jenakuns

[jenakuns]

SpaceX's good week

SpaceX has a lot of good weeks nowadays, but even by those standards this week was pretty radical.

To start off with, IFT-5. Yeah. I'm not sad that I slept through the catch. Not at all. Holy shit it happened what the fuck. This is some bullshit Musk design that actually got to pad and worked what? Aspiring to have no flame diverter my ass. In general I'm pretty happy with the launch. The things that were supposed to happen happened. There are things to improve, but that's what makes following dev of Starship so much fun. SLS is like, yeah we need ~3 year to refurb our pad after that first launch because an elevator door got knocked and different scope. Meanwhile SpaceX pulls this fucking shit and has launched 4 more times in the past 1.5 years and done a catch.

Then Europa Clipper launched on Falcon Heavy. This is SpaceX most important/expensive/performance driving scientific launch yet. The adventure of selecting Europa Clipper's launch vehicle is a journey in it's own right. Very cool and awesome stuff there and definitely not nerve racking giving the second stage issue two launches before.

3 Starlink launches for prosperity. Starlink is borderline misunderstood? People obviously talk about it a lot. What is so good about it is that it's a payload with a relatively steady $/kg in regards to revenue. You can only launch so many Earth observation sats you can launch before you run into diminishing returns, but satellite broadband is a lot more scalable. There are years in which ULA launched less than SpaceX did in this week.

Also SpaceX won $733M worth of Space Force launch contracts for 9 launches. So much for Lane 1 diversifying providers lol. It's not unsurprising per say, it's only between SpaceX and ULA and VC0 is a bad config (0 contracted launches btw).

This is just a week for SpaceX. 5 launches and winning $733M in contracts.

[jenakuns]

Stoke Space Second Stage Solutions So Sick

So Stoke is pretty popular ay. I remember there was this fun time in like 2022 on discord where about once every month somebody would discover Stoke for the first time and it would be this fun little conversation. That boat has well and truly sailed.

Now onto this trade space. I'll begin with engine ISP as that's the easiest to understand. Now despite what this meme and Stoke might have you believe, the engine isn't that efficient in a vacuum at 430s of ISP. Conventional vac optimised expander bleeds can hit 450s, so they're down relative to that. And against a hypothetical FRSC hydrogen upper stage engine which could hit 470s of ISP (and they are capable of building); they're losing 40s. 430s is what open sea level engines achieve.* Essentially, if it was methalox, the ISP would be ~345s, below that of staged sea level engine like Raptor. This is their primary performance loss in the trade space. So what do they gain for that? They get to integrate 3 capabilities into the one engine, effectively decreasing the system complexity and dry mass of the stage.

*Although worth noting that those sea levels need ~100 bar in the chamber to achieve that ISP, Stoke does it with 40 bar. 40 bar means that they get to do things like skip bi metallic jackets on the chambers and can instead just have a copper alloy print. This combines with the 'positively benign' environment that a bleed expander turbine experiences so their engine doesn't experience a lot of stress; it should have a really good lifespan and be easy to reuse.

So it's kinda more of a sea level engine that gets reasonable performance in a vac for how robust and versatile the engine is. Something that restricts the performance of Starship, especially for HLS, is the fact that they need those sea levels in the skirt. Because they can't fit a gimballing Vactor in there, it means they have to deal with the ISP drop that comes with having to use 350s engines alongside your 375-380s vac engines. (not to mention throttling problems).

As a landing engine it’s fairly nice as it’s much easier to deeply throttle the architecture because you don’t have to worry about combustion instability in the pre-burners/gas generators, nor in the combustion chambers themselves as they can just limit the number of chambers that receive propellant. I don’t know what the theoretical limit is, but the renders do show 10% throttling.

Actively cooled heatshields are to ablative heatshields as liquid rocket engines are to solid rocket motors. (don't think about passive tile solutions, they don't fit in the analogy). This is something that people doubt the technical viability of. However, what I think people need to realize is that propulsion engineers already have to deal with thermal fluxes of ~100 MW/m² on the combustion chamber walls in their engines. The ~10 MW/m² of LEO re-entry is actually a breath of fresh air for them. In this case, the Stoke’s ‘aeroplug’ won’t experience that high of a heating due to the relatively low thrust/area of the engine, so re-entry thermal flux will likely be higher in this case. But the point still stands, it’s well within the propulsion engineers experience; get out of here TPS engineers! (just kidding we still need you for the material science of optimizing the alloy for the heat shield).

What does an actively cooled heat shield achieve? It’s no secret that the heat shield is one of the weaker elements of Starship’s design currently, with 18,000 tiles and a secondary ablative heat shield taking a significant portion of the mass and maintenance budget. It’s more robust than I would’ve thought; well at least the vehicle is. However for a fully and immediately reusable orbital heat shield, there’s definitely a lot to be improved upon. Active cooling has a much better shot of getting there, because effectively we’ve already gotten there previously with thrust chambers on reusable rocket engines. Monitoring the heat shield for damage should be easier, given that it’s inherently easier to place sensors and watch the flow of propellant compared to analysing 18,000 tiles. I’m not sure it would be easier to repair per say, but if it is permanently damaged, you could just treat the heat shield as a Line Replaceable Unit (LRU) and slot another one in.

Now comparing Starship flaps to Stokes lifting body heat shield is interesting, because the flaps do provide greater control. The mitigating factor here is that Nova doesn't need the greater control. Because SpaceX has a tile solution and enters on the broad side in an passively unstable state, they need the flaps to provide that flight control during re-entry and landing. Nova is passively stable with its base heatshield, so with the offset to provide a degree of lift they have enough trajectory control.

Chines are just in there for the meme of it (storage space). If they do get rid of the chines on Super Heavy, I’ll definitely miss them. They add so much to the look of that vehicle.

So what about hydrogen? To get something out of the way, I'm tired of people treating hydrolox as a curse on price. Yes the hardware (valves/seals) to support it does cost a % more, but it's not like you're forced into hundred million $ stages as a result. It's like going, 'oh well Shuttle proved reusability could never be cost effective, so Falcon 9 reuse will fail.' You're blaming mixed technical trades when the ultimate driver is managerial and organizational culture.

Hydrogen is fairly important to this architecture. A likely significant reason why methane wasn’t competitive with passive tile solutions for Starship was that by mass, it requires 4.5x as much methane to extract the same amount of heat as hydrogen. Now volumetrically this is less than the hydrogen, but mass is the driver at this stage of the mission. 

I will say that VTVL first stages prefer high mass second stages as staging earlier decreases delta v + thermal flux to get to the landing site. In Falcon 9 case; replacing the current stage with a hydrogen stage of the same proportions would decrease performance to LEO by like ~3 tons for the reusable missions. This is why the higher impulse density propellants like methalox/kerolox are nice in this case and why Starship in particular.

