SpaceX FH : Falcon Heavy Demo : Feb 6, 2018 : Discussion Thread 2

[Kaputnik]

Here is how I see the landing video. The ballistic trajectory has horizontal speed. Pulling numbers out of the air that could be 100[kph] or 110[kph] (undershoot or overshoot) but I can't see the difference. Either way the landing burn needs to halt that motion at or above the X. That 10% difference can't be discerned by the human eye.

Use logic. If the impact spot was before the landing burn, the booster would have to *add* horizontal velocity to hit it. But the booster never changes orientation or gimbal to do add horizontal movement to the right. It is braking its horizontal velocity the ENTIRE time.

Not necessarily so. It just needs to be killing off a greater proportion of its vertical speed than its horizontal, thus prolonging the time until touchdown and allowing the booster to drift onto the pad.
I agree that this does not appear to be the case for the FH demo flight, but it may have been used on previous landings. Would need further video analysis to say I've way or the other.

[tater]

Time matters. With any X velocity, longer flight time moves it in X. You buy flight time by slowing sink rate (in Z).

Booster can be moving horizontally West (inland), and still have an impact point that is off the beach East. Burning vertically slows the descent rate, and gains time for the impact point (still moving West) to cross the LZ. Then you kill all the horizontal velocity over the LZ.

That's a way that boostback can form a trajectory that is safely offshore, then by killing some of the sink rate, it can move it to the LZ only after the engines restart.

I've not bothered to look at the trajectories, so I have no idea if this is what they do, but it's entirely possible.

[Jcc]

Time matters. With any X velocity, longer flight time moves it in X. You buy flight time by slowing sink rate (in Z).

Booster can be moving horizontally West (inland), and still have an impact point that is off the beach East. Burning vertically slows the descent rate, and gains time for the impact point (still moving West) to cross the LZ. Then you kill all the horizontal velocity over the LZ.

That's a way that boostback can form a trajectory that is safely offshore, then by killing some of the sink rate, it can move it to the LZ only after the engines restart.

I've not bothered to look at the trajectories, so I have no idea if this is what they do, but it's entirely possible.

May be ways you test that in simulation.
http://moonlander.seb.ly/

[Kabloona]

Here's a thought experiment that may help.

First, imagine the ballistic trajectory of the incoming stage with no landing burn.

Now, imagine that the booster tries to follow the identical trajectory, ie follow that exact flight path, while executing the landing burn. Is that even possible?

I would suspect that while possible, it's not propellant-optimal, because of gravity losses while trying to follow the arc exactly, and not physically practical because you end up landing with a non-zero horizontal velocity.

The propellant-optimal landing burn truncates the horizontal component of the arc, causing the flight path to "go vertical" sooner than it would ballistically (which would, in fact, be never).

Since we can infer SpaceX uses a propellant-optimal flight path to land, by the above logic it's not possible that the powered flight path follows the un-powered ballistic arc.

Ergo, the ballistic path results in overshoot.

[Jcc]

Nope, the ballistic trajectory results in undershoot. The landing burn provides much more vertical thrust than horizontal, so lacking that the stage falls much faster, landing short of the target.

[Lars-J]

Nope, the ballistic trajectory results in undershoot. The landing burn provides much more vertical thrust than horizontal, so lacking that the stage falls much faster, landing short of the target.

Nope. Look at the image. Compare the thrust axis with a hypothetical straight line between the booster positions. Play KSP.

[fthomassy]

Here's a thought experiment that may help.

First, imagine the ballistic trajectory of the incoming stage with no landing burn.

Now, imagine that the booster tries to follow the identical trajectory, ie follow that exact flight path, while executing the landing burn. Is that even possible?

I would suspect that while possible, it's not propellant-optimal, because of gravity losses while trying to follow the arc exactly, and not physically practical because you end up landing with a non-zero horizontal velocity.

The propellant-optimal landing burn truncates the horizontal component of the arc, causing the flight path to "go vertical" sooner than it would ballistically (which would, in fact, be never).

Since we can infer SpaceX uses a propellant-optimal flight path to land, by the above logic it's not possible that the powered flight path follows the un-powered ballistic arc.

Ergo, the ballistic path results in overshoot.

Emphasis mine ... if you follow the path then you zero horizontal when you zero vertical.

[Kabloona]

Nope, the ballistic trajectory results in undershoot. The landing burn provides much more vertical thrust than horizontal, so lacking that the stage falls much faster, landing short of the target.

It provides more vertical thrust because at that point in the trajectory, the flight path angle is more vertical than horizontal, so there's more vertical velocity to kill.

The question is, what's the stage's attitude relative to the unpowered ballistic trajectory during the burn. Depending on the answer to that question, you can imagine two different scenarios: one powered path that goes "above" the ballistic trajectory, and one that goes "below" it.

So I guess you're imagining the powered path that goes "above" the ballistic trajectory, and I'm imagining the one that goes "below" it.

And I believe Lars-J's photo analysis shows the powered flight path going "below" the ballistic trajectory. Apply this logic, also from Lars-J, two posts above:

Look at the image. Compare the thrust axis with a hypothetical straight line between the booster positions. Play KSP.

[Kabloona]

Emphasis mine ... if you follow the path then you zero horizontal when you zero vertical.

No, because the ballistic trajectory isn't vertical at landing. There's still a small horizontal component.

[fthomassy]

Emphasis mine ... if you follow the path then you zero horizontal when you zero vertical.

No, because the ballistic trajectory isn't vertical at landing. There's still a small horizontal component.

