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

[JonathanD]

A 3 engine landing is more fuel efficient than 1-3-1 burn. Faster "slam on the brakes" means less dV lost to gravity.

Oh I understand the concept, but they had loads of extra capability on this flight and given the extra layer of complexity with a 3-engine landing burn I'd assume they'd only do it when necessary.

[cscott]

'If you're not having explosions, you're not testing hard enough'. (Paraphrased Elon, I think).

Pretty sure that was Gwynne Shotwell who said that. It was memorable to me because my previous mental image had her as the sober safe "business" person.

[Lar]

'If you're not having explosions, you're not testing hard enough'. (Paraphrased Elon, I think).

Pretty sure that was Gwynne Shotwell who said that. It was memorable to me because my previous mental image had her as the sober safe "business" person.

Everything is relative.... once in a great while I am not the most crazed LEGO collector in the room, for example. 

Compared to Elon, she **is** the sober safe "business" person.

[chipguy]

My understanding is that the booster would under shoot, not overshoot, the pad/ship if the landing burn impulse fell short.
This is because the booster is approaching on a ballistic trajectory that falls short of the pad. Most of the velocity is vertical, and some is horizontal. In making the landing burn, the vertical rate is rapidly reduced, prolonging the length of time that the booster remains airborne, and allowing it to cover a greater horizontal distance and thus reaching the pad.

I actually made a simple sim in scratch a few years ago that captures this pretty well.

It depends how the landing burn is done.

A zero angle of attack thrust (i.e. thrust is kept aligned with velocity vector) will slow the horizontal
and vertical components in proportion to the velocity components but the vertical component has
gravity loss so over time the velocity vector bends more towards vertical. This causes the landing
spot to be short of the no burn impact point.

A pure vertical burn (tilt the stage so it standing straight vertical despite the horizontal component
of velocity) will reduce the vertical velocity component while leaving the horizontal component the
same. This means it takes longer to reach the ground so the landing spot is further along the flight
path than the no burn impact point. Of course you still need cancel the horizontal component just
before landing.

[Herb Schaltegger]

At the start of the landing burn, grid fins and body lift provide substantial control over the trajectory. If there was a complete engine failure of any sort, grid fins might easily divert a returning booster by several hundred yards at least, regardless of the IIP at the point of (attempted) ignition. Further, as atmospheric density increases drastically with decreasing altitude, so too will drag. I don’t believe the IIP at the time of planned engine ignition would be on-shore at all, regardless of what it looks like. For any number of reasons, a failure to land safely at LZ-1/2 is far better to fall short than go long.

[fthomassy]

My understanding is that the booster would under shoot, not overshoot, the pad/ship if the landing burn impulse fell short.
This is because the booster is approaching on a ballistic trajectory that falls short of the pad. Most of the velocity is vertical, and some is horizontal. In making the landing burn, the vertical rate is rapidly reduced, prolonging the length of time that the booster remains airborne, and allowing it to cover a greater horizontal distance and thus reaching the pad.

I actually made a simple sim in scratch a few years ago that captures this pretty well.

That may be your understanding, but it does not match reality of what they are doing. (It would also waste propellant) Watch several landings and you'll see. (Did you watch the video I linked to?)

Or do you think that the boosters in the video would have undershot the landing area without the landing burn?

Can't tell from looking at the video. However, overshoot of RTLS would be the waste of fuel and has greater risk to life and property. The few diagrams I find in a simple search show either direct to pad/ASDS or clearly illustrate ballistic undershoot.

[JonathanD]

Can't tell from looking at the video. However, overshoot of RTLS would be the waste of fuel and has greater risk to life and property. The few diagrams I find in a simple search show either direct to pad/ASDS or clearly illustrate ballistic undershoot.

That is my recollection regarding RTLS as well, I remember Elon talking about it specifically.  Cruise missiling into LZ-1 is not good optics!

[Lars-J]

My understanding is that the booster would under shoot, not overshoot, the pad/ship if the landing burn impulse fell short.
This is because the booster is approaching on a ballistic trajectory that falls short of the pad. Most of the velocity is vertical, and some is horizontal. In making the landing burn, the vertical rate is rapidly reduced, prolonging the length of time that the booster remains airborne, and allowing it to cover a greater horizontal distance and thus reaching the pad.

