SpaceX Falcon 9 : PAZ & Microsat 2a/2b : SLC-4E : Feb 22, 2018 : DISCUSSION

[gongora]

I still think that is very unlikely.

[ugordan]

Just rewatching the webcast, there's something about that one pad camera view that makes it clear the two ports on the interstage are spewing out liquid oxygen from the MVac prechill. It's distinctly blue in color compared to the white of water vapor and liquid oxygen really is slightly bluish. It's really noticeable in the last 3 or so minutes of the countdown. Same with the big transporter/erector residual dump after T-1:30.

[Nomadd]

The fairing recovery vessel should be under manual control unless you really trust the parachutes and wind to land within a couple meters of a set position (which doesn't seem very likely).

I've been on ships without dynamic positioning systems trying to maintain accurate positions manually. There's usually much cussing involved and it's not something you want to do if you have a choice.

[Lar]

I guess why wouldn't the ship move to where the parachute seems to be guiding the fairing, rather than hold a fixed position? Baseball analogy, the fielder moves to get under the ball as they understand the ball's likely arrival point better and better.

[ZachS09]

That’s amazing that the fairing was brought back to port despite missing the net by a few hundred meters!

I thought that Mr. Steven would leave it floating in the water just because the boat didn’t catch the fairing.

[Jdeshetler]

Helodriver's stabilized video.

So the upper is active one?

My video of the launch: https://forum.nasaspaceflight.com/downloads/l2/PAZ%20Launch.mp4

Fairing RCS jets start at 3:20 in the video both halves can be seen jetting.

2nd stage at left, fairings middle, 1st stage at right.

[speedevil]

I guess why wouldn't the ship move to where the parachute seems to be guiding the fairing, rather than hold a fixed position? Baseball analogy, the fielder moves to get under the ball as they understand the ball's likely arrival point better and better.

I think a better analogy may be Quiddich, and the Snitch.

The fairing descent (in the face of winds and control systems) is much less predictable to outside observers, if they are not purely assuming it will land in the designated spot.

[cscott]

(Elon's tweet suggests to some that the boat is manuveuring to get under the parachute; it could also be read as "the amount of cross-range compensation possible on the chute was not enough given the descent time", ie discussing rates of x/y versus rate of z, with the implication being it was easier to decrease dz than increase dx and dy.)

So our opinions agree? The boat hold position and the parachute maneuvers.

Honestly, as indicated by my commentary on Elon's tweet, I don't know. My opinion is that the starting point for the recovery attempts was probably the same as for ASDS: boat steady, parachute maneuvers.  But given that things aren't *quite* working out, perhaps they will experiment with some tweaks.  Depends on what their position-versus-time and wind speed data look like.

For example, if the dispersions are primarily in the initial descent from altitude so that by the time the parachute deploys they are too far for the parachute to correct fully, then perhaps it makes sense for the boat to split the difference and sprint to meet the fairing after it has recomputed a new target.  Punching in and driving the boat to a new target position could even be done manually. Perhaps they are already doing that, but the boat couldn't get to the target position in time.

If the problem is cross-winds or ensuring the parachute "always approaches from the back" to protect the cockpit, perhaps it makes sense for the boat to sail into the wind, creating a moving target for the parachute.  But I'm not certain that commercially-available boat Dynamic Positioning systems support a time-varying target position.

If the problem is very low-level winds blowing the parachute off-course in the final minute, maybe some sort of manually-steered game of catch is appropriate, with both parties adjusting.  (Autonomously steered catcher boat?)

Perhaps part of the problem is picking the right GPS location to wait at---or to start at, if their final target is computed by the fairing.  Maybe they just need better pre-launch modelling of the descent...or a faster boat.

TL; DR: I'm certain the simplest thing is to wait at a particular GPS location and let the fairing come to you.  But given that they are having some problems with this approach I think it is *possible* they may try complicating it---a little. They've got great data now, so they've modelled the process enough to know what the simplest thing that will work is, even if we don't.

[Steve D]

Maybe they need a small computer on the fairing monitoring its GPS and radioing the ship where it thinks it can hit. Aim for a specific point initially and then use the computer to recompute where it can get to and coordinate with the ship in the process. I know from my skydiving days that we would attempt to hit a target but when the wind conditions didnt allow that we would pick another spot on the fly and go for that.

[StuffOfInterest]

Wouldn't the easiest thing be to have someone on the recovery boat do the final fly in?  GPS can get the parafoil in close, but it seems like the fine adjustments would be most easily handled by a person with the same remote you use for rc planes.

[Poole Amateur]

That would be a fun episode of Flitetest (You Tube channel) and get Josh Bixler to do it!

