You are not logged in.
The first time I know of for point-to-point rocket delivery was Von Braun attempting rocket-delivered mail in the 1930's, before he came to the attention of the Wehrmacht.
GW
The nozzle needs cooling because the most practical materials to build it are metals, and none of those survive 3000 K temperatures. The carbon-based chamber material can survive 3000 K temperatures, and if it has the strength to contain 5-atm-class pressures at such temperatures, then it needs no cooling.
Thermal expansion is going to be a real bugaboo with such a design, if you try to affix the engine to the vehicle at more than one axial station. I wouldn't do that myself, but I would make the mounting point and hardware very strong and stiff, so as to maintain firm control over the thrust vector direction line passing through the vehicle cg, no matter what thermal expansion occurs.
There is a downside and a work-around: firm mounting precludes easy gimbal-type thrust vectoring! You will need some other form of attitude control.
I would suggest attitude control thrusters. You can tap some hydrogen off the main chamber to run them, up nearer the cooler front end,
so the not-so-hot gas is easier to handle with metal tubing and valves.
Attitude thrusters are how we had positive flight control on the 4-stage solid-propellant "Scout" launcher without trying to do flexible-structure vectorable nozzles on the solid motors! Those were very simple, very lightweight hydrogen peroxide monopropellant decomposition thrusters. Electrically-heated catalyst beds in the thruster chambers. This required high-test peroxide, but the storage time before use was short enough to avoid that danger.
Trying to do vectorable nozzles on solid motors was very expensive, usually way too heavy in small sizes, and added multiple possible leak failure modes to the design. Only the strategic-size motors had them. They are usually hydraulically actuated, requiring both substantial motive power and some very heavy equipment to run them.
GW
The water column height that makes this work is not the top to bottom height of the J-tube. Is is the difference in elevations between the outside entrance surface and the habitat floor level surface inside. There does not need to be a deep bend.
Put the habitat instead just under the crater floor, and the entrance instead 13.5 m (less the hab floor burial depth) higher up the adjacent crater wall, and you do not even need the ladder!
Biggest problem will be the water freezing, or the water evaporating at a high rate from the exposed water surface at the entrance. I'm not sure which will dominate.
But the local atmospheric pressure on Mars is 6 to 7 mbar generally. That's not enough to stop violent boiling away of the water. Even if it is a bit higher in some areas so that the violent boiling didn't occur, you would still have a very high evaporation rate!
The only thing that stops the (non-violent-boiling) high evaporation rate is when the vapor pressure of the water in the local air, not just the local air pressure, exceeds the equilibrium vapor pressure of just-at-freezepoint water. Warm the water and that vapor pressure requirement is even higher.
What you really need for such an application is a liquid with almost no vapor pressure at all to slow its evaporation rate, a much lower freezepoint to limit freezing effects, and preferably a much higher density. The higher density would lower the height difference between the exposed liquid surfaces, making the design far easier to implement.
Mercury immediately comes to mind, but with all the toxicity problems it brings.
GW
Rob:
I don't know what "Starship siphon" refers to. But I think your suggestion of propellant slosh may be right. That surface is variably sloshing around, which may transiently uncover the suction intakes for the turbopumps when the propellant level is low enough. That would be a separate problem from the possible POGO structural instability driven by thrust oscillations.
Bob:
My old silicate composite material has a limited temperature range and while stronger than NASA's shuttle tiles, limited strength. It was something that worked by chance for me 4 decades ago, but was never developed since.
Ed Pope has several modern unique materials for very harsh temperature and structural environments, all of which are better than my old stuff. His zirconium carbide oxide material in particular caught my eye.
GW
Max efficiency by itself is a very misleading measure! It also has to survive extended operation at full soak-out temperatures, and for a duration that suggests a decent lifetime. The ones that meet those criteria have yet to exceed about 27% conversion efficiency. So says AAAS's "Science" Journal, which is peer-reviewed to a fault.
GW
Today I finally did make the video work of Flight Test 8 on the SpaceX website. I DO NOT do X or any of that other social media BS. I never will.
I saw no surprises, except that the Starship upper stage loss occurred right before the end of the ascent burn, with little propellant left aboard. I did see first stage engine losses on both ascent and descent, excepting the 3 gimballed engines dead center. That makes me wonder if we didn't see POGO in both stages.
If the control software is formulated correctly, ascent oscillations risking engine damage might cause shutdowns that do not recur during descent. Except that there were 2 engines that failed to restart during the 13-engine initial boost-back burn. 11 were enough to slow it so that the "design 3" could recover it on the tower.
