totm nov 2023 SpaceX Discussion Thread

[Ultimate Steve]

58 minutes ago, Exoscientist said:

 The point is the 5 second burn was not diagnostic of the problems that would arise in the first test flight. It also didn’t diagnose how many engines would fail during an actual flight burn, eight of them with two actually exploding.

 THIS is how a static burn is supposed to be done:

Where are you getting the 8 from? I counted six with one restart, and we know three of them failed during the startup procedure (and if a fourth one had failed, it would have been an aborted takeoff. Setting the limit assuming nothing else would fail is my second biggest issue wth IFT-1, second to only the failure of the FTS).

 

I think they are justified in doing short static fires in this case. The most complicated parts of rocket engine operation are generally startup and shutdown. There's also the issue of vibrations causing problems as the fuel drains a la N1 but those cannot be discovered when the vehicle is affixed to the launch pad, the dynamics in play are not identical (and would change significantly throughout the flight). That's a situation where problems can only be discovered by flying it or with really really good computer simulations.

Absent of this, after sufficient ground testing, an engine that lasts five seconds will probably last a lot longer.

Of the six engine failures on the starship test flights (assuming I counted right), three were aborted ignitions, the kind that should have been caught by short static fires, and the kind that, imo, should have aborted the launch, but it is now clear that the engine out limit was not set up in a way to deal with what ended up happening.

The other three failures were in flight failures, and they were able to restart one of them (I think that is the first time that's ever happened on a rocket actually). Assuming the rocket had made it off of the pad with those three engines healthy (if the limit had been set to 0, this would have taken several launch attempts most likely) it would have been able to take the remaining 3 engine outs (although it still would have failed when they lost all gimbal control, and later when the FTS didn't work (which is by far the most alarming thing about IFT-1, a few engine outs don't even come close)).

Basically, of the six engine outs, a full duration static fire would have, at best, caught two and a half of them (which may have been flight specific issues, we don't know, it could have caught none), at the cost of having to build a massive test stand, and we know what a kerfuffle it was getting the permits to build their fairly minimal launch pad as is, and that was specifically designed to avoid having to do much time and paperwork intensive business.

This particular static fire was also as much about testing the new deluge as it was testing the new booster.

 

Given the chance to do a full duration ground test at little to no additional cost, anyone would. However, the infrastructure needed to do that does not come fast or cheap. SpaceX judged that the could get the same data for cheaper in other ways. So far I think that judgement is correct.

[magnemoe]

10 hours ago, tater said:

Why?

Fly it, instead. They have a barn full of them.

I agree an short test fire is enough, perhaps 10 second would be better. But it looks like they have problems with the ignition and stable burn of all the engines. I suspect the pressure in the pipes drops then the engines turns on. Raptor is hard to start as you need each of the turbopumps to provide fuel or oxidizer to the others and I'm sure pressure into the pumps are critical. 
Once all the engines are running you are in an steady state. Its many ways to solve this, easiest is probably to stagger the firing sequence more, an more permanent is to redesign the piping to reduce the pressure loss. 
The engines themselves are easy enough to test. It could also be vibration or other issues starting all the engines. 

Edited August 12, 2023 by magnemoe

[tater]

6 minutes ago, magnemoe said:

Once all the engines are running you are in an steady state. Its many ways to solve this, easiest is probably to stagger the firing sequence more, an more permanent is to redesign the piping to reduce the pressure loss. 

Yeah, I agree they probably want to achieve steady state. They'd be limited by the time the deluge can run. It will need to have more duration than strictly needed for a launch probably.

[CBase]

2 hours ago, tater said:

It will need to have more duration than strictly needed for a launch probably.

So opportunity costs to optimize booster start are some additional water tanks vs. early failures in test flights

Seriously I miss the fast testing iterations from early starship. Even small hops of booster would be awesome and probably help to improve startup sequence.

[tater]

1 hour ago, CBase said:

So opportunity costs to optimize booster start are some additional water tanks vs. early failures in test flights

Seriously I miss the fast testing iterations from early starship. Even small hops of booster would be awesome and probably help to improve startup sequence.

