SpaceX launches

In reply to sobe_death :

Thing is- that's all known, simulatable, physics.  The shape of the body isn't unique, the maneuvers are not unique or new, the falling path ins't unique, the engine requirements are not that unique anymore.  As an engineer, I see all of the stated requirements to met easily by a ground based test.  Which also gives the engineers parts to look at after the test was done- making sure the engine issues can easily be debugged by looking at both data and hardware.

This isn't a landing craft for earth- otherwise it would need a total redesign to accept heat shields.  To get earth atmosphere anywhere near Mars, the ship would have to go up another order of magnitude in height for Mars surface dynamics.  Heck, even doing the maneuver in a earth based gravity field is incorrect- since Mars gravity is less than the earth by quite a bit.

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

Man, I'm still running version 1.11.1 of Kerbal Space Program.  Looks like Elon got early access to 2.0.  Impressive stuff.

alfadriver (Forum Supporter) said:

In reply to sobe_death :

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

Which is why you aren't a rocket scientist.

Javelin (Forum Supporter) said:

alfadriver (Forum Supporter) said:

In reply to sobe_death :

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

Which is why you aren't a rocket scientist.

Forgive me for being able to think and consider what they are doing.  It's really not that hard to understand.

alfadriver (Forum Supporter) said:

Javelin (Forum Supporter) said:

alfadriver (Forum Supporter) said:

In reply to sobe_death :

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

Which is why you aren't a rocket scientist.

Forgive me for being able to think and consider what they are doing.  It's really not that hard to understand.

But it is, and you don't seem to be able to comprehend that. Do you really think SpaceX would be performing these very public tests and failures for no reason? They are so far beyond any envelope pushed by NASA that no amount of "ground testing" is ever going to give them the data they need. There is so much more going on here then you either want to or are able to acknowledge. We literally just witnessed the fastest ever tank/detank/retank in rocketry history, the first ever light off to thrust/auto-abort and then launch without rebuilding the whole thing, and the first successful touchdown of a rocket to return horizontally in history and you are asking what the point is. And yes, STarship most certainly is being designed to RTB to Earth as well as Mars.

alfadriver (Forum Supporter) said:

Keith Tanner said:

What's amazing  about this launch today is that it included a new concept in relighting and initiating the flip. Which was suggested, planned and put into practice in a month. Yes, the touchdown was a little harder than ideal and/or something went wrong with the legs that led to LOV, but that's an incredible pace of progress. 

???

That concept was part of SN8 and 9.  During the righting phase in SN8, it blew up (IIRC one of the engines didn't re-light), and SN9, it was not quite vertical.  Maybe I need to look back on those launches, but I very much remember it laying on it's side as it came to earth.  And then the engines fired starting to make it vertical again for landing.

Actually, that's not quite right- at 6:34 into the launch of SN8, it went upright, but one motor was not powerful enough for it to stop the momentum- 1:54:47 of this video https://www.youtube.com/watch?v=ap-BkkrRg-o&ab_channel=SpaceX

And the same thing happened for SN9, almost at the same moment relative to the launch- 6:26- (11:53 of this https://www.youtube.com/watch?v=_zZ7fIkpBgs&ab_channel=SpaceX)

SN8 suffered from fuel starvation during the flip. That was solved with a different header tank pressurization scheme.

SN9 had an engine that failed to relight so it didn't have the power needed to perform the flip. At that time, the plan was to light two engines, flip, then shut down one engine and land on the other. But with only one lit, it couldn't rotate enough. Raptors being a new engine design, they're a little harder to light than some of the others out there.

For SN10, they changed the process. Now they're lighting three engines, just in case. Then one is shut down (if applicable), two are used for the flip, one is shut down and it lands on one. That's a new concept that brings some real changes in control logic and dynamics. Remember that rocket engines don't idle, the minimum thrust is fairly significant and the craft is at minimum weight so things are pretty sporty with three engines lit. Plus you don't know in advance which of the three engines will be the "safety engine" that gets shut down or which of the engines will be used for the final landing burn. That's quite a change to make, and it all takes place in a part of the flight where you cannot get it wrong. And it worked. Three lit, two flipped, one landed but the vehicle took some damage.

They also managed to improve the minimum throttling of the Raptor engine at the same time, I believe - you can actually see the difference in the two engines on the way up. In a MONTH.