I actually made a simple sim in scratch a few years ago that captures this pretty well.

That may be your understanding, but it does not match reality of what they are doing. (It would also waste propellant) Watch several landings and you'll see. (Did you watch the video I linked to?)

Or do you think that the boosters in the video would have undershot the landing area without the landing burn?

Can't tell from looking at the video. However, overshoot of RTLS would be the waste of fuel and has greater risk to life and property. The few diagrams I find in a simple search show either direct to pad/ASDS or clearly illustrate ballistic undershoot.

Sigh. No. What they are doing is the video is the optimal use of propellant. Those diagrams you have seen have not come from SpaceX, they are from "informed" (ironic quotes) speculation.

[whitelancer64]

A 3 engine landing is more fuel efficient than 1-3-1 burn. Faster "slam on the brakes" means less dV lost to gravity.

Oh I understand the concept, but they had loads of extra capability on this flight and given the extra layer of complexity with a 3-engine landing burn I'd assume they'd only do it when necessary.

SpaceX is still experimenting with the landings. Pushing corner cases and getting more flight data on the limits of the first stage's capability. Since core recovery was not crucial (it wasn't going to fly again either way), they could afford to run the risk of losing it.

[Prettz]

A 3 engine landing is more fuel efficient than 1-3-1 burn. Faster "slam on the brakes" means less dV lost to gravity.

Boostback burn because sending out the ASDS 200 km is easier than sending it out 600 km (numbers pulled from the aether).

That's no answer. The question wasn't "what is a 3 engine landing for", it was "why did they use one".

Regarding the boostback burn, that's too glib an answer. I'm sure there's a reason they used it that isn't "we don't have to send the barge out as far."

[mark_m]

My understanding is that the booster would under shoot, not overshoot, the pad/ship if the landing burn impulse fell short.
This is because the booster is approaching on a ballistic trajectory that falls short of the pad. Most of the velocity is vertical, and some is horizontal. In making the landing burn, the vertical rate is rapidly reduced, prolonging the length of time that the booster remains airborne, and allowing it to cover a greater horizontal distance and thus reaching the pad.

I actually made a simple sim in scratch a few years ago that captures this pretty well.

That may be your understanding, but it does not match reality of what they are doing. (It would also waste propellant) Watch several landings and you'll see. (Did you watch the video I linked to?)

Or do you think that the boosters in the video would have undershot the landing area without the landing burn?

Can't tell from looking at the video. However, overshoot of RTLS would be the waste of fuel and has greater risk to life and property. The few diagrams I find in a simple search show either direct to pad/ASDS or clearly illustrate ballistic undershoot.

Sigh. No. What they are doing is the video is the optimal use of propellant. Those diagrams you have seen have not come from SpaceX, they are from "informed" (ironic quotes) speculation.

My recollection matched Kaputnik's (that the IIP is initially an undershoot of the landing zone, with a divert before landing), so I tried to do some research. I came up with a more ASDS-like profile diagram (with pencils and the Empire State Building) from ORBCOMM-2 which does show an undershoot. (See the full launch video at about the 32:15 mark.) Not sure of course if that's really authoritative, or if it applies to an LZ1 landing, but it does seem better to me from a safety point of view to have the IIP offshore as late as possible in the process.

Edit: Didn't realize the video would be embedded, so reformatted a bit.

[whitelancer64]

A 3 engine landing is more fuel efficient than 1-3-1 burn. Faster "slam on the brakes" means less dV lost to gravity.

Boostback burn because sending out the ASDS 200 km is easier than sending it out 600 km (numbers pulled from the aether).

That's no answer. The question wasn't "what is a 3 engine landing for", it was "why did they use one".

Regarding the boostback burn, that's too glib an answer. I'm sure there's a reason they used it that isn't "we don't have to send the barge out as far."

The "why" is "because it's more fuel efficient." Also, per my most recent post, SpaceX is still pushing the limits for the landings. They can afford to risk losing a booster that wasn't going to fly again anyway.

Occam's Razor. Simplest explanation most likely to be correct. Also, I'm reasonably sure (not going to go back to every single mission to check) that all ASDS landings have had boostback burns. If not all, then most.