The upper stage Starship did not seem to suffer problems until right before the end of its ascent burn onto the ascent ellipse. Then it lost 1 vac engine, then all 3 SL engines (with the thrust vector gimbal capability), then another vac engine, then it obviously tumbled totally out of control until self-destruct ended this.
Bear in mind that this ascent ellipse has apogee at LEO altitude, but a perigee at, or very near, the surface. THAT is what ensures automatic de-orbit, if you do not do a circularization burn at apogee. The ascent ellipse perigee is well down in the atmosphere.
GW
Rob:
I tried but could not make the link you provided work. Since I come from the slide rule days, that should be entirely understandable, and yet totally outside anything SpaceX is familiar with. I will try some other way.
Tom:
I never applied at SpaceX in any of the last several years, after finding out some years ago that they hire no one over age 40 to 45, because they demand chronic 70-80-hour weeks. Oldsters cannot do that. So they do not want them.
Both:
The problem is now that SpaceX is in deep kimchee with what looks like POGO troubles with Starship. No one there in California or any other site would have the slightest idea how to recognize this as a problem, much less have any idea how to approach solving it. That is because (1) no one has faced this trouble in some decades, and (2) this is 80-90% art never written down decades ago. All the people who knew are long dead or retired, like me. Even even what I know is only tangentially useful, as it deals with solids, not liquids. The real expert in solid instabilities was my friend and mentor W. Ted Brooks, who died many years ago.
The real experts in liquid instabilities would have been those who designed the J-2's and F-1's for Saturn-5. They are all long dead. Period. 10% or at most 20% of what they knew was ever actually written down anywhere at all. It was development stuff, not production stuff, back then.
Production: 50% art, 40% science, 10% blind dumb luck. Only the science was ever written down.
Development: art and luck factors very much higher.
GW
When the cost of rocket travel becomes low enough, you will see a point to point industry appear and grow. That is true.
I rather doubt rocket travel will ever be as cheap as airplane travel, but there is a value to very short flight times with some cargo items or people, which justifies the higher price.
In my own opinion, we are not there yet, not even SpaceX.
Crudely, $1000 airline ticket for 100 kg person = $10/kg. I think it rather insane to believe rocket flight can ever get to be that cheap, because it is inherently more demanding than cruising in a high-subsonic airplane. But if it gets down near $100/kg, that could create an industry transporting stuff on suborbital flights of only minutes instead of hours. Factor-10 higher price for factor-10 shorter travel time.
GW
Rob:
That's exactly right: the susceptibilities shift dramatically as you burn off propellant. We saw the same things in solids: as the propellant burned back, the internal cavity drastically changed is size and shape, bringing different resonant frequencies into play. You should find it scary. It really is!
With the solids, we had to analyze the shit out of it, with our best shot at instability computer models (which were imperfect at best), and then test the shit out of it, with the right analog instrumentation, all done by people already experienced using that analog instrumentation for instability detection, before we ever let a solid design out the door! Nonmetallized reduced/min smoke required much greater care and attention in this area, because there were no metal oxide particles or droplets to provide any damping action, just the soot. Absolutely NONE of that was a single simple calculation. Test data ALWAYS trumped any analyses, too!
Actually, the same thing is true of liquids, just the details differ.
I suspect the "new space" outfits will be seeing a lot of POGO-type troubles as time goes by, because none of the youngsters at those outfits know anything substantive about instability (since that remained about 80-90% art, never written down). And none of those youngsters knows anything about the analog instrumentation required to diagnose it, or anything about how to go about doing this kind of work with that analog instrumentation.
Myself, I think SpaceX accidentally dodged the instability bullet with their kerosene-fueled Merlins, because of the damping effect of the soot cloud in the chamber flame. A soot cloud is inherent when burning long-chain carbon things like kerosene. Vaporizing droplets can assist with that damping effect. But with the methane-burning Raptors, there is no soot on the flame, and the droplets evaporate far quicker.
Since SpaceX hires no one past age 40 to 45, how would any of their engineers who design these things know anything at all about such instabilities? Or even recognize them? Much less how to deal with them when they crop up? And they DO crop up!
"Build it, break it, build another" is a strategy that will always fail, if you cannot recognize the problem to be corrected.
GW
The nuclear light bulb version of gas core nuclear fission rockets had a nested pair of quartz cylinders as its engine chamber. The cold liquid hydrogen flowed between them, getting heated on its way to the nozzle, by the uranium gas fireball inside the inner cylinder. This was late 1950's early 1960's design concept work. The materials and technologies were never finalized.
GW
Rob:
The process you describe in the F1 is what causes serious thrust oscillations, which can build up and destroy the engine, or might not (just run very rough).
