You are not logged in.
The conversion from wavelength to frequency involves the speed of sound for physical substances and the speed of light for electromagnetic waves. You must know the speed of sound in your material in order to do the wave/frequency stuff. We are looking at liquid propellant for Starship's Raptor engines: LOX and LCH4.
I looked up the speed of sound in liquids from the online "engineering toolbox" site. They list c = 1420 m/s for liquid methane at -170 C, and c = 1056 m/s for liquid oxygen at -202 C.
Just for reference, c = 1246 m/s for liquid hydrogen at -255 C, c = 1324 m/s for kerosene at 25 C, and c = 1402 m/s for water at 25 C. It's 1482 m/s at 20 C, and 1402 m/s at 0 C, so there is also a modest liquid temperature effect. Sea water at 3.5% salinity is c = 1533 m/s at 25 C, and 1522 m/s at 20 C.
GW
From the same issue of Daily Launch" as the previous post:
------
SpaceX Starship failure casts further doubt on NASA’s moon landing timeline
NASA hopes to put astronauts on the moon in just two years, but a critical spacecraft required for the mission keeps exploding or disintegrating: SpaceX’s Starship, the largest rocket ever built. Originally envisioned as a Mars rocket, it is also a key piece of NASA’s plan to outpace China and land humans on the lunar surface for the first time in more than half a century.
------
Just goes to show that resolving this is a very serious issue, despite all the chaos released upon NASA by that crowd in DC.
GW
Found this in today's AIAA "Daily Launch" email newsletter:
----------
article image
Aviation Week Network
Space Ops: Blue Origin Prepares To Leapfrog SpaceX To The Moon
Blue Origin plans to attempt a lunar landing this year. If successful, the MK1, which is 26 ft. tall and 10 ft. in diameter, would become the largest vehicle to touch down on the surface of the Moon, eclipsing the Apollo program’s Lunar Modules (LM) that landed crews six times in 1969-72. Fully fueled, the MK1 weighs 47,000 lb., compared with the 36,200-lb. mass of the extended Apollo LM.
-----------
We will soon see if it works. Between the larger size, and the hydrogen fuel, that is how the Blue Origin lander can do the landing single stage, even from that ridiculous Gateway orbit with the one-way dV that is at least factor 1.5 larger than Apollo's.
Meanwhile, SpaceX is getting into trouble with its basic design approaches, and I'm unsure they actually realize it. Had they had attitude thrusters independent of propellant tank pressurization, Flight 9 might have been under control at reentry! Simple as that! Between Flights 7 and 8, that answer has been staring them in the face.
GW
Getting back to the proper topic here, what is needed is vehicles assembled in circular LEO at a space station facility that does both assembly work and propellant fill/refill work. You do it in circular LEO at fairly low inclination to make it reachable from the surface at the least launch dV.
There needs to be a space tug or multiple tugs based at that same facility. The tugs take the interplanetary craft to just below escape, so that a much smaller departure dV is demanded of the interplanetary craft, while the tug itself stays in an extended ellipse, and returns unladen to the station in LEO for reuse. That leaves more dV available for the interplanetary craft to enter low Mars orbit, should that be desired. Direct landings require less dV, but are not reusable. The direct landing dV is not as much less than the enter-LMO dV as most people think. Why? Aerobraking cannot do it all!
BTW, the same basic idea also works for lunar missions.
GW
The NASA science mission portion of what NASA does has been gutted. They will not even complete the missions already out there.
The NASA manned mission stuff is going to be greatly changed. Trump wants flags-and-footprints on the moon and on Mars, but nothing else. Once he gets that, no more NASA anything. Look to see SLS and Orion killed after the Srtemis-2 flight around the moon. And Gateway is toast, too.
That future loss of NASA contracts is really where the new baby rift between Musk and Trump is coming from. Musk's NASA contracts will dry up when that happens. DOD will not make up the difference, not with the likes of that incompetent Hegseth in charge of it. Musk knows that.
It's already just barely started over DOGE, with Musk essentially complaining that Trump's "big beautiful bill" does not kill enough Americans for lack of cuts to health care, social security, etc.
