of its very thin "air". End of hypersonics altitude is very sensitive to ballistic coefficient, and it is also much lower down! It is near only 25 km at 50-100 kg/m2, and nearer only 5-10 km at 400-600 kg/m2. You are only mere seconds away from surface impact at your Mach-ish speed (on Mars, about 700 m/s). From 25 km, there is just barely time to deploy a ringsail chute at Mach 2.5, but your surface terminal velocity is just barely subsonic, if not still transonic. Rocket braking to touch down survivably is simply required.
GW
Well, my guess for the wrap-around angle was just empirical, based on what I have seen. You can always run a spread of angles and see how much they affect the answers. As near as I can tell, it's not the strongest effect, once you are anywhere in the ballpark. It is the heating rates, pressures, and refractory re-radiation temperatures that help you select the right kinds of materials for your heat protection. Ablatives depend upon heating rate and pressure. Refractories depend upon servce temperature and pressure.
My aerocapture spreadsheet tool is for crude estimates, to be made by people who do not have (or have access to) an trajectory program that includes both thrusts and atmospheric drag, at orbital-class speeds. Most people do not have that. I do not have such.
I come from the slide rule days, when almost everything was done by hand, pencil-and-paper, running numbers with slide rules (later, scientific calculators). Only those things absolutely too complex to do that way, were run on the main frame computer, which for most of my defense career was an enormous thing in a room cooled with an enormous air conditioner, fed batch input by punch cards in steel trays, and with about a 3-hour to 3-day turnaround time for each run, the majority of which bombed out. All the while with management yelling at you, that you were delaying payroll! There were no desktops or laptops in those days! There were no word processors, no spreadsheet software, and no powerpoint to make slides with.
That is when I learned how to do those things that I can do. I did do some scientific programming to automate some of the worst stuff, but the only languages I knew are long dead now. Things like Fortran IV (NOT Fortran 77 or newer; I learned on Fortran II in college, over half a century ago), and some of the advanced BASIC's like QuickBASIC 4.5. I still have access to QuickBASIC 4.5 on my laptop here, but it takes a couple of gamer shells to run it, and to print out the results. It was a Microsoft product that only ran under DOS operating systems, not Windows! The gamer shells get around that. My ramjet cycle analysis codes are written in it, about 3000 lines each.
I was forced out of defense work after 20 years, by the huge drawdown that took place after the fall of the Soviet Union. I went into a second career of mostly teaching, plus some civil engineering work. I taught math, physics, and engineering at levels from high school to university. The civil engineering included both fire protection and environmental. That went on for another 20 years. I have been fully retired for 11 years now.
This last year I did some paid consulting in ramjet propulsion work. But that's about all, other than about 20 years custom-building farm implements that kill prickly pear cactus out of farm and ranch pastures, with no pickup and disposal, and no chemicals. I invented that totally by accident, but it worked, and I have built and sold over 60 of them, in my shop out here on the farm. About half a ton each.
GW
I'm not sure, but I think we are saying roughly the same things, just using different words.
GW
Dr. Muller
I do have a crude spreadsheet tool for estimating aerobraking. It is in the tools topic as a free download.
It uses a best guess for how far around the planet one must go while aerobraking, along with how much velocity you must lose to drag, to get an average speed around that path (which is also your speed at periapsis), and therefore a max deceleration gee at the periapsis of that path.
If you know your ballistic coefficient, you can use that plus the max deceleration gee at periapsis to estimate the max drag force you must have. If you know the drag coefficient, you can use that with your periapsis speed estimate, to find the density of the atmosphere that must exist at your periapsis point. That density, used with a simple exponential scale height model of density versus altitude, tells you just how deep into the atmosphere your periapsis must be.
During entry, peak heating always occurs prior to peak deceleration gees and at a significantly higher speed than the speed at peak gees, but not a large difference in altitude (because the angle must be shallow to be survivable at all). So as a rule-of-thumb estimate, I use the average of speed at entry and speed at periapsis as the peak heating speed. I put that and the periapsis density into the rough correlation equations for convective and plasma radiation heating rates, to estimate stagnation heating rates for those two effects.