From a long term perspective, hydrolox is better for the Moon, methalox is better for Mars. I will say that in order to produce methalox via ISRU on Mars, you can produce hydrolox. However impulse density and storage conditions make it preferable for methalox on the red planet. 

From a Mars and a LEO perspective, the Starship architecture likely does provide better performance. Given that those are the 2 biggest theoretical market's for SpaceX, it makes sense that they optimized around that space. However the presumably more effective reuse and performance in the cislunar space, likely give Stoke's design the edge there. (Obviously there's a degree of separation in that one is a 3 ton to LEO vehicle and the other is a 100 ton to LEO vehicle)

[jenakuns]

Posts like these are interesting to me. This ability to put together a small collection incorrectly used facts into a very opinionated piece and then it does really well. It reminds me of those tiktoks that talk about, well anything really.

Dunning krueger is obviously overused and even by using that term, in turn I am operating under the sphere of it. Idk, I personally like to imagine it as a graph, where you have a graph relating 2 variables.

Starting out in any topic area, people's knowledge of a subject looks something like this. Pretty rad right?

When you learn a couple facts or concepts, suddenly you're able to put together a trend. Bam you can form opinions and make tumblr posts about them. You don't even have to understand the data points.

The actual data scenario looks something like this. Getting a trend from it is cursed and requires basically ignoring segments under assumptions and rules.

NASA engineers considered the failure rate of some critical shuttle parts to be about 1 in 100

Appears to be a reference to the figure in this article (which if it is, would make it a misquote on a couple layers), although I really have no clue, half of what Shuttle engineers did was risk analysis (there's a lot of figures). Honestly it was a lot worse than that, those first missions were like a 1/10 on a LOC/V happening, which is utterly nuts.

Now what is the failure of Shuttle, given that we appear to be discussing this at length. Umm. There's a lot of context that I really can't bring to bear. Early Shuttle is easier I guess. It's a mixture of "operational vehicle" that didn't allow things to be fixed, budget restraints and programmatic launch fever.

Compromises occur in reality and failures will generally happen in those spaces, but that doesn't mean the compromise is at fault. It's just something you have to handle and manage. But anyways; this is history from 30 years ago; it's pertinence to NASA in 2024 doesn't seem that clear, given the learning in between those periods.

Do you know that NASA engineers currently have no idea how many rocket launches the next mission in the Artemis program (in 2 years!)

It's not no idea; it's a rough idea. Starship has a lot of engineering parameter's and challenge's to deal with and it's performance is a little bit varied. But ~18 launches is a number to b

The schedule has been a joke since before the lander was selected; no it's not 2026 or even like 2028; 2030 maybe. I wouldn't be worried about them flying before they're ready in regards to humans, so they'll push schedule way before they'll compromise on safety. This is a big difference from Shuttle era.

has never been attempted before on any spacecraft

How dare NASA fund efforts to attempt something that has never been done before. They should only stick to things that have been done. That's how we have always made progress as a society and in regards to technology.

because said vehicle does not actually exist at time of writing?

Yes and

no.

Obviously the vehicle which is capable of supporting and actually being the lunar lander isn't here yet. But there is clear progression to point to. We have another 2 launches in like the next 2 months. People tend to get caught up in the moment and argue about the technical trades of the second and ignore the trend upwards. Oh wow tents good, factories bad; nah bro, but the factory is here now anyways and producing hardware at a good rate.

SpaceX being contracted to deliver a supply of cryogenic fuel to the crewed Orion capsule in orbit

The fuck are you talking about. Ok that's a little harsh, but this is what I mean bro with the dunning. Saying this shows that you have a fundamental lack of knowledge about the architecture and in general the vehicle designs. Orion is launched by the SLS rocket to a lunar orbit (NRHO). It docks with the Starship lunar lander variant, but that is just to transfer crew from the capsule to the lander (and back after the landing). The propellant transfer is between Starship tankers and the lander (and the depot, dw about it), to give the lander full tanks so it can conduct the mission.

I mean Orion can't even be refuelled by Starship because it uses MON/MMH propellants and Starship supplies methalox. There's a really funny thing you/I could bring up, but it won't happen.

every single mechanism fails

It would take 1 to cause the issue as described. And that's not an issue for Starship, single stage baby; but there's a whole bunch of other things so it doesn't really matter.

swing-out crane as the only entry and egress point

This whole hating on cranes being complex and high risk as a technical is something I've always found hilarious. Like bro this is literally riding on rocket science where you have some of the most crazy compact mechanical engineering pushing the limits of material science and it's like; 'nah bro lunar gravity elevators are complicated.' Apparently SpaceX agrees because that's basically the only HLS hardware they've shown lol.

Artemis' proposed lander, on the other hand, is planned to be a vehicle whose design didn't even include heatshields until it was realised it would obviously need heatshields

This is a bizarre, blatantly wrong and an utterly hilarious point because I think I understand how you came to this misunderstanding. See the early SN5-SN11 flights didn't have full heatshields and I think you saw that and assumed that it was the vehicle as intended. But no, it was always intended to have heatshields, just for those early prototypes that wasn't the priority; developing propulsion, avionics and fluid systems was.

And more to the point, the lander doesn't use/need a heatshield . Again showing your lack of understanding.

are ceramic tiles bolted after-the-fact directly through the steel hull

They are bolted to, not through, but the attachment is a weak point right now so you can keep this.

decided to mass-produce the original-design hull sections all at once for all the 'starships' first, before doing any integrated testing.

No. Running out of energy to explain, but like half of the complaints with the program is too much integrated testing.

The Accords came after Artemis; the program does not exist just to shove these new space operation rules down peoples throats to enable commercial exploitation of the Moon. Rather, the ideas set out in the Accords are for enabling/streamlining the ops on and around the Moon for Artemis. If you hate the concept of cities on the Moon this is bad, but I see it as an absolute win.

Artemis Accords not being signed by Russia/China is unsurprising and not the slam dunk people think it is(?); US is funding a war against one and the other one is it's geopolitical rival. Every single other relevant space power has signed it, so it's obviously doing something right. Outer Space Treaty was written in the 1960s, the expectation that legislation from then would still be entirely accurate after 60 years of development and geopolitical change is unreasonable. Utilisation of space resources is important for the future of space and I don't think it's unreasonable to set up principles to allow for it.

Something bad is going to happen, and it's going to happen for the sake of SpaceX and the military-industrial complex at large.

Look I'm not going to say that something bad is not going to happen, because landing on the Moon is hard and very dangerous.