GW
Heard another Raptor test today. It still has some sort of pressure/thrust oscillation going on, at about 10 Hz.
They fired it just before the bad thunderstorms barely missed McGregor. Dodged the bullet, we did.
GW
I agree wholeheartedly!
GW
The idea behind high AOA during booster entry is that high AOA flight is higher drag that acts to decelerate you. That reduces the propellant quantity required for the landing burn. The problem is that high AOA is a high crossflow velocity with larger wind loads perpendicular to the stage axis. Every supersonic airframe I ever heard of would break up under those crossflow air loads in supersonic flight, if there was more than trivial ambient atmospheric density. Higher AOA is higher breakup risk. Simple as that.
GW
As near as I could tell from the data readouts on the screen (as viewed on SpaceX's site), staging occurred at 60 km altitude, 4700 km/hr (1.31 km/s, which seems low to me), and somewhere near a 40 degree upward path angle (which seems high, so I'm unsure that indicator on the screen really means anything). The directional forces from the specially-shaped hot stage ring seemed to do that flip force job very well indeed. I'm unsure, but I think a few booster engines were lit to settle the propellants.
The failure after the high AOA entry looked to me to be a failed engine ignition. I saw a lot of fire in the engine bay, which did not really spike into the expected engine plumes, and then dimmed just as the view was lost. If I had to guess, I'd say the 45-deg AOA airloads cracked something, venting propellants, so that there was nothing the engine turbopumps could acquire to get proper ignitions. The fires in the bay likely melted open the bottom of the adjacent tank, breaking up the stage.
Early after staging, I could see propellant venting from some sort of opening on the leeside skin of the vehicle. This was right after staging, and I'd almost bet that was not supposed to be happening. The view in the payload bay showed entirely too many little white flakes of something floating around. I was disturbed to see that, it should not have been happening either, certainly not in such numbers with such persistence over time. Unsure what they were, but the odds favor frost flakes. Which points to propellant leaking into the cargo bay somehow.
I do not know what means are used to open the cargo bay door, but if it was propellant evaporation gas pressure, that venting since staging would explain why the door failed to open. That same propellant evaporation gas pressure is what powers the only attitude control thrusters they have, which are cold gas. I thought I saw the vehicle spinning since about the time the door failed to open.
I think Kbd512 is exactly right: they need attitude control that is independent of tank pressurization. They have something on board that independently powers the flaps, I don't understand why that doesn't also power the cargo door. But it cannot do attitude control, you need some sort of thrusters for that.
If I were Shotwell, I'd tell the team to put the Dracos from the cargo Dragon on Flight 10. It would be worth it to recover the vehicle; they'd learn a lot from actual hardware in their hands that telemetry alone simply cannot tell.
As for the plumbing leaks, I don't think they have tried to fix the evident pressure oscillations in the Raptor power heads. They already found some plumbing somewhere that is resonant at that frequency, and have beefed up the things that failed to take the wild oscillation pressure loads. But that is exactly like replacing radiator hoses only one at a time. Fix one, and the next weakest one breaks. After a while, you replace them all. And that's the same pattern as what we have been seeing.
GW
The ULA settling design is basically spin the tank, not attached to anything else resembling a space station. Which means you dock with the tank, and spin both your vehicle and the tank up, rifle-bullet style, as a single unit. I was pleased to see they make the spin settling work at 10^-3 to 10^-4 gees. (That means my vanes inside the tank do not need to spin very fast at all, maybe near or even slower than 0.1 rpm.)
The sunshield part is basically a thermal radiation equilibrium problem. If you are in orbit about the Earth on the day side, you are receiving 1300 W/sq.m solar radiation from half the sky, and you are receiving thermal radiation from the Earth itself at approximately 300 K. Both are too warm to tolerate, even with LOX. Much less LH2. On the night side of low Earth orbit, half the sky you see is at 4 K which is OK, but the other half is still the Earth at 300 K, which is too warm. That's why you need the sunshade, even on the night side.