[wannamoonbase]

Wouldn't the easiest thing be to have someone on the recovery boat do the final fly in?  GPS can get the parafoil in close, but it seems like the fine adjustments would be most easily handled by a person with the same remote you use for rc planes.

Human in the loop, no thanks. 

Let computers do it, with the boat and fairing. 

[Lar]

Maybe they need a small computer on the fairing monitoring its GPS and radioing the ship where it thinks it can hit. Aim for a specific point initially and then use the computer to recompute where it can get to and coordinate with the ship in the process. I know from my skydiving days that we would attempt to hit a target but when the wind conditions didnt allow that we would pick another spot on the fly and go for that.

The fairing presumably already has a small computer, since it's controlling the parachute. Presumably (that word again) a routine to also predict the current impact point is not impossible to add but the computer may be close to the limit of what it can do real time. With a transponder and telemetry data, this computation could be done on the ship (maybe with a computer with more grunt since weight isn't (or almost isn't) an issue on the ship).

[Lar]

And from Pauline Acalin, to one-up the above...

https://twitter.com/w00ki33

Some really nice detailed pics. Strikes me that the intact half isn't going to be reused because in the last image, there's a chain  holding it down which is directly on the composite.

If one were planning on flying that again one would (presumably) at least put some padding on  so the chain doesn't abrade the edges... but for research of how it did coming down, that might not be important.

[ugordan]

Oh my, have to share a few more. She found a different perspective, and I'm in love with the well-worn fairing

First image, cool to see the scorch mark on top due to ascent aeroheating.

[Kabloona]

I guess why wouldn't the ship move to where the parachute seems to be guiding the fairing, rather than hold a fixed position? Baseball analogy, the fielder moves to get under the ball as they understand the ball's likely arrival point better and better.

Seems more likely to me than fixed position. Hitting a fixed position (+/- the width of the net) with the fairing would be quite difficult, if not impossible, given variable winds and a guidance computer that can't know wind conditions below it, and thus can't properly compensate, except trying to fly as tight a circle as it can all the way down, which still wouldn't be accurate enough.

The problem becomes much more soluble GNC-wise if instead of trying to hit a fixed point with the fairing, you try to guide it along a pre-programmed fixed path/azimuth, which is also pre-preprogrammed into the ship's DP system.

Then the problem reduces to the baseball fielder's analogy, like you say, where the path of the fairing is already known and it's just a matter of following the same line and getting underneath it.

[aero]

Didn't Elon write that they needed a bigger parachute to slow the fairing down? What do you suppose the "landing" speed of the fairing is? That is, how fast is it moving along its glide slope while Mr. Stevens is trying to catch it? Or just maybe, trying to catch up with it.

Wind speed and direction may be a significant factor here. Turning into the wind is obvious but it assumes the fairing "knows" the direction of the wind.

[AncientU]

Didn't Elon write that they needed a bigger parachute to slow the fairing down? What do you suppose the "landing" speed of the fairing is? That is, how fast is it moving along its glide slope while Mr. Stevens is trying to catch it? Or just maybe, trying to catch up with it.

Seems that isn't a fixed value, but a factor of wind speed at altitude and flight direction relative to that wind.  Fairing could theoretically be 'backing down' relative to the ground if pointed into a fresh breeze.  One of the challenging parts of the catch ship will be fielding the fairing when surface wind and wind aloft are quite different (which is a lot off the time, I think).

[Space Ghost 1962]

Didn't Elon write that they needed a bigger parachute to slow the fairing down? What do you suppose the "landing" speed of the fairing is? That is, how fast is it moving along its glide slope while Mr. Stevens is trying to catch it? Or just maybe, trying to catch up with it.

Wind speed and direction may be a significant factor here. Turning into the wind is obvious but it assumes the fairing "knows" the direction of the wind.

Inertial guidance systems can "infer" wind speed and direction. (Airlines use it routinely.)

[Lar]

Didn't Elon write that they needed a bigger parachute to slow the fairing down? What do you suppose the "landing" speed of the fairing is? That is, how fast is it moving along its glide slope while Mr. Stevens is trying to catch it? Or just maybe, trying to catch up with it.

Wind speed and direction may be a significant factor here. Turning into the wind is obvious but it assumes the fairing "knows" the direction of the wind.

Inertial guidance systems can "infer" wind speed and direction. (Airlines use it routinely.)

Quite well for where it IS ....but perhaps not so well for 200 meters further down and 600 meters over (or whatever the glide path is). I don't want to rube goldberg this[1] but maybe drones that are relaying inferred windspeed/direction back to the fairing control computer might add some accuracy?

Consider how much easier the fielder's job would be if the balls were being told wind speed and were cooperating with the mitt in being caught.

1 - who am I kidding, yes I do.