-
#700 by IanThePineapple on 31 Dec, 2017 16:37
-
Antares didnt activate the flight termination system, so it came down as two full tanks.My guess is it's around 30 seconds, give or take 10 seconds. That's the moment on the mission clock that I'll be watching for ...
If you look at this fabulous Falcon Heavy launch simulation vid by Zach and freeze it at the right time you should get an idea of when the stack wouldn't tumble back on the launch pad ... Looks like around 15 seconds elapsed mission time to me.
Thanks Oersted for the link to that GREAT simulation.
Zach, could you run that with a view from the side, showing the IIP during the first 30 seconds? (in other words, keep the ground in view, and we can use the stack height to estimate horizontal distance.
Per the simulation, the pitchover doesn't even start until +15 seconds, so +15 secs is definitely NOT long enough (it will fall straight back down, a la Antares as I mentioned). By my guess, looking at downrange distance, altitude and apogee, it's not until about +25 seconds (+/- 5 secs) that the IIP would clear the LC-39A fenceline, 400 meters from the pad. We are converging on a solution
Hopefully SpaceX will be faster on the trigger if it happens.
Propellant dispersal is really important in these situations.
Every gallon of propellant burned in the air is a gallon of propellant not burning the pad up. -
#701 by nacnud on 31 Dec, 2017 16:51
-
It's 975m from the pad due east to the ocean. Assuming that if the vehicle fails it travels ballistically from that point onwards the sim suggests that any time after 30 seconds the pad is safe. After 40 seconds or so the remains would land in the water and after a minute the vehicle is flying over the sea.
Sources: the above linked sim and the ruler buried in the left click menu of google maps. -
#702 by Space Ghost 1962 on 31 Dec, 2017 16:57
-
If you look at this fabulous Falcon Heavy launch simulation vid by Zach and freeze it at the right time you should get an idea of when the stack wouldn't tumble back on the launch pad:
Watch the bouncing IIP as it suddenly has a horizontal component. I'd say a half minute.
Looks like around 15 seconds elapsed mission time to me.
(Zach's video does a fine job. If I were to add anything it would be corrected sways/bends in combined flight, then minor twists during booster sep. Great work!)
One might fire it up on actual FH launch in side by side - then you'll notice the little differences more easily in the actual. -
#703 by Greg Hullender on 31 Dec, 2017 16:59
-
I thought the center core was only going to fire until shortly after liftoff and then reignite when the side boosters separated. Has that changed?
-
#704 by jak Kennedy on 31 Dec, 2017 17:05
-
I seem to recall some rockets start moving actively move away from directly above the pad as soon as they clear the tower. ie sliding in a horizontal direction for a few seconds so that any debris would clear the pad.
-
#705 by mme on 31 Dec, 2017 17:16
-
The center core will throttle downn but it will not shutdown its engines.
-
#706 by ZachS09 on 31 Dec, 2017 17:46
-
My guess is it's around 30 seconds, give or take 10 seconds. That's the moment on the mission clock that I'll be watching for ...
If you look at this fabulous Falcon Heavy launch simulation vid by Zach and freeze it at the right time you should get an idea of when the stack wouldn't tumble back on the launch pad ... Looks like around 15 seconds elapsed mission time to me.
Thanks Oersted for the link to that GREAT simulation.
Zach (was that ZachS09?), could you run that with a view from the side, showing the IIP during the first 30 seconds? (in other words, keep the ground in view, and we can use the stack height to estimate horizontal distance.
Per the simulation, the pitchover doesn't even start until +15 seconds, so +15 secs is definitely NOT long enough (it will fall straight back down, a la Antares as I mentioned). By my guess, looking at downrange distance, altitude and apogee, it's not until about +25 seconds (+/- 5 secs) that the IIP would clear the LC-39A fenceline, 400 meters from the pad. We are converging on a solution
It was a different Zach that ran that simulation. I had nothing to do with it. -
#707 by Oersted on 31 Dec, 2017 17:59
-
Watch the bouncing IIP as it suddenly has a horizontal component. I'd say a half minute.