You need something to protect the tank before you get the sunshade erected, and if the shade itself should get warm, you still need to protect the tank from thermal radiation coming from the sunshield. THAT is why I recommend wrapping the tank with a layer of high R-value fiberglass insulation, and topping it with a very reflective layer of polished aluminum foil.
The sunshade itself can be a simple cone-shaped parasol-like structure. It will work more effectively if both its outer surface seeing sunlight and Earth radiation, and its inner surface facing the reflective tank, are both very highly reflective. I've been through this before, designing shielded thermocouple probes for measuring total air temperature in very high speed ramjet inlets. They only worked when gold plated. Everything else that would be initially reflective would oxidize "dark" when exposed to air that hot (over 1200 F). Oxidation would not be an issue for a sunshade in space. Aluminum foil or similar would work fine. Even highly-aluminized Mylar would work. Just put the extension rods on the shady side, so that they do not get warm.
GW
They have the approval, even though the flight 8 inquiry is still open. The warning area is about twice the size that was in place for flight 8. This one is a 3rd test of the enlarged Starship upper stage, with a larger propellant load capability.
In the static tests leading up to this at the McGregor site, I heard no changes to the pressure waves coming from Raptor 2 tests, that were a periodic noise signal at about 10 Hertz, hidden in the random combustion noise. I thought I heard a Raptor 3 test just the other day, but it had the same thrust oscillations hidden in its noise, at about the same 10 Hertz.
It would be almost impossible to find and diagnose thrust oscillations in digitally-acquired data traces. Unless they acquire (and learn how to use) some analog recording devices of about 1 MHz response, they're not likely to figure this out.
But, we'll soon see if this one has engine-out troubles like the last two. I haven't seen troubles like these since the Saturn-V. Which was before everything went digital.
GW
Looking at this from a viewpoint of flight mechanics and aeroheating:
With TSTO you have a booster stage that supplies a smaller fraction of the total dV, usually at a lower Isp due to the limitations of nozzle design and of propellant selection (primarily density and volume).
You have an upper stage that supplies a significant majority of the total dV, and usually at the higher Isp associated with vacuum nozzle design and higher-Isp propellant selection, with the density-volume problem far less demanding in a smaller upper stage.
If flown to a landing somewhere, because of the low staging velocity that is typical, the booster hits atmosphere at a speed nearer 1 km/s than 2, which would be supersonic near Mach 3 instead of hypersonic near Mach 6. The aeroheating is far lower the slower you are moving when you hit atmosphere. Depending upon what the stage is constructed of, you may or may not need an "entry burn" to slow down to a survivable entry speed, as a bare metal item with no added heat protection generally. There are some heating issues locally around the engine bay, due to hot engine plume gases plus the hot air, there. And it usually needs some sort of landing legs, which add at least a bit to its loaded-stage inert weight fraction.
The upper stage, if it is to be recovered, will hit atmosphere at speeds that are full orbital class. It takes very little dV to deorbit (on the order of 100 m/s from circular LEO). It would take nearly the same 6 km/s dV as the second stage supplied to ascend, to hit atmosphere at the same slow speed as the booster. No one could ever afford that much propellant! So the second stage MUST BE a fully qualified reentry vehicle, if it is to be recovered.
There are essentially 2 ways to design the upper stage of a TSTO. It can be just a standard rocket stage, with the payload housed in a separate payload fairing that gets jettisoned as you go exo-atmospheric, at or just after staging. Or it can be a vehicle that actually contains the payload inside, so that there is no payload shroud to jettison.
If you design it as a stage with a payload shroud, you are going to have to add a fully-qualified heat shield to its forward end, and enter like a blunt object, nose forward. That can be done, but it is heavy, and there is no way to do that and still achieve a 5% loaded stage inert fraction! It'll be nearer 10, maybe 15%. The heavy heat shield complicates the final landing of the stage, unless you jettison it after entry, which creates a serious falling debris hazard. And because the sides have no tumble-home angle, and attitude control is not perfect, those sides will also likely need at least some heat protection. That also adds to loaded stage inert weight fraction.