I put all this into the aerobraking spreadsheet tool. It actually works pretty good. I even have some rules of thumb in it for spreading those stagnation heating rates around to other locations as realistically-reduced values. Given some surface emissivities, the tool even estimates refractory surface radiation temperatures at those sam locations.
You need only know what planet (for the atmosphere model), what speed at entry interface, how much speed you need to lose in order to capture, your vehicle's ballistic coefficient, and the hypersonic drag coefficient that went into that ballistic coefficient. Make the best guess you can, for how far around the plant your aerobraking path will go, in the atmosphere. A quarter turn (about 1.5 radian) is a good starting guess for aerobraking aerocapture.
GW
Maybe my version of the low thrust problem will make more common sense to everyone.
A rocket at some radius R from a central body can thrust in some direction at the end of that radius. Depending upon the direction of that thrusting relative to the position of the vector R, this can be more or less efficient at increasing the kinetic energy of the rocket (measured as its speed). 90 degrees to the radius vector is most efficient, parallel to it outward is least efficient. This is why ....
... The work done by the thrust force upon the rocket can go into either or both of two pots of energy: kinetic or potential. What goes into potential energy increase does NOT go into kinetic energy increase, meaning no added speed for whatever went into potential energy increase. That component of thrust along the radius vector R is what raises potential energy, working as it must against the gravity of the central body. Only that component of thrust perpendicular to the radius vector R can go into kinetic energy increase!
If you are at the periapsis of an orbit and fire an engine in the prograde direction, then at least initially, your thrust vector is perpendicular to that radius vector R, and all the work of your thrust upon the rocket can go into raising its kinetic energy, at least initially. If your burn time is short, that perpendicularity of thrust & R does not have time to change, and essentially all the work done during the burn goes into increasing kinetic energy, which means you actually get the full speed increase from your burn.
If on the other hand, you have a low thrust engine that barely accelerates the vehicle at all, then over the course of a longer burn time, the vehicle has time to move away from the periapsis point. Everywhere away from there, thrust is no longer perpendicular to the radius vector R, meaning some of the thrust work goes into raising potential energy instead of increasing speed! The lower the vehicle acceleration, the worse this effect becomes! As a rule of thumb, you need around 0.1, maybe 0.2, gee acceleration, in order not to suffer this loss in any measurable way, in typical Earth orbits.
This loss can be suffered even with "high thrust" chemical rockets, if your sized engine thrust produces too small an acceleration, leading to a long required burn time. It is just catastrophic with current electric propulsion technologies, because vehicle accelerations typically fall in the milli-gee to micro-gee range! Satellites maneuvering on orbit with electric propulsion can afford the loss of efficiency ONLY because the specific impulse is high, so the mass of propellant required for a relatively small low-thrust burn is still low enough to tolerate.
As another rule of thumb, I use the dV calculated for high acceleration "impulsive" rockets, multiplied by at least a factor of 2 (!), as the low-efficiency "mass ratio effective" dV to use, when sizing electric propulsion systems. Some say use only 1.5, but I would never bet lives on a factor that low!
GW
PS beware of 3rd-body corrections for your dV's based on the formulas for 2-body orbital mechanics. Third-body effects are quite real, and quite large.
Those circles are around where the 4 hold-down pads were located on the heat shield. Apollo had them too. These cause local waves, that affect the local flow pattern approaching them, and where the wave intersects the heat shield surface, enhanced heating rates locally.
That last enhanced heating effect is much larger with actual shock waves, which were seen to produce damage on the X-15 flight that reached Mach 6.7 at about 100,000 feet. Here, it's likely still a subsonic effect, because I think the flow re-acceleration sonic line is near the rim of the heat shield. In the photo, the stagnation point is off-center, to the lower right, more right than down. The directions of the streaks on the heat shield point to it.
GW
Spacenut:
Much of the risk is mitigated if you just do spin gravity. Most of the rest is mitigated if you do solar flare radiation shielding adequate for large flares like the 1972 event. That flare shielding cuts even the worst-case cosmic ray radiation down to levels considered "safe enough".
You can do those things in an orbit-to-orbit transport, but generally not in something that can also do entry, descent, and landing. The design requirements conflict too much!
You would need to do nuclear pulse propulsion to get around that problem of conflicting design requirements imposed upon one do-it-all design, and that would cause a whole host of other problems.