And it's happening because human exploration of Moon good; well that's not convincing, but I've never been good at doing HSF justifications.

Idk, this is kinda just an elaborate tangent for me ay.

Did you know that NASA engineers considered the failure rate of some critical shuttle parts to be about 1 in 100 (significantly greater than what NASA upper-management considered the failure rate to be, and what was considered at all acceptable by the certification process)?

Do you know that NASA engineers currently have no idea how many rocket launches the next mission in the Artemis program (in 2 years!) is meant to involve, because the mission plan relies on SpaceX being contracted to deliver a supply of cryogenic fuel to the crewed Orion (™ Lockheed-Martin) capsule in orbit - a procedure that 1: has never been attempted before on any spacecraft, let alone the Orion™ capsule, not even in uncrewed technology demonstration flights; and 2: would require an as-of-yet unknown number of SpaceX 'Starship' launches, because said vehicle does not actually exist at time of writing?

Did you know they're planning on using this 'starship' as the crewed lander? A design for a lunar ascent vehicle, that is, that does not use hypergolic fuel, that relies on a swing-out crane as the only entry and egress point? During the original moon landings, the LEM had so many redundant methods to make sure it got astronauts off the surface of the moon, that in the most absurd, extreme case, where every single mechanism fails, there's a procedure trained into the astronauts to climb around the outside of the capsule, take a pair of bolt-cutters from the equipment box, physically cut the couplings holding the capsule to the lander stage, and take off to get home. Artemis' proposed lander, on the other hand, is planned to be a vehicle whose design didn't even include heatshields until it was realised it would obviously need heatshields, which are ceramic tiles bolted after-the-fact directly through the steel hull, because SpaceX had decided to mass-produce the original-design hull sections all at once for all the 'starships' first, before doing any integrated testing.

We're seeing the exact attitude that led to the shuttle disasters not being prevented now expressing itself in (and even through) the Artemis program, a project pushed harder and faster through the gates than it should be, by a government (and NASA administration thereby) desperate to advance the eponymous Artemis Accords (that goes unsigned by China, Russia, and much of the world) and reneg on all previous space charters that onsidered ownership, commercial exploitation, and military usage of space forbidden. Something bad is going to happen, and it's going to happen for the sake of SpaceX and the military-industrial complex at large.

[jenakuns]

Starship Flight 2

This is now just another Starship spam channel that's right woo baby all we talk about is this. I mean I am inherently very excited for Starship because it is just a very interesting launch vehicle, although my interests are more broad in space than just that take notes people. Coping I guess. I mean even people who hate the program find it at least very engaging; even more than the things that they like to extent.

Now the question on everyone's minds is how successful was this flight? How does it compare to the previous flight?

Well at a baseline, the pad isn't totalled, so that's nice. The steel plate worked and minimized damage. That side of the infrastructure will presumably no longer be a bottleneck for future flights. Well, ok there will be scrubs and work continues; but presumably it seems like it won't generate month long pauses.

Booster 9 performed it's job as a first stage well; getting to staging with no apparent major issues with it's 33 engines, which would be good enough for most launch vehicles. As a reusable stage however it failed. It appears that relighting one engine failed in the dynamic situation post stage separation for boostback causing a cascade of failures and overall explosion. I'm pretty happy with how it performed; it's not going to inhibit future launches and restrict future flight tests. Falcon 9 first stage recovery took it's time as well; but didn't inhibit Falcon 9 customers as development was happening. The main concern would be that in future, there's not enough Super Heavy's for launches because those are very large and complicated stages with 33 Raptors. And so the associated delays with the recovery effort because of this failure could have longer term schedule impact.

Though recovering early boosters honestly isn't too valuable either. Well ok, it is valuable because you can look at the hardware and see how it performed and some of the most valuable data you can collect in regards launches and reliability. But like; vehicles are inherently going to become less valuable over time given SpaceX rapid iteration and development. Having a vehicle be initially expendable while you figure what is the optimal design points is something I think is reasonable. Even with Falcon 9 Block 5 the later boosters are a fair bit better than the earlier ones.

Starship 25 however less so. A. SpaceX where were the on board views? We were robbed. More to the point though, it failed during the second stage burn which means they have to repeat this flight. Yes you want to test; but you also do want to continuingly expand what your testing if you want a reasonable schedule. If SN10 had of succeeded in landing, SN11 and maybe even SN15 probably wouldn't have flown. Still it did hot stage which was a risk point and burn for a good portion of time; so it would seem like with fixes they would have reasonable odds of going the full way on next launch. To quote Handsome Jack, "that's the sound of progress baby." Tiles fell off but I maintain that tiles aren't important for the reusable launch system while the launch system part isn't working.

Are SpaceX closer to their goals after this flight? Should they be closer? Is this a close to optimal balance of schedule, budget and hardware? Evaluations that I don't really feel like making.

The thing to watch for and baseline as assessment of program execution as I set out in a previous post is turn around time for the next flight.

The current metric to watch for Starship progress will be turn around time from flight 2 to flight 3. Sub 2 months is indicating that the program is going really well, 3-4 months is alright, 5+ months is bad.

I'll adjust it to say 3 months is good, 4 months ok to give myself some margin. High flight rate devalues the impact of individual failures. But again, I don't think it's worthwhile to have to repeat tests instead of spending some extra effort to avoid predictable problems. So if the next flight doesn't reach Hawaii (not successful reentry, but just on a trajectory basis), I would consider that a pretty big failure. Whereas I reasonably happy with this launch. You have to fly hardware some time, but you wouldn't fly B4S20.

[jenakuns]

Alternate crew transportation solutions to NRHO

You have to be very sneaky like when discussing this topic, because if people understand what you're getting at before you arrive at the topic area, they will form arguments. They will feel less obliged to do so if they're already committed to the idea because it got explained to them in fullness. So you kinda have to pretend to be a little old lady and then bam you hit them with the upper cut of the entire architecture. 

The inspiration for this is the recent OIG report and the reaction to this quote.

Although the SLS is the only launch vehicle currently available that meets Artemis mission needs, in the next 3 to 5 years other human-rated commercial alternatives that are lighter, cheaper, and reusable may become available.

Most people I feel dismissed this instantly, because of the Falcon Heavy historical argument, where Orion could be easily launched on Falcon Heavy, SLS bad. In practicality, Orion can't easily launch on Falcon Heavy; it's a very marginal solution. But there is a question here of, well are there now more solutions on the horizon in 2023 than there was in 2018?