If you instead design it as a vehicle that contains the payload inside, where the vehicle structure serves as the "payload shroud", that frees up many possible choices for its shape, and for exactly what attitudes you can use during entry, in turn freeing up more choices for what you can use for heat shielding protection, and for how much of the vehicle needs heavy-duty shielding, and how much does not. But, at entry attitudes that are not nose-first, the wind loads on the structure during entry are much higher and more difficult to design for, compared to simple blunt-shape nose-first entry. Such designs will have loaded stage inert weight fractions in the 15-20+% range, or more (it's 40-50% in jet aircraft, and they still break up if they go broadside at full speed).
If the upper stage really is an entry-qualified vehicle design, you still have two classes of options for its landing: wings or vertical. Wings add yet more inert weight, for the strength to survive entry wind loads at entry attitude, and also for the area needing heat shielding. They also have a long runway requirement. But you can delete landing propellant, if you give up on go-around capability at landing. Or you can land vertically, which means you must have some means to go tail first at lower altitude, and you must have significant landing propellant aboard. Generally, you also need landing legs, too.
Reusability is simply going to increase your stage inert weights, especially the upper stage in a TSTO, there is no way around that ugly little fact of life. You get what you pay for.
Kbd512:
There's no fundamental reason why you couldn't feed with from multiple turbopump assemblies. There might be a practical reason not to do that, though.
For the same flow velocities, flow rate should be proportional to cross section area, which suggests flow rate scales as pump dimension squared. Cutting down to half the flow in each pump would have those two assemblies each about 70% as large in dimension as the original single unit. That makes the packaging of the turbopumps about the engine chamber more crowded.
Since mass scales as dimension cubed, that option would be heavier, too.
I knew there was an extendible bell at least tested on some engine. From what you said, it was a variant of the RL-10. Did that ever fly on anything?
GW
The trouble with mixing two engine types in an SSTO is the very same problem SpaceX's "Starship" would have if one tried to make it an SSTO: you are too short on thrust to begin with, limited by what will fit behind the stage. Adding a second engine type just makes that even worse! You have to have enough thrust to launch "smartly", which is takeoff thrust/weight near 1.5 or more (never less), or your gravity is loss increases dramatically. Why? You end up burning the vast majority of your propellant in only the first ~5 km off the pad, leaving you still moving very subsonic!
Before the recent design updates, "Starship" was in the neighborhood of 120 tons inert and 1200 tons propellant if fully filled. At ZERO payload, that's a launch weight of 1320 tons. It is normally configured with 6 Raptor engines, 3 sea level in the neighborhood of 250-300 tons sea level thrust (call it 275), and 3 vacuum engines which would be nearer 150-200 tons thrust (call it 175), if unseparated at sea level, which strongly limits exit bell expansion ratio.
By the way, raising vacuum expansion ratio with constant chamber and flow rate makes the exit area bigger, so that fewer engines like that, will actually fit behind the stage! The takeoff thrust problem really gets unsolvable very quickly, if you attempt what might otherwise seem obvious!
Thrust is mdot*Vexit + (Pe - Pa)Ae, where the pressure term is quite strongly negative at sea level for an unseparated vacuum engine. (Thrust would be almost zero with a separated bell, which would further destroy itself in a single handful of seconds.)
Separation-limited vacuum engines (like the current vacuum Raptor) inherently have utterly-lousy sea level thrust! There is simply no way around that! 3x275 + 3x175 = about 1350 tons with all 6 burning at sea level on "Starship". That's thrust/weight only 1.02 at liftoff, which is long known to correspond to gravity losses WAY TO HELL-AND-GONE ABOVE 20% (or more) of LEO speed, not the 5% of an efficient system. Add only 30 tons of payload to this example, and this thing CANNOT budge a single inch off the launch pad, no matter how much propellant it has!
And there is NO ROOM behind it for more engines! Making the tankage hold 1300 or even 1400 tons really does not change that picture very much at all.
All SSTO designs face exactly the same thrust problem as trying to make an SSTO out of "Starship"! You cannot have any more engines, because those added would lie outside the stage diameter! That doubles-or-more your drag, and way-more-than-doubles your drag loss, which with a really clean shape of the right L/D ratio is about 5% of LEO speed.