GW
Welcome to the forums, HG Muller!
I think you have encountered Tahanson43206, and there are a few others of us on these forums. I am an old retired defense industry development engineer, and a retired teacher. I got started in the slide rule days, and was interested in spaceflight long bore anybody went up there. Never actually worked for NASA, though. When I was young, the X-15 replaced the X-1 and the D-558-2, just about the same time as NASA was founded from the old NACA. I well remember the hoopla when Sputnik first went up.
Most of my industrial work was in air-launched missiles and the countermeasures that defeat them. But you might remember the LTV "Scout" satellite launcher. It was a 4-stage solid that could put small (by today's standards) satellites into orbit. I got to work on advanced configurations for it at LTV one summer as a graduate student. About that same time, I learned the old Allen & Eggers simplified by-hand analysis for entry dynamics and heating. Did my MS in a hypersonic wind tunnel working on Shuttle nose shapes before that design "gelled". What we grad students found, about the AOA limits on shuttle entry, was later verified at the AEDC facility in Tennessee.
Anyhow, welcome!
GW
Oldfart1939:
You are quite right!
There's both the crash of hardware cratering in, and the crash of budgets cratering in.
GW
Oldfart:
I agree, spaceflight is hard and it is inherently dangerous. Doing new things will always require test pilots and involve high risks, same as with high-speed airplane flight, but worse.
I agree that the FAA is indeed the stumbling block, but it is a necessary one! Musk has proven himself willing to take too many risks with other people's lives, by always trying to push ahead faster than is prudent. That showed up in his repeated violations of his launch license conditions, earlier in the Starship flight tests. Remember how he fumed at the delays in granting the next license? It is the FAA's restraint that keeps the casualty rate down. As it does with new airplane manufacturers, as well.
However, for there ever to be commercial space travel of someone other than test pilots, one must address all the risks and mitigate them to the max extent possible, same as was done with high speed airline travel! It's a big row to hoe! But NO ONE will ever tolerate in commercial space travel of civilians the same high casualty rate that one sees among test pilots in experimental flight test!
Somehow or another, Tom read something into what I said in post 2412 above that I did not say! JFK set a reasonable and realistic goal for the moon landing, despite there being neither space craft nor launch rocket adequate to the task when he set it. He was able to do that because he had advice from people who actually understood and had direct experience with how long it might really take to do such things! Most of them were high up in the defense contractors, relying on how long it took to build new jet airplanes, and how long it took to make a new ballistic missile reliable. No "magic" there! Just lots of experiences to draw upon.
I don't think there ever were any people at NASA who understood such issues, or if there were, they were never in positions of decision-making at that level. No one at NASA set the moon landing goal, JFK and his advisers and industry contacts did!
Today's NASA probably still suffers from the same lack, hampered further, in that goals and programs are set by a Congress totally incompetent to be setting them! Not to mention the fact that unlike "legacy old space" contractors, the "new space" contractors involved with the Artemis landers have no such long track record of experience at developing new spacecraft and rockets.
The closest thing to such a track record would be SpaceX, but there are ONLY 2 such examples to look at! How long was it from constructing the first Falcon-1 until Falcon-9 could launch reliably as an expendable? Something close to 8 years! How long was it from the first flying prototype (Grasshopper) to a Falcon-9 booster core recovery that seems routinely reliable? Also something like 8 years.
Why would anyone actually knowledgeable, ever expect it to be anything less than 8 years for Starship to go from its first prototype to flying reliably? And given how much more demanding a reentry-vehicle second stage (something never done before, this is NOT a space plane like the shuttle) really is, the timeline is probably longer than 8 years! SpaceX would seem to be "about on track", reaching orbit in experimental flight test but still having serious problems. It has NOT been 8-10 years yet! Flight 1 of ship+booster was in 2023. The suborbital flights occupied about 1-2 years before that. They're only about halfway through the process!
Something near 2021/2022 getting started, plus 8-10 years is 2029-2032, before the baseline Starship is safe to ride by anyone but test pilots! The HLS is later than that, maybe mid 2030's! Blue Origin with its bigger manned Blue Moon might beat that schedule, but not by much, if at all! They only started the little unmanned Blue moon a couple of years ago. 2024+(8 to 10) = 2032-2034! Where does THAT put NASA with its Artemis landing?