The angle of the line is not relevant. Stop is stop.

[Kabloona]

Emphasis mine ... if you follow the path then you zero horizontal when you zero vertical.

No, because the ballistic trajectory isn't vertical at landing. There's still a small horizontal component.

The angle of the line is not relevant. Stop is stop.

You don't want to land a booster at an angle, which you'd have to do if you're still trying to zero out horizontal velocity at touchdown.

You want to land it vertically, with zero horizontal velocity before you get close to touchdown. That's what F9 does.

[fthomassy]

You don't want to land a booster at an angle. You want to land it vertically. That's why F9 zeroes out the horizontal velocity well above touchdown.

Very much agree with that. The actual rocket does not follow a ballistic path to the pad, regardless of speed along the path.

[docmordrid]

Looks like the booster undershoots the pad/ASDS then diverts after landing burn startup. If the burn fails, it takes a swim.

http://www.spacex.com/news/2015/06/24/why-and-how-landing-rockets

[georgegassaway]

Anyone claiming to eyeball video evidence of overshoot or undershoot is fooling themselves.

Your reply reminded me of the old Marx Brothers joke:  ”Who are you going to trust, me or your lying eyes?" 
https://tinyurl.com/y9vm6b8x

So, included is a version with lines added.  The boosters did come in on a path that would have them overshoot a few hundred feet if they came in ballistically. I believe my eyes, if you don’t fine.



Admittedly, the grid fins would do as much as possible to try to keep the booster from overshooting, so the impact point wold not be as far out as projected. But I think it is highly improbable that they could maneuver the booster enough in a few seconds to avoid overshooting altogether. Doesn’t need that kind of maneuverability for the flight phases the fins are designed for. Also, in the case of a planned 1-3-1 landing burn where the outer two do not ignite, the center engine thrust vector would also help to reduce the overshoot.

But there is definitely a ballistic overshoot path when coming in for the landing burn (and this should not be confused with the intentional offshore undershoot that the boostback burn aims for, which is extended across the shore after the re-entry burn, by the grid fin aerodynamic steering).

And again, for the Xth time, this is probably why the center core missed the ASDS by a few hundred feet or whatever. Or do you have some alternative information on why the center core missed the ASDS?

[Kabloona]

We've seen that SpaceX trajectory graphic (in docmordrid's post) before, so that's how I imagined the trajectory design. But Lars-J's photo analysis upthread seems to show that, without the landing burn, there would be some overshoot of the pad.

It's possible that they're getting more horizontal component of lift due to improved grid fin control authority that enables a significant angle of attack from the rocket body, moreso than they expected when that graphic was made.

In which case, they may have changed their trajectory design to get more cross-range velocity from the aerodynamics, which then would need to be canceled by the horizontal component of the landing burn.

[Lars-J]

Your reply reminded me of the old Marx Brothers joke:  ”Who are you going to trust, me or your lying eyes?" 
https://tinyurl.com/y9vm6b8x

LOL, so true for this thread. 

But there is definitely a ballistic overshoot path when coming in for the landing burn (and this should not be confused with the intentional offshore undershoot that the boostback burn aims for, which is extended across the shore after the re-entry burn, by the grid fin aerodynamic steering).

Yes! This is the crucial difference... It is fully believable that the *bostback burn* ends with the ballistic impact point short of the destination. The grid fins are then used to lift/glide the impact point forward until it just exceeds the landing point, and the 1-3-1 short and drastic landing burn is able to take care of the rest.

But the evidence certainly shows that when the landing burn starts, the impact point is beyond the landing spot. It is also the most efficient use of landing propellant. The landing can be viewed as a launch in reverse, they are in effect doing a reverse gravity turn.

Note also that the new 1-3-1 landing burn is far shorter. The old 1 engine landing burn started MUCH higher with more control authority. With the 1-3-1 burn, they save propellant but it is going to overshoot if the engines don't light right.

[Kabloona]

The grid fins are then used to lift/glide the impact point forward until it just exceeds the impact point, and the 1-3-1 short and drastic landing burn is able to take care of the rest.

This is the crucial disctinction, I believe. The guided aerodynamic portion of the flight is when the IIP goes from undershoot to slight overshoot.

So both undershoot and overshoot are possible failure modes, depending on the portion of the trajectory being considered.

[AncientU]

So is impacting at 300mph.

[Jcc]

Anyone claiming to eyeball video evidence of overshoot or undershoot is fooling themselves.

Your reply reminded me of the old Marx Brothers joke:  ”Who are you going to trust, me or your lying eyes?" 
https://tinyurl.com/y9vm6b8x

So, included is a version with lines added.  The boosters did come in on a path that would have them overshoot a few hundred feet if they came in ballistically. I believe my eyes, if you don’t fine.

If you notice from the picture, the stage is firing its landing burn by this point.
The ballistic path includes gravity, which is accelerating the stage downward at 10m/s2 (approx). Without the landing burn, it hits the ocean.

[Kabloona]

The ballistic path includes gravity, which is accelerating the stage downward at 10m/s2 (approx). Without the landing burn, it hits the ocean.

And the ballistic path also includes a horizontal velocity component that is not affected by gravity.

From this video it looks clear to me that the landing burn begins when the IIP's are already over land.

Not to mention that, even if the landing burn fails, the grid fins are still trying to guide the stage aerodynamically to the LZ coordinates.

So it looks to me that, even if the landing burn totally fails, the stage is falling onto land, not into the ocean.