When that thrust oscillation interacts resonantly with either (or both) vehicle structural oscillation modes or organ-pipe fluid oscillation modes in tanks or plumbing, the entire vehicle is at risk of breaking up. The feedback is feed pressure changes feeding back into the rough-burning thrust oscillations. The term "POGO" encompasses all of those.
The Apollo-6 problem had thrust oscillations in the first stage kerosene-LOX F1's that shook the vehicle so violently as to damage the LH2-LOX J-2's in both the second and third stages. The Apollo-13 problem was J-2 engines shaking so violently in the second stage as to cause 1 engine to shut down lest it explode. Different, and yet many of the same things.
The Titan-II man-rating for Gemini was delayed by thrust oscillations that broke up test vehicles in flight. Those were NTO hydrazine storables. I'm not sure, but I think Atlas might have suffered from some of this, too.
GW
I've already voiced my opinions in other threads of Elon Musk as "Chief Engineer" of SpaceX. He has a BA in physics and a bachelor's in economics. All his other credentials are honorary. Look him up in Wikipedia.
The real engineer's at SpaceX are Glynn Shotwell and her team. They usually do pretty good, when Musk doesn't over-rule them, or piss off the FAA. But there are exceptions, as with any outfit.
But with the Musk legacy of hiring no one over 40-45 years of age, in order to be able to demand chronic 70-80 hour weeks, and the high turnover rate that causes, there are no oldsters at SpaceX who know much of anything about POGO instabilities! I think that's probably what they've bumped into.
And don't forget, the "science" in rocket science, being the stuff that was written down, is only about 40% or less of what you need to know, especially if you bump into a difficult, relatively-intractable problem. That is EXACTLY where the strategy of "build it, break it, build another" will let you down.
You need some oldsters who know the art, which was never written down because nobody wanted to pay for writing it all down. There would be few to none oldsters who would know the art of treating POGO, as the last time that was a big concern was 6 decades ago. I'm not one who knows that art myself. Solids and missiles are different from liquid-engine launchers.
GW
From AIAA's "Daily Launch" for Wed 3-19-2025:
Ars Technica
Here’s the secret to how Firefly was able to nail its first lunar landing
Firefly Aerospace's Blue Ghost science station accomplished a lot on the Moon in the last two weeks. Among other things, its instruments drilled into the Moon's surface, tested an extraterrestrial vacuum cleaner, and showed that future missions could use GPS navigation signals to navigate on the lunar surface. These are all important achievements, gathering data that could shed light on the Moon's formation and evolution, demonstrating new ways of collecting samples on other planets, and revealing the remarkable reach of the US military's GPS satellite network.
------
my take on it:
The link to the much longer Ars Technica article is where the "secret" really is. They have a photo of the Blue Ghost in a shop before loading and launch. The photo shows the min pad span directly (of 4 legs). A good guess for the cg height "just looking at the photo" says min pad span/cg height ~ 2. The old criterion is 1.5 to 2. They really tried to do rough/soft field landing right, and it shows in the longer Ars Technica article's descriptions of all the things they did t ensure a good landing. Their software and cameras were able to do what Armstrong and Aldrin did on Apollo-11: see and avoid obstacles.
GW
I downloaded the pdf paper/report. Still have to digest it fully. But I did notice there was no surface bearing strength data (they had no way to do that). The particles physically look like Earthly sand dune sand, maybe not quite as rounded. Sharp-edged particles on the moon still behaved for the Apollo astronauts like walking in fine dry sand dune sand, so it seems likely the sharpness of the particles has low effect on surface bearing strength. For Earthly "soft fine dry sand", the allowable bearing load pressure for foundations is about 0.1 MPa ~ 1 US ton/sq.ft. At an estimated-as-common factor 2 reduction failure/allowable, that puts the surface failure pressure for landing pads digging in, at about 0.2 MPa ~ 2 US tons/sq.ft.
For vehicles landing and taking off, you must meet only the higher failure pressure as a maximum value. For roadways and structure foundations, you have to meet the lower allowable as a maximum value. You do your actual designs under those maximum values.
We have observed lots of dusts and sands on Mars with loose rocks in it. Unless these rocks are in direct contact with each other, they do not reinforce the strength of the sand. That is because there are no cementing effects. Solid outcrops flat enough to land upon are quite rare. Places where the loose rocks in the sand actually touch, so as to reinforce it, are also observed as rare. So, this "soft fine dry sand" assumption applies to 99% of the surface of Mars!
Dust adherence to surfaces is affected by the sharpness of the particles, worse if sharp. Dust contamination-caused wearing and damaging of machinery and structures is also worse with sharp particles. We learned that from Apollo. As near as I can tell, the "sharpness" of Martian sand and dust particles falls somewhere in between that of Earthly materials and the lunar materials. Whatever we find that works on the moon should work on Mars.