I think where we disagree is the impact on the test plan from the loss of the vehicle. SpaceX has never reflown a Starship prototype even if it landed successfully. Past behavior would indicate that the actual flight is the last step in its planned lifespan. They've saved three after test flights - one has been turned into a mounting point for cameras and telemetry. The others were all eventually dismantled for scrap.

alfadriver (Forum Supporter) said:

In reply to sobe_death :

Thing is- that's all known, simulatable, physics.  The shape of the body isn't unique, the maneuvers are not unique or new, the falling path ins't unique, the engine requirements are not that unique anymore.  As an engineer, I see all of the stated requirements to met easily by a ground based test.  Which also gives the engineers parts to look at after the test was done- making sure the engine issues can easily be debugged by looking at both data and hardware.

This isn't a landing craft for earth- otherwise it would need a total redesign to accept heat shields.  To get earth atmosphere anywhere near Mars, the ship would have to go up another order of magnitude in height for Mars surface dynamics.  Heck, even doing the maneuver in a earth based gravity field is incorrect- since Mars gravity is less than the earth by quite a bit.

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

The flip is very much unique and it would cost more to build a ground-based test rig than it would to just build a Starship and launch it. The bellyflop skydive did work out of the box although the sheer size of the vehicle (it's about 12 stories tall) makes it unusual.

And this is all for earth-based landings with our atmosphere and high gravity. Mars will do a more traditional hover. The entire purpose of the entry profile is to bleed off velocity and minimize skin heating. There will be heatshields - they're working on robotic installation of them and all of the test Starships carry at least some.

You can see the black heat shield tiles on SN9 in this picture.

What they are testing is everything. You can simulate the heck out of it and do all the ground base tests, but at some point you have to light the candle. They're taking an old-school hardware-rich approach to it where you learn more from actually flying than by simulating, because all you can truly test in your simulations is your simulator design. My "cattle not pets" comment earlier was a reference to how modern servers are treated, there's a lot in common with this.

http://cloudscaling.com/blog/cloud-computing/the-history-of-pets-vs-cattle/

Keith Tanner said:

alfadriver (Forum Supporter) said:

Keith Tanner said:

What's amazing  about this launch today is that it included a new concept in relighting and initiating the flip. Which was suggested, planned and put into practice in a month. Yes, the touchdown was a little harder than ideal and/or something went wrong with the legs that led to LOV, but that's an incredible pace of progress. 

???

That concept was part of SN8 and 9.  During the righting phase in SN8, it blew up (IIRC one of the engines didn't re-light), and SN9, it was not quite vertical.  Maybe I need to look back on those launches, but I very much remember it laying on it's side as it came to earth.  And then the engines fired starting to make it vertical again for landing.

Actually, that's not quite right- at 6:34 into the launch of SN8, it went upright, but one motor was not powerful enough for it to stop the momentum- 1:54:47 of this video https://www.youtube.com/watch?v=ap-BkkrRg-o&ab_channel=SpaceX

And the same thing happened for SN9, almost at the same moment relative to the launch- 6:26- (11:53 of this https://www.youtube.com/watch?v=_zZ7fIkpBgs&ab_channel=SpaceX)

SN8 suffered from fuel starvation during the flip. That was solved with a different header tank pressurization scheme.

SN9 had an engine that failed to relight so it didn't have the power needed to perform the flip. At that time, the plan was to light two engines, flip, then shut down one engine and land on the other. But with only one lit, it couldn't rotate enough. Raptors being a new engine design, they're a little harder to light than some of the others out there.

For SN10, they changed the process. Now they're lighting three engines, just in case. Then one is shut down (if applicable), two are used for the flip, one is shut down and it lands on one. That's a new concept that brings some real changes in control logic and dynamics. Remember that rocket engines don't idle, the minimum thrust is fairly significant and the craft is at minimum weight so things are pretty sporty with three engines lit. Plus you don't know in advance which of the three engines will be the "safety engine" that gets shut down or which of the engines will be used for the final landing burn. That's quite a change to make, and it all takes place in a part of the flight where you cannot get it wrong. And it worked. Three lit, two flipped, one landed but the vehicle took some damage.

They also managed to improve the minimum throttling of the Raptor engine at the same time, I believe - you can actually see the difference in the two engines on the way up. In a MONTH.

I think where we disagree is the impact on the test plan from the loss of the vehicle. SpaceX has never reflown a Starship prototype even if it landed successfully. Past behavior would indicate that the actual flight is the last step in its planned lifespan. They've saved three after test flights - one has been turned into a mounting point for cameras and telemetry. The others were all eventually dismantled for scrap.

I thought you were bringing up the entire maneuver of gliding back down  on its side and then trying to land upright was new.