And SpaceX still has to experimentally test and verify any significant variants, such as tanker versions, cargo versions with big clamshell doors, or the HLS for Artemis! That variant verification problem adds yet more time (not necessarily 8 years once the baseline ship and booster are "right", maybe closer to 2 or 3), and it CANNOT be done in parallel with the development and verification of the baseline vehicle!
And if there is a fatal accident (test pilot or civilians on the ground) anywhere in any of these efforts, they will take longer! It is more than just wise to proceed prudently, and not "crash" the program! Crash programs have higher casualty rates! They always have! They always will!
GW
Oldfart:
The difference between JFK committing us in 1962 to a landing before 1970, and what NASA did with its Artemis schedule, is proper background data plus good advice. JFK had some knowledgeable advisors on his staff, plus informal access to the heads of the industrial contractors who would be doing the work. Plus, there were actual experiences with new aircraft developments from about 1940 to 1962, and with ballistic missiles from about 1950 to 1962.
That set a background for what kind of timelines might be reasonable and realistic, presumably resulting in a best-estimated landing date about 6 to 12 years away. That range bracketed 1 Jan 1970, and JFK was good at public speaking, so he picked that date as one everybody in the country could easily understand. It was reasonable and realistic, plus it was challenging, which his speech implied he wanted it to be. It was not NASA that made that decision, it was JFK.
And yes, NASA was taking higher risks with crews than anyone commercial today would ever be allowed to do. We ended up beating the deadline by less than half a year, despite the year-long debacle after the Apollo-1 fire with its dead crew. We did it precisely by taking those risks. Very nearly losing some more crews (Apollo-13 and the Gemini that went out of control that Armstrong piloted, both come to mind. As does John Glenn's Mercury, and maybe Gordon Cooper's Mercury).
NASA working with "new space" contractors did not have that background history of how long it would take for those companies to actually develop new flight vehicles, although there was the 8 years demonstrated by SpaceX getting from Falcon-1 to Falcon-9, and the next 8 years getting to Falcon-9 stage recoveries. Government development contracting has been long infamous for hyped schedules, and it's worse today than it was then for Apollo. Unrealistic schedules hyped to win the contracts. Yet NASA failed to double, triple, or quadruple those hyped times to actually estimate how long it would really take to produce a new flight vehicle.
(BTW, those two 8-year examples in the previous paragraph demonstrate conclusively just how idiotic the Artemis timeline really was, and still is. That's SpaceX. Blue origin is no better. Their number might be a bit different, but it will be comparable.)
Why did NASA fail to allow for that? Simple: all the people who did the technical or the managerial work on Apollo before, were long-retired or dead (more dead than retired), unavailable to pass-on the 50++% of the required knowledge and skills that was never written down in the contract reports! And do not kid yourself, that effect is quite real! No manager EVER wanted to pay for writing EVERYTHING down, only what the contract required! That's about 50% undocumentd in production work, and far higher in development work. Been there and seen it and done it myself! I was roundly hated by upper management for trying to write down everything that I could.
GW
Bob:
As I have repeatedly said, there exists simply NO definitive estimate of Starship v3 payload capability to LEO (or anywhere else) until it finishes experimental flight test! There is only a projection of what they intend it to be. Asking for a firm estimate, during experimental flight test aimed at configuration development, is pointless! Essentially by definition, there is NOTHING you can trust! There CANNOT be!
But complaining repeatedly about there not yet being a definitive value, just shows your own lack of understanding about what experimental flight testing does, and does not, do! To do it successfully, you test your significant changes one thing at a time! Lest you obscure the effects your changes made! That has been true in all sorts of development testing since long before the Wright brothers!
You have to test and verify a whale of a lot of things to get your first successful vehicle configuration. Then there are fewer changes to verify in test, when you come up with each and every variant (those would be dedicated tankers, freighters with big clamshell doors, and the HLS lunar lander). But you still have verify each and every one of those changes! And if something does not verify, you have even more changes and tests to do! Which ALWAYS happens!