There is simply way-far-more to this entire question than just Isp and mass ratio in the rocket equation! I have long tried to communicate that, but unsuccessfully!
And by the way, if sea level thrust gets reduced by the backpressure term, so does the corresponding sea level Isp, for the same combustion chamber design and total propellant flow rate. Which is EXACTLY why you need to look at engine/nozzle ballistics, and not just pull Isp's out of some table in some reference.
I have provided the spreadsheet tools and the instructional lessons, for free, to be able to do this work correctly. That's the stuff accessed by links posted right here on these forums.
GW
All the hydrazines are similar in toxicity, and very similar to the toxicity of anhydrous ammonia, such as what farmers put on their fields. Even the decontamination is similar: heavy dilution with water, while wearing protective gear to keep the alkaline corrosive off your skin, and while breathing with a self-contained air or oxygen supply.
It's the NTO oxidizer used with any of the hydrazines that is truly dangerous. The toxicity rating is far higher, because even a whiff of it causes lung damage that cannot be treated. You will die from it, just slowly. That requires a fully-sealed suit that is essentially a pressure suit, with a self-contained air or oxygen supply. Ordinary protective clothing is not adequate. You must have the sealed suit.
You need the oxidizer because monopropellant Isp with hydrazine decomposition is far lower Isp. It's down near 200 sec, when hydrazine-NTO Isp is up near 300-330 sec.
GW
Those other problems might not have toppled it over, if it had had the wider feet. This thing was very tall and narrow. That makes it susceptible to toppling if anything, anything at all, goes slightly wrong.
Intuitive Machines is not the only outfit failing to face up to that risk. SpaceX has exactly the same problem with its "Starship".
GW
In the Mil Standard tests for tactical weapons and equipment, there is a fragment impact test, which investigates what happens if your item is struck by shrapnel from another explosion.
It is specified as metal fragments impacting at some 8000 ft/sec (which is 2.44 km/s). Such would be typical of a solid propellant motor that explodes for whatever reason, or a warhead exploding.
We're talking about hard metal fragments leaving the scene of the crime at about Mach 8! This stuff is not just supersonic, it is in fact hypersonic.
GW
I'm not so sure. I think the time scale of your look-back is involved.
Meanwhile, recorded history says there are only 10 previous 100+ F days recorded for Waco, Texas. That record goes back more than 150 years. Today may be the 11th. So, yes, at least this year, things are bit more extreme than usual.
There is an old saying: "climate is what you expect, weather is what you get". That is a way of saying day-to-day weather variability is a seemingly-random noise "hash" superposed upon the longer-term climate trend line, a "hash" of high frequencies and large amplitude.
When you look back at previous centuries, the effects of the "hash" and the trend line, become jumbled together and confused. The longer the time scale of the look-back, the worse that effect becomes. It's really hard to tease apart.
Especially when there are also small changes in the trend line, like "the little ice age" vs "the medieval warm period", or very large changes in the trend line, like "deglaciated" vs "glaciated" during the ice ages.
GW
What I did was show that a single stage vehicle of very modest payload fraction (near 5%) could indeed make the round trip between the Martian surface, and low Mars orbit, with any of the common propellants that we know and use. And that includes the storables: NTO and any of the hydrazines.
The higher your Isp, the larger your payload fraction can be, but single stage it is still modest even with LOX-LH2. It's just very small with the storables. Over 5% with LOX-LH2, under 5% with storables.
The only reason I did this as a single stage item was to look at reusability/long service life. If you stage the lander, you cannot re-use the first stage because you jettison it. And on a wild and uninhabited planet, recovering such a jettisoned stage is unlikely in the extreme. But, the overall payload fraction of a two-stage vehicle is very much larger. It's a tradeoff: bigger payloads vs one-shot throwaway first stages.
Whether based on the surface, or based from orbit, this single stage lander is possible. Surface based, the numbers are a bit different from orbit-based, but they are still comparable. However, things sent to low orbit around Mars inherently involve a higher dV for the trip from (and back to) Earth, because of what's needed to decelerate into (and accelerate back out of) low Mars orbit, not to mention rendezvous budgets. Direct landings do not have that dV requirement, because hypersonic atmospheric drag deceleration can do the majority of that job, even in Mars's thin atmosphere. That's part of another tradeoff you have to make: reusable orbital transports plus reusable landers and their logistics, vs throwaway one-shot direct landers.