I'll start with the Lockheed pill. So what do Lockheed have, or will have in the next five-ish years. They are the prime contractor for Orion, so they have the most interest in having the most effective version of that vehicle in face of any potential competition and the more it can launch, the more money they could make. Currently SLS has a cap of roughly 1 launch/year on it near term, if they do arrive at effective reuse of Orion, presumably they could do better than that. What else do Lockheed have. Well they gunna have a Cislunar Transporter, a zero boil off hydrolox depot, developed as part of the Blue Moon HLS contract. And if the rumours are correct they are interested in buying ULA out from Boeing, which would give them full ownership of the Vulcan rocket. They also have no stakes in the SLS rocket itself, so if they were able to develop their own launch solution, they would stand to make a fair bit more money per launch from it. 

Vulcan has a hydrolox upper stage called Centaur V, which is planned to be implemented with long duration and refuelling capabilities, which in turn with a ZBO depot in LEO would enable it to be refuelled. Vulcan has about 27.2 tons to LEO (Orion weighs like 27 tons) and is designed to be able to be human rated, it's a glove that fits. Technically Orion weighs more than that with the launch escape system, but there will be performance enhancements to the engines (BE-4 Block upgrades and RL10C-X) that should cover that. Diameter wise, Orion is 5 metres wide, Vulcan Centaur is 5.4m, so there's no weird aero covers required.

So we can launch Orion to LEO on Vulcan, what next? Well we send it to the Cislunar Transporter depot and refuel it there. With a single New Glenn launch worth of propellant to the depot, we can then transfer that propellant to Centaur and launch Orion to NRHO (needs 35 tons of hydrolox propellant, min, so should be margin there). 

The good thing about this architecture is A. there are clear incentive structures for the companies involved to do this. B. it leverages the tools available in such a manner that doesn't require them to go widely outside of their existing operations. C. it does so with contractors that a lot of the SLS fans have a reasonable amount of respect for. Like any Starship proposal is going to be met with eye rolls, but a Lockheed/ULA/Blue one that uses their capabilities in reasonable manners is something they're going to find harder to reject.

The other big thing is that it doesn't require a 2 billion/year overhead, because all aspects (Vulcan, New Glenn and Cislunar Transporter) have existing operations to cover their baseline costs. So it could be like a $200M Vulcan launch, a $100M New Glenn launch and the cost of the depot operations. There's an interesting question of whether the depot can return to LEO from Blue Moon refuelling in time for Orion launch ops, but you could also spend the capex (idk like 0.5-1B) on just building another. The development is human rating Vulcan Orion and Centaur refuelling, which is like a billion $, but nothing too unreasonable (+Blue Moon/Cislunar Transporter work will help out with the latter). In addition, it develops the LEO Orion launch and reduces the cost for adding additional Artemis flights in a year (or just makes it possible in the first place). 

So to summarise this concisely, Lockheed could reasonably leverage their own hardware/IP to develop a cislunar Orion launch solution with solid costs and flight rates. They are also somewhat incentivized to do so from an optimising potential $ perspective. The main additional critical risk is a single docking and refuelling op which is not insignificant, but in the scope of what it offers is reasonable.

Ok there my motivation is gone lets get through the others quickly. Blue Origin refuelling Lunar Clipper and launching Orion. Basically the same as the Lockheed one but probably less popular because of Blue's execution to date. And SpaceX Starship with expendable Starship stage and maybe another stage on top to get a TLI payload of some sort that would be Orion capable. People really don't like this because SLS fans are inherently going to be sceptical of Starship. Also SpaceX isn't interested inherently unless you come with $.

There you have 3 potential options for Orion launch.

Anything like a Relativity Space Terran R option is no good; start ups don't have the program of large scale execution to worry about giving crew cislunar transportation to. There's the skip Orion people, use CCrew vehicles or Starship entirely; if you think the SLS people dislike alternate Orion launch solutions; just you wait until they see alternate Orion solutions. I'm sympathetic, requirements for Cislunar ops are a fair bit different from LEO transport requirements. And Starship crew rating is not something to emphasise right now. There's some thinking that you could add a habitation module to Cislunar Transporter and have it act like a taxi between LEO stations and Gateway. It does appeal to me on a thought basis; the main criticism is required propulsive to LEO and in general LOC/human rating problems.

IDK, sometimes Orion feels underwhelming from a capsule point of view; is it even capable of launching 3 times in a year without a large increase in $?

[jenakuns]

High thrust hydrogen fuel rich staged combustion vac engine thoughts

So I just thought I might expand upon my thoughts on a particular engine architecture and discuss the US launch vehicle meta. So I'll start with Ares 1 and the SSME. The original plan with Ares 1 was to use 4 segment SRB first stage and an air lit SSME second stage to launch the Orion crew vehicle into LEO. This maximized shuttle heritage and theoretically made the transition from shuttle to Ares 1 as quick as possible. The problem came with the fact that the SSME was designed to be lit once on ground and then burn like 8 minutes until orbit; changing it to be air lit was hard and the engine was expensive. (handling purging is I think the big issue). So they switched from the staged combustion SSME upper stage to a J-2S gas generator upper stage. Unfortunately, it's lesser performance as an engine cycle meant that they needed they needed a 5 segment SRB first stage, meaning both left shuttle heritage and went into unknown territory as requirements built up. This led to a whole host of performance issues, vibration issues and 10 Gs abort modes.

There's also generically a lot of dislike for Ares 1 just under the principle of why not use Atlas V/Delta IV Heavy variants which are already flying and have an existing cadence. (Half the issue with Ares-1 is that it had no real market beyond NASA, so it's flight rate would've been terrible and had an absurd cost per launch for what it was). Now if you're unhappy with Congress/NASA, you could say that well it's not deriving from existing shuttle contractors (well it is and isn't) and so it's politically dead on arrival. But from an architecture point of view, both vehicles needed a new upper stage because the existing vehicles used lofted trajectories because of the RL10 low thrust upper stage meaning that it produced black zones because of high g's in abort situations. (New upper stage vs an entire launch vehicle is some certain math though)

This gets into a tangent of what the DOD launchers have optimized around (Atlas V, Delta IV, Vulcan). That DOD market is designed around edge performance at high energy orbits like GEO, which means the vehicles are high energy optimized. From an expendable point of view, the estimated cheapest way to do that is SRBs + sustainer core (long burning) + high energy upper stage (which doesn't need much thrust because gravity losses at time of staging are low because the first stage burnt for so long). This leads to the best upper stage engine cycle being closed expander, because generically it's the simplest while maintaining relatively high ISP, with the thrust restrictions not mattering. This is why the RL10 rocket engine has been flying for so long; because DOD has been the main game in town.