Yeah, there might be other constraints as well. Capture of spent water, etc. But again, many other launch vehicles never do a full duration static fire. SLS has never done one, the Shuttle never did, or indeed any rocket with SRBs that I know of.

We also have no idea what data they have about how their failures in large groups of engines have happened. I'm in wait and see mode.

[Exoscientist]

 Has there been any consideration of the 4 engines failing after only 2.7 seconds in the static fire being due to water deluge coming *upwards* from the center, thus with an upward force against the engines? Usually the water deluge comes from the outside horizontally.

Water deluge by SpaceX:

 

Water deluge by NASA:

 

    Bob Clark 

 

[tater]

4 minutes ago, Exoscientist said:

 Has there been any consideration of the 4 engines failing after only 2.7 seconds in the static fire being due to water deluge coming *upwards* from the center, thus with an upward force against the engines? Usually the water deluge comes from the outside horizontally.

Nah, SpaceX considers nothing, they just build random junk and blow it up.

Seriously, if water impingement was a likely cause, they would likely alter GSE procedures as the fix. Ramp deluge pressure differently, the same way they adjusted engine start timing on FH.

 

[darthgently]

Around the 1h45m+ mark they start poking around the launch mount and deluge area.  Paint held up so water is converting a lot of heat to steam.  Some surface cracking in the high temp Fondag concrete layer

 

1 hour ago, tater said:

 

I'm  leaning to the methane Tanker SS guess in the comments

[tater]

 

[CBase]

12 hours ago, tater said:

But again, many other launch vehicles never do a full duration static fire.

I never proposed a full duration static fire, that was some other guy. Just your comment that they don't reach the stable state of engine startup in static fire due to deluge limitations is the point more water could help.

And since Raptor is deep throttable iterative testing could target lower stable throttle levels first and then keep increasing it. G-Forces are quite low with the full rocket, so pad tests should be quite close to early flight phase. I thought SpaceX is all about speed, but somehow for booster development they stopped to rethink testing.

[king of nowhere]

a worry for further in the future, but there is an issue with the full plan for a potential mars landing.

on start, starship broke because the rockets were too powerful, they tore chunks of concrete from the launchpad that hit the rocket. they want to solve this with the water suppression system.

but what about the final goal of landing on mars? true, starship will be a lot less powerful by then, but it will have to land on gravel. in a thin atmosphere, so there will be nothing to stop those shrapnel going everywhere. and they can't reinforce the engines, they need to be thin to cool off.

so, how do they plan to survive landing on a hostile surface?

[darthgently]

2 hours ago, king of nowhere said:

a worry for further in the future, but there is an issue with the full plan for a potential mars landing.

on start, starship broke because the rockets were too powerful, they tore chunks of concrete from the launchpad that hit the rocket. they want to solve this with the water suppression system.

but what about the final goal of landing on mars? true, starship will be a lot less powerful by then, but it will have to land on gravel. in a thin atmosphere, so there will be nothing to stop those shrapnel going everywhere. and they can't reinforce the engines, they need to be thin to cool off.

so, how do they plan to survive landing on a hostile surface?

As you note, it will only be SS, so far fewer engines.  And lower gravity too.  But good points.  I'm imagining that scouting the exact landing location beforehand and doing an F9 style pinpoint landing would be part of the recipe if they can find a flat clean rock surface.  Maybe a wall-e rover could sweep the LZ beforehand?

Maybe a hover at a "safe" x meters for n seconds could clear most of it.

Interesting problems

[RCgothic]

The last variant of lunar Starship we got a render of showed landing engines high in the fuselage, so I presume Mars Starship will be similar. This isn't something SpaceX won't be aware of.

Edited August 13, 2023 by RCgothic

[Exoscientist]

16 hours ago, tater said:

Nah, SpaceX considers nothing, they just build random junk and blow it up.

Seriously, if water impingement was a likely cause, they would likely alter GSE procedures as the fix. Ramp deluge pressure differently, the same way they adjusted engine start timing on FH.