Javelin (Forum Supporter) said:

alfadriver (Forum Supporter) said:

Javelin (Forum Supporter) said:

alfadriver (Forum Supporter) said:

In reply to sobe_death :

The more I think about the stated testing outcomes, the more I question what in the world they are even testing.

Which is why you aren't a rocket scientist.

Forgive me for being able to think and consider what they are doing.  It's really not that hard to understand.

But it is, and you don't seem to be able to comprehend that. Do you really think SpaceX would be performing these very public tests and failures for no reason? They are so far beyond any envelope pushed by NASA that no amount of "ground testing" is ever going to give them the data they need. There is so much more going on here then you either want to or are able to acknowledge. We literally just witnessed the fastest ever tank/detank/retank in rocketry history, the first ever light off to thrust/auto-abort and then launch without rebuilding the whole thing, and the first successful touchdown of a rocket to return horizontally in history and you are asking what the point is. And yes, STarship most certainly is being designed to RTB to Earth as well as Mars.

Do I think they are doing tests for nothing?  No, of course I don't.  

Do I believe the excuse that it's not a big deal that the ships crashed?  No, I don't.  Based mostly on what they say to us what the goals were/are.

alfadriver (Forum Supporter) said:

Keith Tanner said:

alfadriver (Forum Supporter) said:

Keith Tanner said:

What's amazing  about this launch today is that it included a new concept in relighting and initiating the flip. Which was suggested, planned and put into practice in a month. Yes, the touchdown was a little harder than ideal and/or something went wrong with the legs that led to LOV, but that's an incredible pace of progress. 

???

That concept was part of SN8 and 9.  During the righting phase in SN8, it blew up (IIRC one of the engines didn't re-light), and SN9, it was not quite vertical.  Maybe I need to look back on those launches, but I very much remember it laying on it's side as it came to earth.  And then the engines fired starting to make it vertical again for landing.

Actually, that's not quite right- at 6:34 into the launch of SN8, it went upright, but one motor was not powerful enough for it to stop the momentum- 1:54:47 of this video https://www.youtube.com/watch?v=ap-BkkrRg-o&ab_channel=SpaceX

And the same thing happened for SN9, almost at the same moment relative to the launch- 6:26- (11:53 of this https://www.youtube.com/watch?v=_zZ7fIkpBgs&ab_channel=SpaceX)

SN8 suffered from fuel starvation during the flip. That was solved with a different header tank pressurization scheme.

SN9 had an engine that failed to relight so it didn't have the power needed to perform the flip. At that time, the plan was to light two engines, flip, then shut down one engine and land on the other. But with only one lit, it couldn't rotate enough. Raptors being a new engine design, they're a little harder to light than some of the others out there.

For SN10, they changed the process. Now they're lighting three engines, just in case. Then one is shut down (if applicable), two are used for the flip, one is shut down and it lands on one. That's a new concept that brings some real changes in control logic and dynamics. Remember that rocket engines don't idle, the minimum thrust is fairly significant and the craft is at minimum weight so things are pretty sporty with three engines lit. Plus you don't know in advance which of the three engines will be the "safety engine" that gets shut down or which of the engines will be used for the final landing burn. That's quite a change to make, and it all takes place in a part of the flight where you cannot get it wrong. And it worked. Three lit, two flipped, one landed but the vehicle took some damage.

They also managed to improve the minimum throttling of the Raptor engine at the same time, I believe - you can actually see the difference in the two engines on the way up. In a MONTH.

I think where we disagree is the impact on the test plan from the loss of the vehicle. SpaceX has never reflown a Starship prototype even if it landed successfully. Past behavior would indicate that the actual flight is the last step in its planned lifespan. They've saved three after test flights - one has been turned into a mounting point for cameras and telemetry. The others were all eventually dismantled for scrap.

I thought you were bringing up the entire maneuver of gliding back down  on its side and then trying to land upright was new.

Well, it IS. But it's not new for the SN10 launch. The actual method of performing the flip changed, though. It wasn't that obvious to see but it was a fairly fundamental modification that would have had some serious work behind it. And it may also be an illustration of some pretty good code underneath. There's a lot of closed loop feedback going on in that engine control.

alfadriver (Forum Supporter) said:

Do I think they are doing tests for nothing?  No, of course I don't.  

Do I believe the excuse that it's not a big deal that the ships crashed?  No, I don't.  Based mostly on what they say to us what the goals were/are.

Hopefully we don't have to outsource to the darkside (relevant XKCD):

In reply to Flynlow (FS) :

My point is that Space X isn't quite being totally forthcoming.  