But we are now in reuse town and this DOD architecture is bad for that. The SRB's are cheap and kinda small and so aren't worth recovering. But then the first stage which is worth the most stages really late, meaning it has a lot of energy to remove from a very large stage. And it ends up being that there isn't a good way to recover the stage without gimping your performance. You can still do stuff like SMART reuse (I don't like that acronym), but it just ain't as satisfying.

And now LEO mega-constellations are driving the market; which emphasizes early staging reusable first stages and high mass and high thrust second stages. This has meant that the 9-1 engine config Falcon 9 pioneered has become really popular.

What this is all getting at is that for a current suitable upper stage hydrogen engine, you want a high thrust, high ISP solution; which would be enabled by a vac optimized fuel rich staged combustion hydrogen engine. This would be good for the HLV crew launch vehicle, because it would be a high thrust solution. It meshes the LEO optimized/DOD vehicles together, because it enables a high mass upper stage while also doing high ISP for the GEO missions.

And if you had a SHLV, the upper stages of the HLV and SHLV could be common. This would be better than the gas generators historically used because of that ISP dif and size. Don't feel necessary to derive from RS-25, we can do clean sheet engines up in here. That design just isn't pertinent here, set up new paradigms. Of course there are any number of problems (dev cost, existing tech and fitting in), but it makes me happy. Give me my 470 seconds of ISP.

There's also the expander bleed people over there doing their own thing, but we're staying conventional for general applicability.

All of this is essentially saying that Saturn IB and Saturn V are still the kings of optimal NASA based rocket architectures 50 years on. Titan 3 and it's consequences or something.

edit:

Oo and a final point; something you hear time and again from stans/whoever is that hydrolox makes launch vehicles too expensive. This take is driven by the perspective of SpaceX getting by on single propellant combinations for mixed reasons; with then there being derived explanations that hydrogen must be too expensive.

It's like going, oh well Shuttle proved reusability could never be cost effective, so Falcon 9 reuse will fail. You're blaming mixed technical trades when the ultimate driver is managerial and organizational culture. Like do you really think a cheap hydrogen stage is an impossible technical challenge for SpaceX after everything they've done?

[jenakuns]

Misc

(I wanna make 1 point about Starship, but if I just make that the post, I would be limiting my visible self unusually)

It's always funny seeing launches get delayed by months. Well not funny, more very sad. So opposite. But it's like, you see the fully built rocket on pad and it's like; just fill er up and go ay fella. Not the case. Every launch campaign is an epic journey with many trials and tribulation that unfortunately will not see the light of day. The reason I'm currently thinking about this is Firefly Alpha with Victus Nox that had what appeared to be a ready vehicle on pad in March and we're now in August.

The current metric to watch for Starship progress will be turn around time from flight 2 to flight 3. Sub 2 months is indicating that the program is going really well, 3-4 months is alright, 5+ months is bad.

The actual time on repairing the OLM and getting ready for flight 2 I think has caught a fair amount of people off guard. Like the gut reaction to that flight was 'oh no shot; next launch is in 2024.' The fact that it looks like the hardware is reasonably setup for September is like damn. Like there was 69 days between the first 33 (31) engine static fire attempt on Booster 7 and the actual launch. This is the next launch, so I expect the process to go a bit smoother than that.

It's good to say predictions to yourself and provide justification as it can be really hard to judge your headspace in the future. This understanding can then help with current circumstances.

[jenakuns]

Damn niche reddit communities have members that bother to read the source. Shout outs to them.

I still think that this quote has legs, but I just need a more emotional or at least a less technical community. I feel like the right anti-nuclear person on twitter (or well X now I guess) is probably the right space for it. It seems hard to community note. I suppose you would just have to emphasize the overall risk being small.

I remember a person from the thorium documentary that would probably be perfect fit for it. Do I just have to send them a message?

Like come on if this bullshit can do this, this should be able to do something right?

Experiment time with a meme

Ok, so I've got a great factoid and I think it has legs to have a little bit of a run.

NASA underestimated radiation doses of certain accidents for the launch of their nuclear Mars rover by a factor of 3,200 (that's a comma not a period)

It think it's got a real grab and clickbait potential. So I'll post it to an anti-nuclear subreddit and see what happens. Will it spread beyond or just stay there and get a couple hundred upvotes or I am overestimating it and it just dies. If it dies, I'll try it again elsewhere just to see whether it was a timing thing.

But anyway, lets talk about memes, and when I say meme, I'm referring to the older definition, not just the current one of funny image. As I understand it, the older definition considers a meme more of an idea; with information on how it spreads and grabs people and it's position in the minds of people.

See I will post these words in this order and then it will spread as an idea as people share it on their own volition. It's an infection of sorts; mind virus is definitely an underwhelming term as it's been restricted to focus on one specific thing when it could be considered in a more general context.

It is very interesting to think about how much power words have and how much possession of someone's brain you can have despite never meeting them.

Putting NASA in their just gives phrases for some reason a certain staying power. NASA does so much in so many things that it sort of has lost meaning to me to say "NASA does something" given how broad the organization is.

For the record, I am not anti-nuclear; in fact I am generically positive as most space people given the engineering links. I think storage and disaster concerns are overblown, but actual implementation and socially acceptable power concerns under. The actual risks in the NASA report are still fairly insignificant and didn't matter in the end; I fully support the launches of Perseverance, Curiosity and Cassini.

I am just really curious to track how this spreads. (sorry nuclear fans)

[jenakuns]

Experiment time with a meme

Ok, so I've got a great factoid and I think it has legs to have a little bit of a run.

NASA underestimated radiation doses of certain accidents for the launch of their nuclear Mars rover by a factor of 3,200 (that's a comma not a period)

It think it's got a real grab and clickbait potential. So I'll post it to an anti-nuclear subreddit and see what happens. Will it spread beyond or just stay there and get a couple hundred upvotes or I am overestimating it and it just dies. If it dies, I'll try it again elsewhere just to see whether it was a timing thing.

But anyway, lets talk about memes, and when I say meme, I'm referring to the older definition, not just the current one of funny image. As I understand it, the older definition considers a meme more of an idea; with information on how it spreads and grabs people and it's position in the minds of people.

See I will post these words in this order and then it will spread as an idea as people share it on their own volition. It's an infection of sorts; mind virus is definitely an underwhelming term as it's been restricted to focus on one specific thing when it could be considered in a more general context.

It is very interesting to think about how much power words have and how much possession of someone's brain you can have despite never meeting them.

Putting NASA in their just gives phrases for some reason a certain staying power. NASA does so much in so many things that it sort of has lost meaning to me to say "NASA does something" given how broad the organization is.