[georgegassaway]

My understanding is that the booster would under shoot, not overshoot, the pad/ship if the landing burn impulse fell short.
This is because the booster is approaching on a ballistic trajectory that falls short of the pad. Most of the velocity is vertical, and some is horizontal. In making the landing burn, the vertical rate is rapidly reduced, prolonging the length of time that the booster remains airborne, and allowing it to cover a greater horizontal distance and thus reaching the pad.

Certainly true for the ballistic path from the end of boostback, thru the re-entry burn.  Note that in several of the RTLS landings with good  based camera views , that after the re-entry burn you can sometimes see the booster change its angle of attack (grid fins steering it a lot), to extend the path from ballistic impact at sea, towards the landing zone. Which is pretty much what you are referring to.

 By the time of the landing burn beginning, the descending booster angle is something like 15-20 degrees from vertical. The landing burn does more to kill that horizontal component to zero than the grid fins are likely be able to, for that velocity,  aerodynamic maneuvering capability, and altitude left. So, failure to ignite the outer two for  1-3-1 landing, the booster would crash by overshooting, not undershooting. 

Which back to my original reply, was  a strong reason why the FH core could have missed the ASDS a few hundred feet - overshoot (if there is NOT programming for the F9 to self-ditch in a hopeless landing attempt).

There's better videos but I did not have time to search much, so here is a landing compilation I found and took two screenshots from, during a segment beginning at 1:05.  The first screenshot is well into the burn, at about the time that the outer two were ignited. They are tilted a lot. You can see where their horizontal overshoot would have been more inland, by a few hundred feet, had the landing burn thrust not been able to stop that horizontal momentum. Indeed they go from that angle to very close to vertical pretty quickly, by the time of the second screenshot.

That's going to be the nature of the 1-3-1 RTLS landings, the incoming diagonal path HAS to aim to overshoot the LZ more than for a single engine burn.  Don't have time now, but imagine a straight line projected from the centerline of the lower booster, to the ground. That line would end up a few hundred feet inland (to the right), but still well inside of the LC-13 area.

[fthomassy]

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

[Kabloona]

Don't have time now, but imagine a straight line projected from the centerline of the lower booster, to the ground. That line would end up a few hundred feet inland (to the right), but still well inside of the LC-13 area.

I'm not taking sides here, just cautioning that the flight path is not necessarily parallel to the booster centerline. In fact, it's most likely not perfectly aligned until the booster goes vertical just above the pad.

Until that point in the flight path, the booster is partially slip-sliding using its body as a lifting cylinder in flight, which requires the booster to have a non-zero angle of attack in order to generate lift, and that angle of attack may continue to be non-zero even after the landing burn starts.

The only accurate way to judge flight path angle is to watch the video and plot the booster's location frame by frame, and make a rough extrapolation from it.

[fthomassy]

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.

[Lars-J]

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

Sigh. How about this image, then? A series of screen caps, adjusted and overlaid with matching landmarks to adjust for camera movement and rotation.

Don't have time now, but imagine a straight line projected from the centerline of the lower booster, to the ground. That line would end up a few hundred feet inland (to the right), but still well inside of the LC-13 area.

I'm not taking sides here, just cautioning that the flight path is not necessarily parallel to the booster centerline. In fact, it's most likely not perfectly aligned until the booster goes vertical just above the pad.

Until that point in the flight path, the booster is partially slip-sliding using its body as a lifting cylinder in flight, which requires the booster to have a non-zero angle of attack in order to generate lift.

The only accurate way to judge flight path angle is to watch the video and plot the booster's location frame by frame, and make a rough extrapolation from it.

Look at the image. Look at it. The first frame (for the trailing booster) starts just after engine ignition.

[fthomassy]

The only accurate way to judge flight path angle is to watch the video and plot the booster's location frame by frame, and make a rough extrapolation from it.

Actually, I think you'd need to plot the ballistic path from before the landing burn but even that could be false due to grid find control.

I don't think anyone can win an over/under argument from video analysis.

[Lars-J]

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.

[Lars-J]

The only accurate way to judge flight path angle is to watch the video and plot the booster's location frame by frame, and make a rough extrapolation from it.

Actually, I think you'd need to plot the ballistic path from before the landing burn but even that could be false due to grid find control.

I don't think anyone can win an over/under argument from video analysis.

I give up.