Bear in mind that we have yet to find what really works on the moon! Apollo found the problems but not the solutions.
GW
There are 3 options for going to Mars (or anywhere else). Those are (1) a flags-and-footprints mission (or missions) for geopolitical reasons, (2) to establish a scientific base only for research purposes, and (3) to establish a scientific and engineering research presence with an eye to permanent settlements. The hardware requirements, and the science objectives, are quite different for those 3 options.
Accordingly, one has to figure out which of those 3 overall objectives one is going to presume. No one has really done that since the voyages of exploration 500 years ago, and they generally didn't do it very well then.
GW
Update: I would only add that the "3-phase process" I defined applies ONLY to the overall objective of permanent settlement. It does not apply directly to establishment of a science-only base (although it overlaps some), and it absolutely does NOT apply to flags-and-footprints.
Those phases for permanent settlement were (1) exploration (what all resources are there, where exactly are they, how much is really there, and how hard will getting it be), (2) experimental base (learn how to live off the land with the resources found in exploration), and (3) start building the settlement ONLY when the experimental base phase is fully successful and the appropriate infrastructure put in place! That infrastructure has to grow as the settlement grows.
The distinction here is OVERALL OBJECTIVE (flags-and-footprints vs science base only, vs create-a-permanent-settlement), vs the 3-PHASE PROCESS that I identified for creating-a-permanent-settlement.
Apollo was only flags-and-footprints.
The Antarctic scientific bases rely entirely on supplies shipped in. None of the research they do has anything to do with living off the land in Antarctica.
Exploration with an eye to settlement hasn't been done since 300-500 years ago.
The fraud is perpetuated by those whose livelihood comes from selling fossil fuels, which inherently increase CO2 content in the atmosphere.
I do NOT believe there is a "saturation" level of CO2! That notion is a new part of the fraud-for-profit. It has been far higher in the distant past on Earth, and it IS far higher on Venus and on Mars. It was life on Earth that helped pull CO2 out of the atmosphere and deposit it as limestone on the bottoms of the oceans. That effect may have been too small over geologic time on Venus, and it was likely overwhelmed by atmospheric erosion on Mars by the solar wind without a string magnetic field.
I know the right-wing and far-right media will do or say anything to deny climate change, but their motive for that fraud is very simple and easy to understand: short-term profit.
GW
Comparisons to Shuttle are totally invalid, because it was NOT a "classic" two-stage vehicle! While none of the main engine propellants were in the orbiter, its main engines fired from launch all the way through the initial ascent. Their propellants were in the external tank. The only engines with propellant aboard the orbiter were the OMS engines and their storable hypergolics. These were used for orbit insertion at the end of the ascent, for on-orbit maneuvering, and for the de-orbit burn. The thing could not thrust greater than weight for launch without the solids. It was a sort of complicated stage-and-a-half. Unique is what it really was.
Myself, I think talking about Starship inert masses under 120 metric tons is utter rank nonsense! Are you listening, Musk?
But very substantially higher is also nonsense! Higher, maybe, but not dramatically so.
Growing the length makes that inert mass higher, yes, but it also makes propellant mass higher, at a faster rate. It is simply too early in the development flight test program to reliably predict what the final inert mass of a Starship (whichever version) is going to be. Mass ratio goes up when you do that, not down. And mass ratio determines the propellant mass fraction. For which if inert fraction remains the same percentage, gives you a higher payload fraction.
That being said, it is really easy to increase the inert fraction, which increase comes directly out of payload fraction.
Let's see if they can solve the POGO oscillation problem. There's a very significant thrust oscillation exciting longitudinal structural resonances in the longer updated Starship structure, and in the propellant feed plumbing. They really need to smooth out that thrust oscillation. Methane leaks easier than kerosene. It's almost as bad as hydrogen.
The new Raptor-3's are being tested at McGregor, and they have a rather substantial oscillation at 5-10 Hz, which rattles the ever-loving hell out of my doors and windows, but is too low a frequency to hear with the ears (the low-end cutoff for humans being about 20-25 Hz). The Raptor-2's had a similar oscillation, but at less amplitude. I heard/observed nothing like that with Raptor-1, or any of the Merlins.
This thrust oscillation in the Raptor-3's is large enough that I consider it "combustion instability". At the old rocket plant, we never allowed a design out the door with oscillations that large.
GW
I found some dimension numbers for the Intuitive Machines "Athena" lander, plus some information that its altimeter failed before landing, as did its predecessor on "Odysseus", of similar dimensions. I used those dimensions to estimate crudely the min pad span to cg height ratio, and from that the critical static tip-over angle (roughly only 14 degrees). That plus the dimensions gave me the tip-over radius gone vertical, which on the moon implies a max horizontal speed at touchdown of only about 0.5 m/s.