For the record, I am not anti-nuclear; in fact I am generically positive as most space people given the engineering links. I think storage and disaster concerns are overblown, but actual implementation and socially acceptable power concerns under. The actual risks in the NASA report are still fairly insignificant and didn't matter in the end; I fully support the launches of Perseverance, Curiosity and Cassini.

I am just really curious to track how this spreads. (sorry nuclear fans)

[jenakuns]

Starship HLS performance hole

Hey so, Artemis is pretty cool. SpaceX bid a modified variant of the Starship upper stage to act as the lander that is currently manifested to take US astronauts back to the Moon. I mean schedule wise it is completely screwed, but that was obviously from the day it was selected. The problem here isn't whether it happens in 2030 instead of 2026, rather how it happens at all.

Now with the protests that followed the sole sourcing of SpaceX, we got a bid launch count of 16 for Starship HLS; 1 depot, 1 lander and 14 tanker (to the depot). This is a large number that has continuingly generated arguments about it's largeness not aided by the napkin math of Musk. It does represent architecture complexity and in currently reasonable world would represent a not small expense (at $2B a year, with 50 launches of Starship, that would still have a cost share of $640M).

But this is just what they bid in 2020 December, how is it going in 2023? Not good. Because of the nature of the architecture; Starship has an extra 3km/s to travel as a single stage; which really tightens the belt on the supposedly large margins it should have as a 1300 ton stage. + Raptor is a large high thrust engine, good for launch vehicles fighting Earth gravity; bad for in space stages where throttling gimps your ISP and thus performance. So launch count hasn't gone down from 16 like people wanna say; it's actually floating even a little up. Now exact launch numbers I don't believe in, because there's a lot of TRL numbers that aren't 7-9.

Launch count isn't only thing; margins are just straight tight. No ZBO + high delta V gives low remaining %s after doing it. Architecturally you can do thing to plug the hole kinda, but fairly large increases launch count and complexity with multiple refillings of Starship in various orbits.

Also reusing a cryogenic stage without zero boiloff in NRHO and which the Gateway people will riot if you leave attached to their station is hard.

Solutions? Innately I just wanna make the lander a bit smaller. That reduces the launch count; although makes the performance hole bigger; meaning more scuffed positions in the architecture to fix reducing the decrease in launch count. To elaborate, performance hole would increase because dry mass ratio increases as the individual components begin taking up more. But if the lander was like 500 tons wet mass instead of 1200 tons that would be nice. Also a smaller engine that Raptor. Closed expanders may be too small at 100 kilonewtons; unlikable burn times without large engine counts. So maybe you do like a 200 kilonewton full flow staged engine?

But like, 8-9 launches and not nail biting margin architecture is nice.

Any custom built lander will also be expensive and increase cost estimates way beyond what 4 bil was bid on. Will SpaceX persevere? Well if Starlink is the money printer they want it to be; expending the resources to develop their lunar capabilities is reasonable. (and also the right thing to do!)

There this thing where like people who dislike SpaceX like their cancelled plans, or at least hypothetical versions of it. I don't think I'm falling into that category.

I really don't like the tumblr post editor. Give me the entire screen please.

[jenakuns]

Why not both? Because it's violating the LAW you reckless individual.

NASA logo time

Some folks like the classic meatball logo.

Others like the futuristic Worm logo.

But for me, I say "why not both?" I give you the Worm ball:

[jenakuns]

Recruiting people to the idea

A fact that recently annihilated me came up in this podcast. Robert Zubrin is a fairly well known vocal engineering figure in the space community, most popular for his Mars Direct architecture. I don't factually agree with everything he says, but he speaks very engagingly and often brings up points and facts that do make do the think. An idea that he frequently brings up is the concept of recruiting people to the idea. Basically the words and actions of someone can inspire others to partake. And those secondary individuals can be very capable people; giving you outsized impact.

And this is something he really emphasizes as a space advocate. He has his own small engineering lab which keeps him occupied, but in his spare time he does many things, but in this context we're discussing his book writing. (I will admit, I haven't read any of his, probably should, but besides the current point.) They cover a whole bunch of things, but in this case we're discussing the space manufacturing study in The Case for Space. In that podcast above, Delian Asparouhov, the founder of Varda Space, states that what got him interested into space manufacturing was that book. Varda Space is a space manufacturing company valued at like $200M that is launching it's first space factory (sort of) to orbit in 30 days.

And that just hits me like a truck.

Oh wow such an original thought Barry I sure hope that it hasn't been around for millennia oh wait "pen is mightier than the sword" and many things older than that

But come the fuck on dude. Sure we've heard and seen it demonstrated, but this is just so explicit.

"I want somebody to do a space manufacturing business, so I am going to write a business case for it in a book. Hello I am somebody capable let's do this business case"

And this gets me onto to the point of how much of what we think, was crafted in like thought center's or at just someone else head. Like are most thoughts the generation of a concerted effort by a small group of people vs an organically forming in a community stirring pot.

One thing that I thought of when thinking about justifying diversity beyond just people are cool was 'diversity of thought.' Which would be a desired thing getting to an optimal solution where 10 different people will give you 10 different solutions to a problem. I organically came up with this term. But then it hit me about the other insinuations and I would be not first one to arrive at such a phrase. And sure enough yes it's been discussed a fair bit. And this leads me to the question; well did I organically arrive at this same place, was brokens bits of the machine left around the place with scraps in articles or did I just read it somewhere and subconsciously pick it up?

I don't really know.

[jenakuns]

Artemis CLPS is very chaotically exciting

Artemis is a really fun program to follow. You never know what is going to happen really and all the schedules are so shot that it makes predicting how it happens a really fun time. There are a bunch of fascinating technical elements and it just has bounds of nuttiness.

CLPS is great. I've come quite a fair bit down from my 2020 thought structure where I thought CLPS companies would be providing bids for HLS. But there is still the after glow and so I can't help but have a positive thoughts about it. For the past 2 years the prediction has been every CLPS mission will be delayed or fail. So far it's been good. This year the wheels touch the road; or legs touch the surface, to provide a more contextually accurate metaphor. The velocity at which that happens is the question.

Essentially CLPS is faces month/year delays on every time you sneeze. Astrobotic has been playing schedule chicken with ULA for like 2 years now. Remember the good ol days where it was IM-1 being delayed to 2022. Now it might not even launch in 2023.

The NASA estimated reliability for these initial is 1/3. Yes. So NASA is expecting a lot more impactors than actual landers initially. I genuinely think there is good odds of a repeat of Mars 1999 and Faster Better Cheaper, where a congressional review asks why the repeat failures; sort it out. In this, I expect unfavorable comparisons to the Chang'e program.