What they are saying was the goal would cheaper and more effective to do ground based testing.  So there's a lot more to these launches than just going up to 10km, and coming back down on it's belly.  And I honestly think that landing is a lot bigger deal than how they are making it out to be.  So that the lack of landings is a bigger problem than how they are making it out to be.

If you watch the Lab Padre footage of the SN10 attempt, you can see that half of the landing legs don't lock into place. It keeps trying but they're just cycling. So that means that they probably only landed on half of the legs, which are designed to crumple and absorb impact. It's also possible that the landing was a bit harder than ideal and missing the legs would not help at all.

This one did land. It just took some damage, which meant that the pressurized oxygen tank eventually failed and it unzipped.

They do a bunch of ground based testing. Those Raptors are on the test stands pretty much constantly. They did multiple static fires in a short window earlier this year. Flaps flap, gimbals gimbal, tanks get pressurized and detanked. But eventually, you've gotta fly. Something you don't see is that they're also pushing the limits on the Falcon 9s too - they're expanding the entry speed envelope and the allowable windspeeds as time goes on, and they're also trying to see how many times they can really reuse one. Losing one of those is a bigger hit to the business, though, as they're an income-producing asset.

I was wrong about only using two engines for the flip - it looks like the current process is using three to flip before shutting down two. This also means a fair bit of driving around over the landing site. 

Keith Tanner said:

I was wrong about only using two engines for the flip - it looks like the current process is using three to flip before shutting down two. This also means a fair bit of driving around over the landing site. 

What the commentators mentioned yesterday was a 3 engine start up, and choose the best two.  

BTW, the speculation on what they are testing is exactly my point.  Which is why I don't believe that the lack of landings is 100% ok.  Acceptable, but it will push things back.

That was my understanding as well but the video shows a 3 engine start and flip, then a two engine shutdown. It's possible that one of those was not supposed to shut down quite so early which may have led to the slightly fast landing.

I thought it went down straight, so the landing struts failed, and then punctured the tanks.  Which also made me wonder where the water was coming from- had to been automated spraying for that purpose.

Previously, they failed because they couldn't get one to light for a two engined landing.  They MAY have planned to slow it down enough with the three to allow a one engined landing?

Who knows (well, they do), but you learn a lot more from failures than successes, and they are clearly learning and improving.  Next test might be 80%+ predicted success?

alfadriver (Forum Supporter) said:

I thought it went down straight, so the landing struts failed, and then punctured the tanks.  Which also made me wonder where the water was coming from- had to been automated spraying for that purpose.

There's a water flood system for the pad, they were trying to deal with the flames I think. There's no water on board. It was kind of funny to watch, the sprayer couldn't quite reach the damaged ship because it didn't hit the bullseye.

The tank didn't fail immediately. But it sure looks like it unzipped starting at the bottom at the low side, so it took the most damage there. Could have been a poke from a fully compressed landing leg.

I don't think they can hover the empty ship on three Raptors, the minimum thrust is too high. It's always been the plan to land on one but you need two in order to perform the flip, and failing to get both of those is what pancaked SN9.

Once again, Scott Manley has done a nice video utilizing footage from  a few different sources and is worth the watch. He thought it may have come in a bit hot which could indicate the second engine shut down earlier than expected. He also pointed out that something failed in the methane plumbing as those engines shut down after the flip which probably led to the eventual failure of the tanks and fire ball.

other than the armchair rocket surgery second guessing everything because internet people always know better than the engineers doing the work.... highly interesting and entertaining. 

Elon Musk on the hard landing. 

"Thrust was low despite being commanded high for reasons unknown at present, hence hard touchdown. We’ve never seen this before.

Next time, min two engines all the way to the ground & restart engine 3 if engine 1 or 2 have issues."

SN11 rolling out to the pad today, so another 3-4 weeks until launch again?

bobzilla said:

other than the armchair rocket surgery second guessing everything because internet people always know better than the engineers doing the work.... highly interesting and entertaining. 

When the engineers doing the work aren't saying what happened, pretty much all you've got is educated guesses based on the available video.  Scott Manley has had a pretty good track record with such in the past.

But the engineers doing the work have now said what the problem was :)

codrus (Forum Supporter) said:

bobzilla said:

other than the armchair rocket surgery second guessing everything because internet people always know better than the engineers doing the work.... highly interesting and entertaining. 

When the engineers doing the work aren't saying what happened, pretty much all you've got is educated guesses based on the available video.  Scott Manley has had a pretty good track record with such in the past.

 

I wasn't referring to offsite people