If this gives you the impression that there can be no reliable hard-estimate schedule for this process, then you got the right impression! Trying to maintain a hard schedule during development testing is a bullshit notion at the very best. The wise government program manager knows this, and at least doubles, preferably triples or even quadruples, the contractor's proposed development test schedule, based on prior performances in previous development contracts. Apparently there are no longer any "wise" government program managers any more! And with "new space" contractors, there is little or no prior development contract performance by which to judge! So quadrupling is the best choice for your schedule to be realistic! NASA DID NOT DO THAT!
Instead, NASA "believed" the Blue Origin and SpaceX proposal hype about how fast they could do the work! It is hyped in order to win the contract. That is just the fundamental nature of government contracting. THAT nonsense is what underlay NASA's Artemis landing schedule, that has so clearly proven to be utter nonsense!
Spacenut:
I pointed out just above in post 2407, and again in the paragraphs just above, that NASA's Artemis schedule is nonsense, and I have repeatedly pointed out why: projecting a landing date when you have no lander ready to start experimental flight testing, and preferably nearing the end of experimental flight testing, is essentially an egregiously stupid thing to do!
Both SpaceX and Blue Origin are still some years away from having lunar landers we can trust with crews aboard, with SpaceX more years away than Blue Origin is, in my opinion. Landing on the moon next year is as nonsensical as was last year's setting that date for this year! Neither Blue Origin's bigger crewed Blue Moon nor SpaceX's HLS will even be flying experimentally by next year, not with Blue Origin's launch pad blown up, and SpaceX still trying to make its first basic orbital freighter work right! (SpaceX still has serious bugs in its Raptor-3 engines, and it is painfully clear from flights 12 through 14 that they do not yet understand those bugs! Without that understanding, those bugs CANNOT be fixed!)
At least Blue Origin's lander design concept comes closer to something that might actually make successful landings on soft dirt with rough terrain features, and they have a mock-up at least, and are just starting to build a first prototype of the smaller unmanned Blue Moon lander. The SpaceX lunar lander design (HLS) is still just a paper design, and I personally would not trust it to land successfully in soft dirt with rough terrain! A non-successful landing really means a dead crew, and we all know what a dead crew costs! We've seen it twice before: $billions and years!
Neither outfit is flying anything like their landers experimentally yet, much less anything remotely "operational". That's true today, and was even more pathetically true 2-3 years ago when this Artemis schedule crap was first announced! And when it was first announced, they did not even include a lander checkout mission in Earth orbit, before trying a manned landing on the moon! Recipe for a dead crew, that was! So, how stupid was THAT?
GW
If by "A" you are talking about the Artemis program, actually I agree with you.
I think the Blue Origin lander has a far greater chance of working, than SpaceX's HLS variant of Starship (which still does not yet exist, even as a prototype for test). Blue Origin at least seems to understand wide stance legs with low cg height, although I think their cg height is still too high.
Whether they (Blue Origin) have enough pad area not to sink in, is an open question. It's not about static pressure loads, it is about transient dynamics and also the effects of the unevenness of the loads put upon the supporting surfaces.
I have seen NOTHING from SpaceX that indicates any understanding of the issues one must address when attempting rough field landings in soft dirt. They have a lot of stage landing experience now, all of it on smooth, level, hard reinforced concrete landing pads. Precisely ZERO of it was landings on soft dirt, much less soft dirt with surface slopes and grade irregularities, even boulders.
I think SpaceX is still some years away from having a lander we can trust. I think Blue Origin is also still a few years away, too. Not as many years as SpaceX, though!
I'm not much worried about the ultimate operational payload capacity of SpaceX's Starship/Superheavy, as you know. Nor am I worried about that of Blue Origin's New Glenn. Both will probably perform approximately as advertised, once all the bugs are worked out.
I am very, very worried about either of them being able to supply NASA with a manned lander for Artemis, especially on NASA's BS schedule. But I am far more concerned about SpaceX's lander ever being viable, than I am Blue Origin's lander!
GW
Spacenut:
I quite agree. They have not thought all this through properly yet. The flight tests should include landings on desert sands and soft muds here on Earth. The desert sands should include terrain with big rocks and creek beds or washes, plus surface slopes.
GW
Care to tell us what your second elephant in the room might be?