Yup, true, 15 seconds is a bit too early. At half a minute into the flight the projected path has moved quite a bit away from the launch complex, in my estimation. Still a bit too early for comfort, though!
Here is a screenie of the dots you need to keep an eye on...
(Please note that the lift-off is 20 seconds in from the beginning of the movie. This screenie is consequently 11 seconds into the flight, where - if the engines were to stop immediately - the rocket would just rise by a bit more than its own length before it would fall back down.)
-
#708 by nacnud on 31 Dec, 2017 18:07
-
It's easier just to keep a note of the downrange distance displayed in the top left of the sim.
-
#709 by speedevil on 31 Dec, 2017 19:01
-
It's easier just to keep a note of the downrange distance displayed in the top left of the sim.
(altitude/downrange) gives a rough estimate of the current launch angle from vertical.
perigee*2/(altitude/downrange) is a conservative estimate of how far it will travel horizontally.
At 100m altitude, it's gone 20m, apogee is 87m.
200 24 233 55
400 24 540 64
1000 36 1562 112
1500 54 2300 165
2500 124 4200 416
3500 200 6000 685
5000 350 8400 1176
So, ~5km, ~40s in means the IIP is likely to be on water.
But, I don't believe this at all for a few reasons. Unless the stage just stops thrusting, and falls intact, it's really, really not going to be ballistic.
Once apogee hits ~1km or so, 12s in, if the FTS triggers, you're getting shards, not any part of the rocket.
So, hold your breath, and when you let it out it's fine.
If it explodes, once you're over a few hundred meters, most of the cloud of debris is going to be shards of sooty fuel tank, with the COPVs and engines falling intact.
Very small tweaks to the trajectory that are essentially free variables and have no impact on launch performance can change the impact dramatically.
Still, a minute in is certainly pad safe,
-
#710 by wannamoonbase on 31 Dec, 2017 21:26
-
If it makes it past the first few seconds it’s likely fine till staging.
Then, hey, who knows, if it survives staging the side boosters I think it will be completely successful. -
#711 by zlynn1990 on 31 Dec, 2017 22:29
-
Oersted I'm the Zach who made the sim Most of that trajectory was done by OneSpeed, and he has the gravity turn starting at 10 seconds. Take the entire thing was a large grain of salt because there are a lot of unknowns right now.
-
#712 by Surfdaddy on 31 Dec, 2017 23:11
-
If it makes it past the first few seconds it’s likely fine till staging.
Then, hey, who knows, if it survives staging the side boosters I think it will be completely successful.
I want to believe you, but I've seen a lot of launch videos where problems happen in your "safe" time period. -
#713 by Oersted on 31 Dec, 2017 23:27
-
Oersted I'm the Zach who made the sim
Most of that trajectory was done by OneSpeed, and he has the gravity turn starting at 10 seconds. Take the entire thing was a large grain of salt because there are a lot of unknowns right now.
Yes, I realise it is a good guess more than anything else. Nevertheless, splendid effort. Thanks so much for putting the vid together. It is clear that a lot of love went into it. Kudos and happy new year! -
#714 by meekGee on 01 Jan, 2018 00:56
-
My guess is it's around 30 seconds, give or take 10 seconds. That's the moment on the mission clock that I'll be watching for ...
If you look at this fabulous Falcon Heavy launch simulation vid by Zach and freeze it at the right time you should get an idea of when the stack wouldn't tumble back on the launch pad ... Looks like around 15 seconds elapsed mission time to me.
Thanks Oersted for the link to that GREAT simulation.
Zach (was that ZachS09?), could you run that with a view from the side, showing the IIP during the first 30 seconds? (in other words, keep the ground in view, and we can use the stack height to estimate horizontal distance.
Per the simulation, the pitchover doesn't even start until +15 seconds, so +15 secs is definitely NOT long enough (it will fall straight back down, a la Antares as I mentioned). By my guess, looking at downrange distance, altitude and apogee, it's not until about +25 seconds (+/- 5 secs) that the IIP would clear the LC-39A fenceline, 400 meters from the pad. We are converging on a solution
Usually the IIP is the thing to watch, but this assumes the rocket continues as an inert solid body - as if the engines are shut off and nothing else breaks.