Without an altitude figure, there is no way the lander could adjust its flight to touchdown at near-zero vertical and horizontal speeds. The far greater likelihood is that those vertical and horizontal speeds were multiple m/s. What that says is that neither machine could possibly land upright. The first "fix" is to get a reliable altimeter. But there is more to be fixed:
Even had the altimeters not failed, so that touchdown speeds could be zeroed, there is still a high probability of a tip-over, with the critical static tip-over angle being only about 14 degrees! At the dimensional scale of the lander, slopes on the moon are already known to often be much higher than that 14 degree figure, anywhere near the rims of craters! That critical tip-over angle needs to approach 40 degrees, which it did with the Apollo landers and the Surveyor probes, long ago. And recently with Firefly Aerospaces's "Blue Ghost".
And THAT HIGH TIP-OVER ANGLE is "short-and-squat", not "tall-and-narrow"!
SpaceX's lunar Starship variant has exactly the same risk, inherently being tall and narrow: something like 50-60 m long, with a pad footprint "diameter" closer to 10-12 m. Maybe 15 m. Maybe. Doesn't matter: any of those proportions are at least as bad as the ones for "Athena" and "Odysseus".
Are you listening out there, SpaceX? You ought to be, you're supposed to land humans on the moon with your vehicle in the next year or three! It would be nice if they survived that landing. Or doesn't your contract specify that?
GW
Firefly's "Blue Ghost" is a "short and squat" design, with feet farther apart than the height of its cg above the surface That is painfully obvious just looking at the photo. It meets the old criterion for rough field landing reliability that was so successful for Surveyor 1 and Surveyor 3, and all the Apollo landers that touched down.
Both of the Intuitive Machines landers were rather "tall and narrow", with the cg height larger than the pad span by far. Those designs violated the old rough-field landing criterion. That one is 1.5 < cg height/min pad span < 2.
Anyone else see the pattern here?
GW
You won't see POGO in a static test of stage on a stand. It is necessarily fixed to the stand, and that fixity precludes many of the vibration modes available in flight. That was precisely why they didn't see POGO until they flew, decades ago. It afflicted a lot of rockets. Titan-2 and Saturn-5 being only 2 examples.
I agree with you that SpaceX needs a true Chief Engineer who is a real engineer. Musk does not qualify, never did. He earned a bachelor of arts (not science!) in physics, and a bachelor level degree in economics. All his other "qualifications" and "certifications" are only honorary. Go look him up in Wikipedia. It's all there.
At one point, he was going to move SpaceX headquarters to Texas, but he never did. I think I might know why. Under Texas law (the engineering practice act), it is seriously illegal to publicly call yourself an "engineer" in Texas, unless you really are one. He calls himself SpaceX's "Chief Engineer". His ego would not let him relinquish that title.
Engineers who build public works in Texas must be board-certified and registered as professional engineers; it is illegal to practice public works engineering otherwise. For companies building things that are not public works, the engineers to not have to be registered professional engineers, but the top dog who signs the drawings needs to be. Some outfits get away with violating that, and apparently SpaceX is one.
GW
I posted about POGO in the "interplanetary transportation" topic, "starship is go" thread. I found a decent Wikipedia article about POGO while researching which Saturn-5 flights had POGO problems. The article quoted George Mueller (formerly of NASA) as saying it's some sort of thrust oscillation in the engines that starts this. The article quotes him as saying it's some sort of "combustion instability". That frequency of oscillation can interact with longitudinal structural modes, and with organ-pipe resonances in the feed plumbing, as a sort of pressure difference thing affecting propellant flow rate. I presume some liquid slosh modes in the tanks could affect this as well, as they would oscillate the feed pressure into the propellant feed plumbing.
The article indicates the unmanned Apollo-6 test of an Apollo-Saturn-5 suffered severe first stage POGO issues that fatally compromised the mission by damaging the second and third stages. The Apollo-13 second stage suffered POGO severe enough to shut one of the five second stage engines down. The other 4 burned longer, as did the 3rd stage, to compensate. This difficulty, which came close to an abort, was overshadowed by the oxygen tank explosion that crippled the craft and forced the emergency abort home.
The article also explicitly mentioned that man-rating the Titan-2 booster for Gemini was delayed because of POGO difficulties in the Titan-2. The article didn't say, but I seem to remember some Titan-2's blew up from this problem, as they were being tested during the Mercury flights.