GW
That "soft dirt" here is dry, fine, sand dune sand. The min allowable stress for that material is 0.1 MPa, or in US units, one US ton per square foot, which corresponds to rounded sand grains. The max is only twice that at 0.2 MPa = 2 US tons per square foot, corresponding to angular sharp grains. Those are allowables for long-term resistance to settling, the actual immediate failure pressures are usually factor 2 to 2.5 lower. EDIT 10-3-2026: meant to say higher! Not "lower".
I successfully used criteria like that for designing the thrust-blocking and dead-manning of underground fire water mains, compliant to the fire codes, which must be unburied for the fire marshall to witness. Test pressure is usually specified as 175 psi, unburied in any way, and absolutely NOTHING can move! If he wants the pressure held like that for days, you do it! Most of that stuff is 8-inch or larger slip-joint pipe. There is essentially ZERO strength in slip joints! The water supply people get away with burial, and only about 60 psi. That's why water mains blow out and get repaired all the time, while fire mains work without failure for a century or more.
Both Mars and lunar regolith are similar to dry, fine, Earthly sand dune sand, with lunar the sharp particles, and Mars the rounded. The presence of rocks and gravel in that regolith does NOT help with strength, until and unless there is so much rock and gravel that all the stones touch each other, everywhere within it. Most places they do not!
I got this understanding from the civil engineering texts and references that deal with foundation design, where the load is the static weight, to be supported upon a certain area. I originally trained in aerodynamics, thermodynamics, and propulsion, but building fire mains was something I had to do, and professionally, after the big drawdown layoffs after the fall of the Soviet Union. So I learned it.
I also learned how to build slab foundations for high-expansion clay soils, and I did that very successfully. It is very rare when a customer is willing to spend for the concrete and steel to do that job right! But about a quarter century ago, one did and built what I designed, and those two houses have NEVER, EVER moved and cracked in that quarter century! That's for a clay plasticity index near (or above) 30, when the federal loan rules say that a plasticity index of 15 is "extremely expansive"! Worst I have seen around here is 60 to 70. Most of the clay around here is Bentonite, which is what everybody else around the country uses to keep the water in their ponds from sinking into the Earth. That delays the effects of rain under a foundation by about 6 months to a year or so. That kind of clay can swell when wet by almost 10%.
From both mechanical and civil engineering comes the practice of multiplying the static load by certain factors to account for low-speed dynamic effects, and certain other factors to account for the unevenness of load application across the intended supporting area. I typically use factors of 2 for both effects, for a factor of 4 applied to the static load. Works well enough, and without the need to run other people's computer codes. Fits right in with my pencil-and-paper engineering abilities.
Unless SpaceX comes and talks with somebody like me, their first attempt at a soft-dirt landing is going to be really, really disappointing. That kind of stuff is NOT taught in aero or mechanical engineering schools. It's civil engineering. I did not originally train in civil, but I learned some of it to continue having a way to make a living after aerospace-mechanical engineering effectively died about the mid 1990's. I retired before it ever really came back. And it came back in the form of kids underpaid to run other people's computer codes, without any ability to do the pencil-and-paper stuff that I do.
Some of the hazardous material experience from aerospace-mechanical stood me in good stead, though. Once of the things I had to do was dispose properly of over-age science lab materials from public schools. I did have to recognize and properly treat both explosives and deadly poisons doing that work.
GW
Multiple layers of bricks with non-overlapping joints does help, but not as much as we might wish, until the bricks are glued together in some way. The load only spreads at the angle from brick to brick vertically, if not bonded together. You need a lot of layers (many dozens) to spread it enough to put concentrated loads onto soft dirt.
GW
Not every payload item Starship will ever carry will be a 2-ton flat slab of a Starlink satellite, deployable by a pez dispenser.
Of course there will be versions that fly with clamshell doors, to let large things ride up and be deployed! The heat shield requirements will put the limits on that clamshell design. The current heat shield extends past the 90-degree point on the sides, and a little ways past the nose on the leeside, with some further extensions near the flap hinge lines. The hinges for any clamshell doors need to locate where the heat shield is not located, almost by default!
There will also be versions either with larger main tanks, or with interchangeable payload propellant tanks in the cargo bay, to be used as dedicated tanker vehicles. Interchangeable tanks in the cargo bay makes more sense to me, since ultimately, other vehicles using different propellants will need their refueling capabilities. And you ALWAYS load the same propellant into any one tank, since it gets into the pores of whatever material you use to fabricate it.