Although, at higher ballistic coefficient, you are looking at a powered landing of much larger dV than you might otherwise think, so the direct-landing dV advantage is nowhere near as attractive as most people seem to think it is!
That advantage almost disappears when you are too large and come out of hypersonics too low to use parachutes! Smaller unmanned probes (at or under 100 kg/sq.m ballistic coefficient) can use chutes, but manned vehicles or items with significant payload mass (300-1000+ kg/sq.m) cannot use chutes, because they come of hypersonics at or under 5 km altitudes, while the small probes come out at or above 20 km altitudes.
If for the sake of illustrative argument you assume end of hypersonics at 0.7 km/s on a straight path angled 45 degrees down, then at 5 km altitude, you are only 10 sec from an un-decelerated impact. If instead you come out of hypersonics at 20 km, same speed and angle, you are almost some 50 sec from an un-decelerated impact. 0.7 km/s is still too fast to deploy a chute, and it takes about 5 sec to deploy one. And it still takes more time (at least 20-30 sec) to slow you down further by any significant amount! See the difference altitude makes?
Even SpaceX's "Starship" must face this issue: I really do not believe they can execute the near-surface pull-up deceleration to subsonic without rocket thrust. The reliable lift is just not there. The extra landing propellant is going to displace payload, even if they make the vehicle work as an Earth orbital transport, which they still have yet to do. Plus, they still have to make tanker refueling work! And, in my humble opinion, very few of these tall narrow vehicles will survive touching down on a rough landing surface that is also soft. SpaceX has never made a dirt landing yet, even with its Falcons.
I'm not saying that this can't be done! I'm saying that doing it "right" is a very difficult thing. And getting those answers is way-to-hell-and-gone more complicated than just running a rocket equation calculation or two.
GW
I guess I do not understand clearly what the words refer to.
Never having used the forum's search tool, I cannot say, but I would think it should locate everything a user posted, if you search for his username.
Does it not do that?
GW
Welcome to the enhanced extremes of weather due to global warming. It will only get worse as the years go by.
Even down here in central Texas, we usually do not get mid to high 90's F until the end of May. But it is supposed to be near 105 F here tomorrow (that's about 40-41 C). That kind of event is not missing from all records for mid-May, but it is unusual.
I've also noticed the weather forecasts are not as accurate as they used to be. The computer database the forecast models are based upon do not apply as well, anymore: because the climate is shifting.
GW
I know there has been an engine with a two-piece bell developed and tested. I am unaware whether it ever flew an operational mission.
The sketch shown in the previous post is there to indicate the problems that the makers of that engine faced, and anybody else who attempts this will have to face. They are serious problems, and the solutions are not trivial.
In supersonic compressible flow, the driving temperature for heat transfer is the recovery temperature, which in a rocket engine, is only somewhat lower than the chamber temperature, throughout the engine and its bell. The heat transfer coefficient varies quite strongly through the nozzle, peaking near the throat, and reducing sharply through the bell toward the exit. Which is why with a vacuum engine that has a long bell at high area ratio, a significant portion of the exit bell does not need regenerative cooling, as long as it can glow incandescently to cool by re-radiation.
While the pressure down the bell is quite low compared to chamber pressure, it is not zero, while that outside is, or is very nearly, zero. That pressure difference across the wall of the bell is where the risk of a leak comes into play, at the joint where the bell extension attaches. That is likely pretty near the location beyond which you do not need regenerative cooling to survive, so as to avoid very serious flexible plumbing issues with cryogenics. Thus inherently, you are faced with installing a reliable seal that has to operate while very hot.
Be aware that any flow captured in any way from that supersonic bell flow reverts to a temperature pretty much identical to chamber temperature, once decelerated to a stop, or subsonic, relative to the bell wall. That is the kind of gas you face within the spaces or passages of any sort of seal design.