Combustion instability in a solid is a little different, being usually "something" that causes a pressure oscillation that affects burn rate. If these phase "correctly", the oscillation can grow, raising average pressure, until the motor bursts. Reduced and min smoke solid motors suffer more from this. Smoky motors with aluminum in their propellant see a significant oscillation-damping effect from the inertia of the metal oxide particles resisting the oscillation by their aerodynamic drag in the chamber gases. Those are orders of magnitude larger than the soot particles.
Combustion instability in a ramjet is different yet. Usually it has to do with vortices being shed from something solid. When the vortex detaches, it then "sucks in" the surrounding gas from the region into which it travels as it spins. A fuel-rich vortex "sucking in" air-with-oxygen then ignites and explodes. If the pressure wave from that explosion causes an enhanced rate of such vortex shedding, you have the necessary unstable positive feedback of energy for instability.
But these all share the same characteristics: there is random combustion noise with a very broad spectrum of frequencies. Its pressure amplitude is usually nearer 1-2% of the basic pressure level. Where the oscillation frequency matches some resonant mode in the chamber or associated passages, you will see those pulses of higher amplitude with a definite frequency you can measure, on a pressure-time trace, if you have the right equipment to see it. It will stand out above the "background" noise. There can be multiple modes resonating, more than just one is common, actually. The are usually fundamental modes or first octave modes, of the many resonant modes the chamber shape might have. Noise at 10%-or-more of basic pressure level, we deemed to be "instability" requiring remediation. Which was usually a complete redesign.
You will NOT be able to see these phenomena in situations where the signals are acquired with digital data acquisition, because nobody can afford to buy the equipment that has the requisite 1 Megahertz-or-higher response band! As a result, you need to record the pressure with an analog tape machine based on FM data processing, that has 1 MHz response or better. Then you can play the test record back and plot it in multiple different formats, until you actually see what you are looking for. Digital will never do that, it is too pixelated. But FM tape recorders are old-time equipment that few outfits have anymore, and even fewer outfits would have employees knowledgeable about it, even if they bought some. You really have to know what you are doing, to diagnose combustion instabilities with such stuff. I used to do it all the time. It takes a lot of experience to get it "right" every time! Much more art than science.
We would parallel-feed our test firing data to both digital data acquisition and the FM tape recorder, making both records at the same time during the test, but only during my second tour a the rocket shop! My first tour, FM tape and real-time photo paper galvanometer oscillographs were all we had. The FM tape data recording is like the 5000 frame/second movie footage: if you have it, you can figure out what happened from a bad test. If you don't, you cannot. Guesses about this stuff are usually wrong!
GW
Oscillations causing propellant leaks, fires, and explosions? GWJ 3-12-2025
I found a good Wikipedia article on “Pogo Oscillations” which explains (1) what happened on multiple Saturn-5 flights, (2) how longitudinal mode oscillations occur due to feedback, and (3) that this has been an issue with many rockets. Text from that article is appended below.
Bear in mind that there other oscillation modes in any flight structure besides longitudinal (usually down in the few to several Hertz range), but that lateral modes are generally going to be higher frequencies. I am familiar with combustion noise in ramjets and solid rockets, as being a few hundred Hertz, but usually only 1-5% of the average pressure level. If it exceeded 10%, we usually considered that “combustion instability” needing elimination, especially in solid rockets.
You don’t measure that sort of thing with digital data acquisition! The digitization (pixelation) totally obscures the real physics displayed in the pressure-tine trace! What was required was an analog tape recorder using FM data processing, and with at least 1 Megahertz response, because sometimes the instability oscillation frequencies could be several thousand to around a hundred thousand Hertz.
You play it back from the tape and plot it by analog means in different formats, until you can see that which you need to see in order to understand what happened. Usually you can pick out the frequency of the mode that is oscillating with feedback, as it will have the largest amplitude of all the other noise. This ain’t easy to do! And few do it that way anymore! Most of the young newbies out there would have no clue how to do it, and no experience at all with the antique analog equipment that is absolutely required to do this!
This phenomenon is flying-vehicle-related, something you would NEVER see in an engine-only ground test because none of the affected hardware is even there! And, you likely would NOT see it in a vehicle-mounted static test on the launch pad! That is because with the vehicle fixed to the pad, most all the flight-related oscillation modes are suppressed by the fixity, including any dynamic slosh effects inside the tanks that cause their own pressure fluctuations in the propellant feed lines. Basic message: what happens fixed to a launch pad is simply quite different from what happens flying free!
With it only showing up in a flying vehicle, plus having only youngsters on their staff who are oriented only to digital data acquisition (they hire no engineers over 40-45 years of age), I am entirely unsurprised that SpaceX has blown up two “Starships” during ascent without really understanding the cause yet! I hope they read this on the forums. They need it to work through this issue.