GW
Part of the payload carried by the Starship includes the pez dispenser of those v.3 satellites. That's a few more tons, in and of itself. But I stand by what I said in post 2396 above! It is stupid to carry max payload in a vehicle still deep in experimental flight test. You test the effects of only 1 thing at a time. Known since the Wright brothers, if not before.
Most of the stuff posted on SpaceX's website as I just now visited it, is data for version 2 Starship/Superheavy, not version 3, and for Raptor-2 not Raptor-3. There is no distinguishing a sea level Raptor from a vacuum Raptor on their website (most of the data looks like sea level version 2 to me). There is (and never has been) any information about stage inert masses (what I used a few years ago was from Wikipedia).
A few years ago I used what I could find to estimate the payload capacity of a version-1 Starship/Superheavy to low-inclination LEO. It was near 150 metric tons payload. But, the version-1 Raptors never achieved their design max chamber pressure, and so never achieved their design thrust or specific impulse. That shortfall made my numbers junk. Doubly so, since a thrust shortfall lowered liftoff thrust/weight, and that correlates to GREATLY-INCREASED gravity loss that the first stage must "cover".
So, we are arguing angels on the head of a pin here, with no reliable data as to stage inert masses, or enough engine details to estimate their thrusts and Isp's. And apparently, not even reliable figures for propellant capacities on the SpaceX website! Without reliable numbers, all performance and payload capability estimates are BS, no way around that!
I do NOT see the point of this argument! Not until there are reliable numbers available by which to check the payload claims on the SpaceX website, which as I saw them today, are "100+ metric tons" to (implied low-inclination) LEO. They know what their numbers are, we do not! And they do not publish them, precisely to keep competitors (and people like us) from reverse-engineering what they are doing!
GW
Oldfart1939:
If I understood the video correctly, they are claiming that Starship HLS's cg will be low enough to be stable on fairly short legs at 8 degrees effective slope. The numbers I ran for Starship on guessed short landing legs showed tip-over static instability at 5 degrees, so there is that issue. At least, the video indicates they (SpaceX) are aware of the issue.
There is the separate issue of rocket blast kicking up dust and rocks. I thought that SpaceX was planning not to use the Raptors for the touchdown (for one thing they do NOT throttle down to the 5-8% thrust the video talked about). Instead, there were to be a ring of things resembling Super Draco's, as canted tractor rockets way up near the nose.
I was a little taken aback by the video talking about "high wind landings" on the airless moon! It may be more AI slop than anything generated by a knowledgeable human. That is one thing to be aware of. There is no policing for truth on Youtube. Or pretty much any other site.
There is another serious issue for landing leg design that the video seemed to mention too briefly and then quit addressing. That would be unwanted/unintended horizontal speed at touchdown. Which is mostly what tipped over 2 commercial landers so far.
As the vehicle rotates about the effective contact point on the leading side of the leg pattern, its cg will rotate a bit upward in an arc about that contact point. That's a rise in potential energy. If the horizontal kinetic energy exceeds that potential energy gain, you WILL tip over due to the unwanted horizontal speed! Plus, if the leading leg (or legs) digs in, the contact point rotation center is pushed below grade, reducing the PE gain that you need, for overcoming the unwanted horizontal KE.
That brings up two things about the pads at the ends of the landing legs: they have to be big enough not to sink into soft regolith, and they need to be tipped inward off of horizontal, so that the lead pad(s) leading edge(s) cannot dig in and "trip" the vehicle as it touches down. It needs the ability to "skid" on the leading side, while digging-in ONLY on the trailing side. The video ignored these issues entirely.
There is also a serious experience problem to consider here: SpaceX has ZERO experience landing anything on soft dirt. The only things they have ever landed on were smooth-and-level, hard-surfaced, reinforced-concrete landing pads. Landing in the soft dirt is entirely a different experience!
The robot that makes bricks put of lunar regolith is not the solution for landing large things on landing pads it constructs, unless you can bond the bricks together into a single piece. The landing pads will just push the loose bricks down into the soft regolith, locally, right under the pad. The pads still sink in, just maybe not quite as deeply.
GW