In an early failure scenario, two things will likely happen - off-axis thrust will move the IIP very quickly in some random directions, and then FTS will make the whole thing "stop cold" and fall straight down, but influenced by wind.
At some altitude, the liquid fuel will have time to burn on the way down, or disperse enough that it doesn't pool on the ground, and that will already be a "safe" situation. -
#715 by joek on 01 Jan, 2018 02:44
-
Usually the IIP is the thing to watch, but this assumes the rocket continues as an inert solid body - as if the engines are shut off and nothing else breaks.
IIP is a point spread function based on numerous parameters. You can only state "probability of IIP at point X,Y is Z based on parameters T, P, Q, R, ...". Constrain the parameters sufficiently and you can assert that it "continues as an inert solid body". However, most real-world models (probabilistic-time-based) do not apply such constraints or make such assumptions--unless the parameters are well bounded and understood (e.g., close to lift off). Which is why the models tend to vary their parameters over the flight path (in particular, probability vs. time, as spread-uncertainty tends to increase with time). -
#716 by meekGee on 01 Jan, 2018 05:16
-
I forget where I saw a realtime IIP displayed, but I got the impression it was simply ballistic.Usually the IIP is the thing to watch, but this assumes the rocket continues as an inert solid body - as if the engines are shut off and nothing else breaks.
IIP is a point spread function based on numerous parameters. You can only state "probability of IIP at point X,Y is Z based on parameters T, P, Q, R, ...". Constrain the parameters sufficiently and you can assert that it "continues as an inert solid body". However, most real-world models (probabilistic-time-based) do not apply such constraints or make such assumptions--unless the parameters are well bounded and understood (e.g., close to lift off). Which is why the models tend to vary their parameters over the flight path (in particular, probability vs. time, as spread-uncertainty tends to increase with time).
You're correct that a true IIP estimate is a stochastic thing. -
#717 by gadgetmind on 01 Jan, 2018 09:47
-
I wonder how FH would handle an engine failure on a side booster? I guess it would need to throttle down the other side? Throw in scenarios of failures of two engines in various configurations and it could all get complicated.
-
#718 by SLC on 01 Jan, 2018 13:57
-
I wonder how FH would handle an engine failure on a side booster? I guess it would need to throttle down the other side? Throw in scenarios of failures of two engines in various configurations and it could all get complicated.
Would it be too off-topic to give this link to the last time a rocket with this many engines made its (short) first flight - the Soviet N1 moon rocket 49 years ago next month:
http://www.russianspaceweb.com/n1_3l.html
On that occasion the KORD control system did indeed diagnose an engine failure on one side, it did shut down the symmetrically opposite engine to compensate, and then it all got complicated ... -
#719 by speedevil on 01 Jan, 2018 14:11
-
I wonder how FH would handle an engine failure on a side booster? I guess it would need to throttle down the other side? Throw in scenarios of failures of two engines in various configurations and it could all get complicated.
Maybe.
From memory, it has been said that the first launch will be at 82% thrust.
This is considerably more than 9/8ths, so you can lose certain engines, and simply throttle up on others with no external changes in thrust vector.
For example, lose an inboard engine on a booster, and if you throttle the remaining outboard ones to 92%, the centre of thrust moves out from the core by perhaps half an engine diameter.
If you throttle the outer engine to nearly minimum, the ones next to the core to ~105%, then it may not move at all.
Only once you can no longer keep a booster or cores thrust vector within narrow limits would you need to make changes to the thrust profile of the stack as a whole.Would it be too off-topic to give this link to the last time a rocket with this many engines made its (short) first flight - the Soviet N1 moon rocket 49 years ago next month:
Yes.
More seriously, N1 failures are not comparable in so many ways, as overall QC and management was so terrible that it loses all meaning as a reasonable comparator.