The one mission they actually care about, VIPER, has had a cost increase of $121M from $199 million to $320.4 million. This indicates that the initial bid had nowhere near enough mission assurance and fiscal reality to have any good chance of landing the $0.5 billion rover on the surface.

The Sierra person I've seen speaking on occasion said Sierra investigated CLPS but decided against participating because it didn't think it would be profitable at the margins they are flying. CLPS bankrupted Masten Space System

But if it succeeds; dude this shit is crazy. This is 2-3 robotic landings on the Moon every year. Basically Artemis in a nutshell. I was planning on talking about everything in the program in 1 post; but I think it would be better to spread it out. Good luck ispace, humorously the realest CLPS company that can't really directly take part in the program.

edit: Well, we're beginning on a strong failure note. We'll be seeing more I'll say.

[jenakuns]

Starship flight: the SLS comparison

I've been thinking a lot about the flight (and so has everyone + a fair bit of talking). So here we're doing another post.

One thing that's kind of been under the radar in the discussion is the timeline. Because everyone is discussing the outcome. So large scale development of Starship began in 2018. You might start with Raptor dev; but given the main comparison will be SLS; trust me, you want 2019. We have this first all up flight in April 2023. 4.5 years to first flight (failure).

Now Saturn V went from initiation in 1961/2 to first flight in November 1967, so 6 years. And that flight was successful. But that was with $46 billion, but that was with slide rules; let's just not discuss it. (the comparison breaks down with that much of a time gap)

SLS is the modern NASA SHLV, so how did it go. From program start in 2011, it achieved it's first launch in November 2022. 11 years. Now this flight involved a entirely successful launch of incredibly complex spacecraft to a lunar orbit. So the launches themselves aren't comparable, but the programs are. The next SLS launch is in late 2024, maybe 2025. This is a crewed lunar launch; again not really comparable to anything that Starship is doing near term. But when is the second Starship launch? How many orbital launches will have been done before Artemis 3 in like 2027 (Gateway mission).

Had the OLM not been significantly kicked in the shins, B9 and S25/6 would have been launching in like 3 months, compared to the 2 year gap between A1 and A2. SLS will have a 100% success record no doubts; but Starship will have a much larger book of launches in that same timeframe.

11 years from 2018 is 2029. What do people think Starship is going to be doing in 2029?

But there was a beatdown on the OLM. It got curb stomped as the kids might say. Climbed 222 stairs just to fall down 222 stairs. This is going to make the gap a fair bit longer than 3 months; how long I don't really know.

Now, had they waited for installation of the water cooled steel + water deluge, they would be in a better spot no questions asked. But being able to make these judgements when there are 1000s to make and dozens of different people saying it's my subsystem that will cause it to end it all is hard. Or maybe Elon overruled everybody who stated otherwise and it is all his fault. Shrug.

Installing the steel plates and deluge would have taken like 2 months knowing work continuing on the OLM. By which time you might well skip B7S24. But the potential learnings would be useful now for all the work they're doing. The concrete wasn't too bad after the B7 static fire. So let's do a launch

How the concrete faired after 31 engine 50% throttle static fire.

Surely it could take 1 launch. I mean sure it would be fairly eroded, but we're going to have to dig it up anyway for implementation of the cooled steel plates.

When you're trying to do a fully reusable SHLV on a tight schedule with not absurd budget; you're going to run into walls sometimes. But they're still walking.

[jenakuns]

Starship launch woot woot 3am discussion

Yeah so uh holy shit. That was definitely a launch that happened. It's such a weird program; extreme by most metrics that it's kinda hard to think about. The main thing that appears to be catching people up here is RUD vs explosion. I mean, RUD been used for a while now and is sort of a tongue in cheek joke, but I could understand why this being the first exposure could catch people.

There's also this post which is cosmically funny. All the complaining about Blue Origin no orbit has come back to bite the SpaceX stans because jokes on them; people are now confusing the two and saying SpaceX no orbit lol. Cheers @greetings-inferiors you gave us a good laugh.

But that's tumblrs reaction, what's mine? Idk, I laughed a fair bit. Schedules were always shot to hell, so I'm not too fussed in that regard. The rocket itself doesn't matter at this stage; it's the OLM damage, which I suppose separates this other launch failures. I still think H3 failure really reduced my expectations of first launches outcomes. If a country that's been doing launches for 30 years can't successfully launch a fairly conservative evolution of an existing rocket; then all bets are off for everyone else.

It's always nice to have a reset of expectations for the entire program so that means when things happen they can be appreciated rather than trodden over. (cough cough Falcon 9)

Launch wise, it was incredibly scuffed. 3 engines out from the get go which resulted in an extended burn at OLM to gain sufficient TTW resulting in that fat crater. The engines out also resulted in a slide that is scuffed and scary for a vehicle of Starship size. But the flight continued. Then more engines out, an hydraulic power unit explosion, some more engines, and oh yeah the tiles. (I don't care about the tiles). And yet the vehicle still flew. It's showed robustness in the scuffed of situations of it's own creation. When you roll around with B9 and the reliability enhancements, it actually is fairly interesting. They took engine explosions or at least flame outs in their stride, but you know; don't have them in the first place am I right.

The question is now when next flight and that really determines the value of this one. Because it's always the next product™. But presumably the Raptor ops and propellant management will be good outcomes. Also not to use HPU, but that was already a learned lesson. This failure just feels unusual because of the scale I suppose.* (I mean it's automatically notable because Musk rocket failure, but that's beside the point). Like we've had RS-1, ZQ-2, Terran 1, LauncherOne, H3 and Vega failures recently; what's different. Well, entirely different classes. But still, nowhere near media circuses. *+the damage to the ground infrastructure.

To summarize, the vehicles job was to gain experience on the performance of the many subsystems that make up this vehicle. Like autogenous pressurization on a vehicle of this scale as it handles many Raptors (failures) is no small feat. The next 2 Starships have given up heat shields and payloads, because they're also now exclusively marked as flight test vehicles as well. There was a lot of things that went wrong, but it still gave a fair bit of insight into the vehicle for SpaceX. And if that was just it; it would be fine because the next vehicle would be rolling to pad in a couple months. But this happened.

And that means a fair bit of work in repairing, upgrading and potential legal issues to grind out, which will cause delays in its own right. This is what pushes the needle in the direction of not worth it. But shit has happened and will continue to happen; the program will just move on.

Ooo forgot; no new HLS renders; fuckin bullshit world we live in.