The longer upgraded “Starship” has a lower fundamental longitudinal structural vibration mode than the original shorter version, sort of like a longer organ pipe. But as the propellant load burns off, any propellant slosh mode frequencies can come to interact with rising longitudinal mode frequencies as the vehicle mass reduces. Same sort of thing affects all the modes, actually. Oscillation frequency is sort of a stiffness-divided-by-mass thing. And mass is dropping rapidly with time.
That time variation in natural frequencies also helps explain why the explosions occurred later in the ascent burn, by the way! That’s when the excitation better matched some natural mode frequency, allowing positive unstable feedback of energy. The oscillation amplitude then increases dramatically and quickly to dangerous levels.
GW
Quoted text:
From Wikipedia, the free encyclopedia
(Redirected from Pogo oscillations)
Pogo oscillation is a self-excited vibration in liquid-propellant rocket engines caused by combustion instability.[1] The unstable combustion results in variations of engine thrust, causing variations of acceleration on the vehicle's flexible structure, which in turn cause variations in propellant pressure and flow rate, closing the self-excitation cycle. The name is a metaphor comparing the longitudinal vibration to the bouncing of a pogo stick. Pogo oscillation places stress on the frame of the vehicle, which in severe cases can be dangerous.[1]
NASA Associate Administrator for Manned Space Flight George Mueller explained Apollo 6's pogo oscillation to a congressional hearing:
Pogo arises fundamentally because you have thrust fluctuations in the engines. Those are normal characteristics of engines. All engines have what you might call noise in their output because the combustion is not quite uniform, so you have this fluctuation in thrust of the first stage as a normal characteristic of all engine burning.
Now, in turn, the engine is fed through a pipe that takes the fuel out of the tanks and feeds it into the engine. That pipe's length is something like an organ pipe so it has a certain resonance frequency of its own and it really turns out that it will oscillate just like an organ pipe does.
The structure of the vehicle is much like a tuning fork, so if you strike it right, it will oscillate up and down longitudinally. In a gross sense it is the interaction between the various frequencies that causes the vehicle to oscillate.[2]
In general, pogo oscillation occurs when a surge in combustion chamber pressure increases back pressure against the fuel coming into the engine. This reduces fuel flow and thus chamber pressure. The reduced chamber pressure in turn reduces back pressure at the pump, causing more fuel to come in and repeating the cycle. In this way, a rocket engine experiencing pogo oscillations is conceptually operating somewhat like a pulsejet or pulse detonation engine. If the pulse cycle happens to match a resonance frequency of the rocket then dangerous oscillations can occur through positive feedback, which can, in extreme cases, tear the vehicle apart. Other situations that can induce fuel pressure fluctuations include flexing of fuel pipes.[3][4]
Pogo oscillation plagued the Titan II first stage during its development, which delayed man-rating the rocket for the Gemini program. The Saturn V first stage (S-IC) experienced severe pogo oscillation on the flight of Apollo 6, which damaged the S-II and S-IVB stages above and likely would have triggered an abort if the flight had carried a crew. The second stage (S-II) had less intense pogo on other flights. The oscillations during Apollo 13's ascent caused the center engine to shut down about two minutes earlier than planned. The loss in thrust was compensated by longer burns from the second and third stages.
Hazard
[edit]
If the oscillation is left unchecked, failures can result. One case occurred in the middle J-2 engine of the second stage, S-II, of the Apollo 13 lunar mission in 1970. In this case, the engine shut down before the oscillations could cause damage to the vehicle.[1] Later events in this mission (an oxygen tank exploded two days later) overshadowed the pogo problem. Pogo also had been experienced in the S-IC first stage of the uncrewed Apollo 6 test flight in 1968.[5] One of the Soviet Union's N1-L3 rocket test flights suffered pogo oscillations in the first stage on February 21, 1969. The launch vehicle reached initial engine cutoff, but exploded 107 seconds after liftoff and disintegrated.[6] There are other cases during uncrewed launches in the 1950s and 1960s where the pogo effect caused catastrophic launch failures, such as the first Soviet spacecraft to the moon Luna E-1 No.1 and Luna E-1 No.2 in September and October 1958.[7]: 440–446
Modern vibration analysis methods can account for the pogo oscillation to ensure that it is far away from the vehicle's resonant frequencies. Suppression methods include damping mechanisms or bellows in propellant lines. The Space Shuttle main engines each had a damper in the LOX line,[4] but not in the hydrogen fuel line.
End quote
I saw a confirming report today (Fri 3-7-25) that Athena is on its side, and unable to charge batteries. Both Intuitive machines ended up on their sides making rough field landings on the moon. Both of the Intuitive machines were too tall for the minimum leg span dimension, thus violating the rough-field design criteria that worked since Surveyor 3. The Firefly machine met the criterion, and is upright.