I think the sum of all that explains why this kind of extendible bell technology does not already dominate the launch business. There has to be good reasons for that absence, and in my opinion it is (1) much greater design difficulty (which costs), and (2) very difficult-to-achieve reliability (with a very high cost of failure).
GW
There is no possible way to convert water to hydrogen and oxygen real time during a vehicle rocket burn!
The electrolysis energy efficiency is 10% or less, likely substantially less, as you attempt to speed up that process. The energy efficiency of combustion heat energy release is within a percent or three of 100%, during any chemical rocket engine burn. That's a ratio of energy efficiencies of at least 10, perhaps 100.
You will have to split the water into hydrogen and oxygen long before you ever attempt to use either the oxygen or the hydrogen in any sort of propulsion burn. That being the case, what is the point of the paraffin that needs the oxygen? Especially if you have to split the water into hydrogen and oxygen, long before you ever depart on your mission? We already know the performance of hydrogen burning with oxygen is far higher than any hydrocarbon burning with oxygen.
Sorry, I have to ask these very pertinent questions.
GW
Those average ages of SpaceX engineers matches up with what I found out several years ago when I contacted them looking to see if they needed somebody well-versed and experienced. They did not want anything to do with me. I was told they hired no one over about age 40 to 45, because folks that old or older could not withstand the chronic, year-round 70-80 hour weeks they demanded.
As near as I can tell, that practice and policy continues to this day. Well-paid for overtime or not, that's actually employee abuse when it is chronic for all 52 weeks of the year. I see a lot of SpaceX employees at lunch time in McGregor. I have never seen a one who looked to be much over 30. And that high workload is exactly why.
I'm about to turn 75. There is zero chance they would ever hire me, even as a part-time or occasional consultant. The ageism is simply too strongly built into their corporate culture. Musk did that, and forces Shotwell to continue it.
Without any experienced types, and with the strong "not invented here" culture that they have, there is no one on the staff able to help them not repeat the mistakes of others, past or present. That is exactly why they (1) seem to be running into problems others have already addressed, and (2) must use the "build it, break it, build another" shotgun approach to development. As long as the troubles they run into are easy to spot and understand, that approach works, as we have seen. But it fails when you run into something subtle. And I think that's where they are with Raptor.
They still do not recognize and understand what's wrong. I just yesterday heard/endured what I believe was a Raptor-3 test on the tower stand where it is really loud, and the pressure waves created by thrust oscillations can reach bystanders less attenuated. It gets magnified a bit more under a cloud deck, which reflects rising wavefronts back toward the surface at remoter distances. This was a 6 minute burn, featuring what sounded like the "explosion" of a near full power start, and several throttle-downs and throttle-ups.
It shook the ever-loving shit out of my house, stronger than any tests seen before, with about a 10 hertz pressure wave oscillation, probably induced by a 10 hertz thrust oscillation whose amplitude is still barely hidden within a general "hash" noise amplitude of about 3-5% of pressure and thrust levels. You simply won't find this problem with digital data acquisition, because of the digital pixellation effect wiping out the signal. You see it with analog equipment of 1 MHz response. Been there and done that. For many years.
And being a young outfit, I'd bet real money they have no such analog equipment, and nobody on their staff who would have a clue how to use it, especially when searching down in the noise for a hidden signal. Everything is digital-only with the young.
Until they stumble on this answer, or see it posted here or somewhere else (and believe it!), they will continue to have engine-outs, fires, and explosions with Raptor. It may require a change to the geometry of the propellant injection devices in the final combustion chamber. That's what mitigated the initially-fatal thrust oscillations in the Saturn-5 F-1 engines, down to a level that could be dealt with most of the time.
GW
Analog works best and easiest, as I already described. 1 MHz FM tape recorder response is plenty good enough to capture even those instabilities at multiple 10's of thousands of Hz.
I know a lot about this, because I used to do it for a living (among many things) at the solid rocket shop decades ago. Back then, we would never let a design with a noticeable thrust oscillation out the door as a product. THAT was the lesson of the POGO difficulties with Saturn and Titan stages in the 1960's! And many others before that.
I'm sorry to say that those ethics and that kind of capability are long gone now.
GW