 

 By that argument they would have realized doing the test flight without a flame diverter was a bad choice as well. And would have realized the FTS system was inadequate to cause immediate destruction of the vehicle. And would have realized using a flipping motion to do stage separation was a bad approach.

 You can not assume a approach SpaceX decides to take is a good one just because SpaceX chooses it.

 Again, rather than SpaceX dismissing the lessons of Apollo, they should learn from them.

    Bob Clark

[darthgently]

4 hours ago, Exoscientist said:

 By that argument they would have realized doing the test flight without a flame diverter was a bad choice as well. And would have realized the FTS system was inadequate to cause immediate destruction of the vehicle. And would have realized using a flipping motion to do stage separation was a bad approach.

 You can not assume a approach SpaceX decides to take is a good one just because SpaceX chooses it.

 Again, rather than SpaceX dismissing the lessons of Apollo, they should learn from them.

    Bob Clark

And yet, they succeed where others shouted it was not possible.  They did learn the biggest lesson Apollo taught by way of cautionary tale:  Spend a lot less money (with some costs) by learning how to spend a lot less money and you can't plan everything in advance when the unknowns abound.  You iterate.  SpaceX has learned from every preexisting space program from what I can see

[magnemoe]

8 hours ago, Exoscientist said:

 By that argument they would have realized doing the test flight without a flame diverter was a bad choice as well. And would have realized the FTS system was inadequate to cause immediate destruction of the vehicle. And would have realized using a flipping motion to do stage separation was a bad approach.

 You can not assume a approach SpaceX decides to take is a good one just because SpaceX chooses it.

 Again, rather than SpaceX dismissing the lessons of Apollo, they should learn from them.

    Bob Clark

Yes the test fire without diverters or deludge was an mistake. As I understand the damage from the static fire was pretty minor so they assumed the launch would be the same, but it was longer and probably more power.  Yes this was stupid as they stated how important stage 0 is, so protect it. 

Flight termination system was probably calculated using aluminum, but steel probably is far more resistant to explosions. FAA also approved it. Using stainless steel of this thickness is new for rockets. 
I say the flip was not stupid, if the flip was attempted I assume some software stopped it or more likely they was so far outside the flight envelope FTS was required. 
They might have attempted it but it failed.
So they go for hot staging instead, on an reusable rocket who is nuts. On the other hand they are the only one who has reused an orbital first stage. 

Now will they launch with the holes in the interstage / extension? or have some aerodynamic cover?
They will obviously need an cover to protect the top of the first stage who hold the grind fin control and other stuff and protect that from the flames. 

 

[darthgently]

1 hour ago, magnemoe said:

So they go for hot staging instead, on an reusable rocket

If the only company that has repeatedly landed reusable boosters decides to try it a different way, I'm going to assume they are the best qualified to explore that different way.  I'm just glad I've lived long enough since Apollo to see this  renaissance of space exploration

[Exoscientist]

 

Edited August 14, 2023 by Exoscientist

[darthgently]

9 hours ago, Exoscientist said:

 

Maybe from vibrations, from deluge flow and thrust.  Maybe laying the Fondag in hex tile pattern would alleviate.  Either way, Fondag is surface only and quite patchable

Some deeper thoughts...

 

[RCgothic]

Yeah, the fodag's just a cap layer, a few cracks in probably aren't cause for concern.

The reinforced concrete foundation is now 2.2m thick, a few surface cracks aren't going to trouble it.

It's really notable how well the paint on the OLM withstood the static fire this time, looks like a near 100-fold reduction in stripped paint. The OLM engine-camera also survived this static fire intact which it has never done before.

It's not zero damage, but it's much, much better from what we can see.

Edited August 14, 2023 by RCgothic

[sevenperforce]

On 8/12/2023 at 1:02 AM, Exoscientist said:

 The point is the 5 second burn was not diagnostic of the problems that would arise in the first test flight. It also didn’t diagnose how many engines would fail during an actual flight burn, eight of them with two actually exploding.

 THIS is how a static burn is supposed to be done:

{Saturn V S-IC}

 SpaceX should stop dismissing the lessons of Apollo and learn from them.