[jenakuns]

Approaching any internet discussion

I maintain that anyone entering a discussion/argument on the internet with the intent of convincing the other person of their point of view is a moron beyond belief. I myself have fit this characterization on occasion (well maybe not just on occasion). The idea that your words are going to be convincing to the person hearing most of the time is delusional. People have a hard time understanding simple statements on the internet and the meaning behind them; there will be no assumptions of good doing when there is conflict. Anything you say can be brushed off at a moments notice and will be. Because what do you know? The people you're arguing are more likely going to be ideologically convinced of their point; they already have all the answers anyways; you're wrong.

But that doesn't mean these discussions are without merit. It just needs to be defined around the impacts on you; because you are having none of the impacts you want on the other person. If you enjoy arguments on the internet and annoying people, good for you. That's not me personally; I don't enjoy conflict of any kind and even the smallest criticisms cut deep to the core with me. The reason I enter is to gather facts and sharpen my knowledge base. Anytime I want to say something, I generally make a point of googling or looking through the relevant pdfs to ensure what I'm saying is not without basis (which jeeze luoise; anyone who asks for a sauce when you're sitting on top of the greatest archive of knowledge the human race has ever created; take some fucking initiative).

The amount of times I've done this and found that I've completely misread something scares me; how much of what I know is wrong just based on my consumption of the material. Which makes me scared thinking about other people; maybe I'm more scatted brained than the other people; but then again, maybe they're walking around with a bunch of wrongs in their heads as well.

I'm also practicing my ability to gather those resources and present them in a clear format. Misinterpretation is rampant, so saying things in a lucid manner that is able to get across to the other person clearly, is impressive. A side note to include, these discussions also tend to scale. 1 person says 1 sentence; person b replies with 3 sentences to counteract A, A replies pointing out inaccuracies in those 3 sentences with 2 paragraphs and pretty soon we're out of space for the argument on the internet. Try to avoid getting argument sniped/scope changed where you end up discussing completely unrelated points to the first point. It's in these scenarios you end up saying the dumb stuff.

Ultimately, the opinions which you hold right now; the ones defining the argument, are probably worthless. Opinions are merely a means to the end; giving you a frame of reference from which you can use to travel around in to find new information. Because that opinion is generally formed in the absence of counter ideas and without the whole factual picture. So thinking you're already right entering is foolish; you should expect to hopefully learn something or reevaluate something else. If you're not getting any of the above out of the discussion; you're convinced you're flagrantly right; leave. It's your precious time on this Earth you're wasting. Well, I suppose you could still be doing it on a public platform; for the clout of winning.

Facts are everything, but also nothing without a common line that is able to handle the context that generates those facts and produce a cogent point of view.

[jenakuns]

Launch vehicle reliability

I feel like there's a mistake people make in that they believe they can have a big impact on reliability with architecture decisions. In recent memory I think of this slide released by Virgin Orbit (RIP) in one of their investor decks.

Like I can simply tell you a launch vehicle's reliability via first stage engine count. This is superficial. LauncherOne didn't really have a large history to call on, but it's reliability is below theoretical with 4 out of 6 and was detached from whether they had 1 engine or 9. RocketLab has actually now crossed to 91%, beating theoretical. And both of the recent failures of Electron involved the single engine upper stage, not the 9 engine first stage. Astra was at 50%, which I suppose is technically correct as it's <88%. But myopic view as none of the failures are directly caused by the 5 engines. And there was more failures of the dirt simple single engine pressure fed upper stage than of the first stage.

These failures aren't baked around engineering margins which you can calculate; they are design, management and QC issues. And there isn't a formula that can output reliability from inputted cultural and management approaches at a company.

My favorite one is where SpaceX paid Futron Corporation back in 2004 for some analysis that showed that their single engine Falcon 1 was actually optimal from a reliability perspective and Falcon V was even better cause it could sustain engine out.

Engine Reliability It was with this in mind that we designed Falcon I to have the minimum number of engines. As a result, there is only one engine per stage and only one stage separation event – the minimum pragmatically possible number.

In the case of Falcon V, there are five first stage engines, but the vehicle is capable of sustaining an engine failure at any point in flight and still successfully completing its mission. This actually results in an even higher level of reliability than a single engine stage. The SpaceX five engine architecture is an improved version of that employed by the Saturn V Moon rocket, which had a flawless flight record despite losing an engine on two of its missions.

Which is funny because they then arrived at a 9 engine setup (and later a 27 engine aha (AND NOW A 33 WOOO BABY)) and have to like it for reliability.

Falcon 9 has nine Merlin engines clustered together. This vehicle will be capable of sustaining an engine failure at any point in flight and still successfully completing its mission. This actually results in an even higher level of reliability than a single engine stage. The SpaceX nine engine architecture is an improved version of the architecture employed by the Saturn V and Saturn I rockets of the Apollo Program, which had flawless flight records despite losing engines on a number of mission.

Now in this regard eh. Falcon 1 flight 1 was lost because who knew metal valves would corrode under salt water spray. Flight 2 was lost because Elon thought they wouldn't need slosh baffles on the second stage as that would be wasted mass and overruled his engineers. Flight 3 was lost because of performance enhancements with Merlin 1C causing collision after staging due to residual thrust.

Falcon 9 then went on to have an engine failure on flight 5 and while it did still manage to achieve the launch, demonstrating the engine margin aspect, they did loose the secondary payloads. The two proper launch failures had to do with struts and COPV, which are entirely independent of engine count. We have had a more recent mission in early 2021 which had an engine fail and it still succeeded, although booster failed to land.

Another failure group that stick with me is The operationalization of Air Force space programs. 'We need to operationalize our launch vehicles such that a history major can run them instead of an engineer.' And so the upper ranks of the Air Force went ahead and dismantled the engineering culture that made these programs at least reliable and got rid of those "guys in white lab coats." And so they lost $5.6 billion worth of satellites and launch vehicles due to 3 failures that can all be attributed to poor quality control.

And that linked article makes a good point: "It is tempting to say that management is the cause of all launch failures, and in one sense that is true." Because if you just gave your engineers more time and money, they probably would have fixed the issues right? Well A. there's any number of things to spend your money on; you need to avoid bankruptcy and B. there are genuine engineering screw ups with time and money anyways. So better to just get off the pad and flying?

A final note on the R7 or Soyuz now. As a design, it has perhaps one of the worst conceptual reliabilities. It has 5 engines on the first stage with hydrogen peroxide turbopumps and not a single one of them can fail as they are each a booster. And because there is 4 boosters that means 4 separation events that also have to work. There is also a liquid third stage which of course has to separate and ignite successfully + the occasional 4 stage. 5 critical separation events and 6 engine burns that all have to go well. And yet it's reliability is kinda just fine with modern variant having 95.5%, with some older variants having better records. There was a stretch in the 1982 with 182 successful R7 family launches in a row.

You can't determine reliability from design parameters; because there is so much nuance in executions of programs