Amazing how that design criterion correlates with results, isn't it?
Those who fail to learn from history are doomed to repeat old mistakes made before.
GW
From AIAA’s “Daily Launch” for Friday 3-7-2025:
AVIATION WEEK NETWORK
B-21 Needs Faster Fuel Transfer From Air Refuelers
Northrop Grumman’s B-21 Raider needs a faster gas pump in the sky when it goes into service, a top U.S. Air Force mobility leader said on March 5. The stealthy Raider needs to top off its gas tanks faster than the fuel transfer rates of the Air Force’s current aerial refuelers allow, Gen. Randall Reed, Transportation Command chief, told the Senate Armed Services Committee during a hearing.
My take on the B-21 refuel: it uses fuel faster than they can put it in. Not a good place to be! And the new tanker based on the 737 does not yet function correctly as a tanker! -- GW
NEW YORK TIMES
SpaceX’s Starship Rocket Disrupts Florida Airports With Unsuccessful Test Flight
Starship — the huge spacecraft that Elon Musk says will one day take people to Mars — failed during its latest test flight on Thursday when its upper stage exploded in space, raining debris and disrupting air traffic at airports from Florida to Pennsylvania. It was the second consecutive test flight of the most powerful rocket ever built where the upper-stage spacecraft malfunctioned. It started spinning out of control after several engines went out and then lost contact with mission control.
My take: disrupting air traffic if debris falls is part of the FAA-approved launch license. Sometimes crap ha[ppens, and you have to plan effectively for it. -- GW
AEROSPACE AMERICA
Latest commercial moon spacecraft makes troubled landing
A lunar lander that’s about as tall as a giraffe touched down on a lunar plateau around 12:30 p.m. Eastern time today, but exchanges among staff at the manufacturer’s command center showed that it experienced technical issues, and its status on the surface was left unclear when the internet broadcast ended.
SPACE
Hope is all but lost for private asteroid probe in deep space — 'the chance of talking with Odin is minimal'
The first-ever private asteroid mission appears to be over, just a week or so after it left the ground. California startup AstroForge launched its Odin spacecraft on Feb. 26, on the same SpaceX Falcon 9 rocket that sent Intuitive Machines' IM-2 mission toward the moon. Odin ran into trouble just a few hours later, however, and AstroForge has pretty much given up hope of recovering the 265-pound (120-kilogram) probe.
The video on the SpaceX website is disabled to black. I cannot watch it.
As for the two failures, bear in mind that most flight vehicle failures in flight are due to multiple things acting together. Having only a single cause is unlikely in the extreme.
That being said, after flight 7, they made alleviating changes to the vehicle for flight 8. Those did not work, apparently. Which lends strong support to the suggestion that the real causes (plural!) were not what they assumed after flight 7, which was about leaks in engine plumbing.
There was a vehicle design change after flight 6. The vehicles for flights 7 and 8 were longer, with larger propellant tanks. Which raises the specter of some sort of slosh or other unstable-surface mode in the propellant tanks causing the vibration instead, and perhaps causing the fatal leaks to be in weld joints of the lower tank aft bulkhead instead of the feed plumbing. The larger propellant masses involved would likely amplify any such effect. It's only speculation, but a real possibility that they need to explore.
If it were me, I'd re-fly the older Starship upper stage design on flight 9, with the smaller tank volumes and shorter length, but with the improvements they made to the flaps, hinges, and heat shield. If that makes the ascent successfully, when the larger design did not twice in a row, that would pretty well nail it to the dynamics of the larger tanks vs smaller tanks.
If it's some sort of liquid-sloshing in the tanks that is the source, then a change to the baffles is indicated. And the tank structures in and adjacent to the aft bulkheads may need reinforcement to better resist the unanticipated loads. They need to instrument for these possible effects, probably to include in-tank camera views, and some strain gages and suitably-located pressure sensors.
That's not to say there might not also be failures in the engine plumbing too. But whatever is going on there, is apparently being overwhelmed by something else they have not identified yet.
This sort of thing happens often in experimental flight test. It shows up more often when you make too many changes to the vehicle design between tests, too early in the flight test program. In this particular case, it may also trace to believing too strongly in computer code outputs, by engineers who cannot detect a garbage-in/garbage-out problem because they have not done enough pencil-and-paper design analysis. Or cannot, because they were only taught how to use software packages.
That problem has existed for some long time now. I saw it during the Rogers Commission hearings after the Challenger loss. The thermal-structural model they had for the O-ring joint said it was OK to fly cold. But their model did NOT match what they built! And THAT is where the garbage got into the computer, making the output garbage, too!
GW