If that's the case, then the same would apply to the Shuttle (would this have prevented Challenger?), Delta IV Heavy, Falcon Heavy, SLS, and many many other rockets that don't have full-duration static test fires.

[Exoscientist]

 Something just occurred to me. The comparison has been made of the SpaceX SuperHeavy/Starship approach to the Soviet multiple failed N-1 rocket in that they both wanted to test by actually flying the full rocket until it works. 

 The comparison was criticized on the grounds the N-1 engines were not tested individually. Instead, the engineers selected an engine at random from a batch to see if that worked. If it worked the entire batch was chosen. The engines could not be tested individually because the testing was destructive. That engine could not be used if it were tested.

 The SpaceX Raptor engines on the other hand are tested individually. But here’s the major failing of the Raptor: even if the engine is tested successfully there is still a quite high chance the engine will still fail when used on a flight. That is a major flaw in a rocket engine. No rocket engine would be considered successfully developed with that flaw.

 Because of the numerous failures of the Raptor both on the test stand and in short test hops of the Starship landing methods prior to the April test flight, I estimated the chance of engine failures of the SuperHeavy/Starship test flight  was 1 out of 3. SpaceX claimed prior to the test flight their Raptor 2 was more reliable. The result? Only 1 in 4 of the engines failed. That is still a stunningly high percentage.

 The upshot of this in a very real sense the Super/Staship is just like the failed Soviet N-1 in flying with engines with poor reliability.

  Robert Clark 

[.50calBMG]

The difference here is that the raptor is a completely novel engine design that had never been flown before, so failures should be expected. They are also testing and iterating on every engine, with raptor 3s starting to show the fruits of that labor. Raptor 3s are proving that they are substantially more reliable than the 2s, which in turn are more reliable than the 1s. You can't point to a few failed engines early on in a test program and say they are unreliable and always will be.

Take a look at how long the RS-25s took to get right, and even after all those flights on the shuttle, sls had multiple failures with the same engines, so are those unreliable as well? Same goes for the F-1s

[Terwin]

It seem to me that so long as the only cargo being launched is telemetry equipment, it is clear that any launches are just further testing, and things like engine failures are to be expected.

A test-launch seems like it would help examine and validate better than test-fires, and considering how hardware-rich the starship program is, it is probably cheaper to dispose of some older engines and an old airframe in a test-launch than to perform the testing needed to get the same data on the ground, especially when one of the bottle-necks is launch-pad repairs.

With 33 engines, you could even have batches of engines with different optimizations all lunched together to see how well they work.  It might be very dangerous to do this without the engine separation hardware or a large throttle capacity, but both of those are present on SH.  I would not be surprised to learn that the failed engines all had similar optimizations that have since been adjusted or discarded. 

 

[Kartoffelkuchen]

1 hour ago, .50calBMG said:

The difference here is that the raptor is a completely novel engine design that had never been flown before, so failures should be expected. They are also testing and iterating on every engine, with raptor 3s starting to show the fruits of that labor. Raptor 3s are proving that they are substantially more reliable than the 2s, which in turn are more reliable than the 1s. You can't point to a few failed engines early on in a test program and say they are unreliable and always will be.

Take a look at how long the RS-25s took to get right, and even after all those flights on the shuttle, sls had multiple failures with the same engines, so are those unreliable as well? Same goes for the F-1s

I think the point he's critizicing is that SpaceX choose to go ahead with the first orbital flight test, while they knew their engines still were pretty unreliable and the vehicle had a high chance of multiple engine failures / loss of control because of that.

No one blames them for exploding engines when they are still in development and are properly tested on the ground. However, knowingly sending your rocket up with engines which are likely to fail on ascent is a bit "meh". Obviously we don't have the full picture though. It could be that they were done simulating everything they can, and just had to know if their launch vehicle design worked as their simulations shown to progress their work. At some point you have to make that jump.

Still, I agree with him here that perhaps the test feels a bit rushed, and if they had spent a couple more months proving their engines work halfway reliable it would have greatly increased the chance of a successful first test flight.