SpaceX Falcon 9 v1.1 - Dragon - CRS-5/SpX-5 -Jan. 10, 2015 - DISCUSSION

[Herb Schaltegger]

I expect that it's not the fluid per se that went low, but the pressurant. Remember, without an APU or an engine running to power a hydraulic pump, you have to have something else to provide the pressure in the system.

[malu5531]

How does that work? I didn't know you could run out of hydraulic fluid, unless there's a leak somewhere. If it were pneumatic, I'd understand, but where does the hydraulic fluid go?

I believe they use RP1 as hydraulic fluid.

[cambrianera]

How does that work? I didn't know you could run out of hydraulic fluid, unless there's a leak somewhere. If it were pneumatic, I'd understand, but where does the hydraulic fluid go?

You can have an hydraulic system run by an accumulator (without a pump).
Self powered and sometimes lighter, but with a limited life.

[Robotbeat]

How does that work? I didn't know you could run out of hydraulic fluid, unless there's a leak somewhere. If it were pneumatic, I'd understand, but where does the hydraulic fluid go?

It bleeds out (no return plumbing and I think it's pressurized just with some gas canister, not a pump). Many rockets use this type of hydraulic system, and running out is a fairly common failure mode. SpaceX uses the kerosene for actuating the Merlin's, so running out isn't likely to happen there.

[mmeijeri]

Ah, so in a sense it's both pneumatic and hydraulic, or at least compressed gas is used to provide the power.

[cscott]

Ah, so in a sense it's both pneumatic and hydraulic, or at least compressed gas is used to provide the power.

Yeah.  I doubt it uses RP-1 as the working fluid, since the grid fins are on the 'wrong side' of the core for that.

Probably dedicated high-pressure fluid; maybe shares the LOX pressurant.  (In the latter case you might deliberately kill the fins a little early to ensure you can shut down the main engine safely.)

I like SpaceX's incremental approach: they are refining the modelling of some of this stuff "in real time" as they get testing data for various parameters.  The first version of their hardware is under- or over-sized in various ways (over strength, excess pressurant, etc) and then they dial in the size of the various components as they get real-world testing data.

FWIW, I wouldn't be surprised if the leg redesign is intimately related.  If early leg deployment leads to static instability, they might compensate using the grid fins to provide active stability control.  But they are still trying to figure out how much pressurant/working fluid the grid fin system requires, so they are deferring early leg deploy until they have better models for grid fin fluid use through reentry.

The exciting implication might be that the test on the 29th already had "50% more hydraulic fluid" because they were already planning for early leg deployment on that flight, so had supersized the fin reservoirs.  (Hopefully their refined modelling continues to indicate that 50% is enough even with early leg deploy.)

[nadreck]

The exciting implication might be that the test on the 29th already had "50% more hydraulic fluid" because they were already planning for early leg deployment on that flight, so had supersized the fin reservoirs.  (Hopefully their refined modelling continues to indicate that 50% is enough even with early leg deploy.)

Elon eluded to that test slipping to February in his second to latest tweet:

"Upcomming flight already has 50% more hydraulic fluid, so should have plenty of margin for landing attempt next month"

[Herb Schaltegger]

Yeah.  I doubt it uses RP-1 as the working fluid, since the grid fins are on the 'wrong side' of the core for that.

Probably dedicated high-pressure fluid; maybe shares the LOX pressurant. 

Isn't the LOX tank self-pressurizing?

[rcoppola]

So is he saying that losing grid fin authority was "the" cause of this off-nominal landing? Or a contributing factor?
I mean, if they got that close and this quickly determined the only thing keeping them from success is more Hydraulic fluid, which they already have planned, then that is quite remarkable.

[JBF]

So is he saying that losing grid fin authority was "the" cause of this off-nominal landing? Or a contributing factor?
I mean, if they got that close and this quickly determined the only thing keeping them from success is more Hydraulic fluid, which they already have planned, then that is quite remarkable.

This is probably just the first issue identified. 

[OxCartMark]

Lets go through the possible causes of a hard landing:

Run out of fuel before altitude reached 0.
More fuel than expected decreased the acceleration.
targeted altitude was below the surface of the barge at impact (waves?)
targeted altitude was above the surface of the barge, causing a fall after v=0 and engine cutoff.
control software didn't get a reliable altitude input.
engine failure or slow ignition.
one or more legs failed to fully deploy.

And possible consequences of the hard landing:

Legs compress beyond design limits and break, allowing the rocket to tip over
Legs don't break, but the bounce tips the rocket over.
Engines bottom out into the deck.
Tank walls buckle.
Tank walls rupture then buckle.

Anyone see a possibility I missed?

I don't want to pick on this post specifically but I do want to point it out as an example of the thinking exhibited here by many members.  The thinking seems to over simplify what is being done to something along the lines of projecting a RC (or autonomous) helicopter downward unpowered with great velocity controlled only with grid fins then at the last minute turning on the helicopter (which has a low battery) and landing it.  Its far from being that simple because of the limited range of throttlability and especially because of the fact that even at the lowest available thrust its accelerating upward at more than 1G.  That makes for a tremendous math problem (from the standpoint of the level of my math education at least), and it makes for many many more interesting failure modes than I've seen being considered in this thread. 

To unmath the landing problem and put it into words (yes, a fools task at self deception) it plays out this way in my mind - For a given vertical velocity, mass, and an assumed average throttle setting (??85% if it goes from 70-100% IIRC???) you have one ideal engine start altitude, and the range of available throttle settings allows you some limited ability to compensate for any errors in calculation or perturbations.  For any given horizontal error in targeting as it passes through the ignition altitude you have a certain amount of angular divert maneuvering that needs to happen (at least once to initiate horizontal corrective velocity and at least once to reduce the horizontal velocity to just zero a few seconds before touchdown).  This horizontal position correction needs to take place within the time and thrust settings allocated for the vertical work and it needs to get the stage to zero horizontal velocity above the target with a few seconds / meters to spare before the landing time.  The landing time can not be varied significantly to allow for additional horizontal adjustment (the helicopter fallacy) So you either get horizontal and vertical to come together correctly using very few variables that affect both outcomes or you have some form of failness.   Oh, one place where it seems to me that the system could be gamed would be that if at a high altitude you realized you needed to reduce thrust beyond what was nominally possible an extreme divert then undivert maneuver might be possible in which the vertical component of thrust could be less than weight for a significant amount of time.           ....All words so don't trust them, wait for the math to come out.

So now that I've said that I'll delve into the speculation pool myself with this entry- Landing was near perfect but it came to Vz=0 at an altitude of a few single digit meters, and in that scenario the programmers had chosen to shut down the engine and fall down rather than the other option of flying off to the side and going into the ocean.

Mark

[mvpel]

I would have thought they'd have used an electro-hydraulic actuator for the grid fins instead. Strange tradeoffs at work there, I'm sure.

[lark]

At 7:42 in that video, you can see a deployed grid fin on the monitor at upper left. At 11:10, I believe you can see what looks like the landing burn video reflected off the control room's back window, also at upper-left.

[ArbitraryConstant]

NASA Coverage (vs SpaceX above) of Launch on YT:

Missed this when it was live, they staged just after they called out T+2:30, alt 58 km, speed 1.8 km/s, downrange 33 km. Lower and slower than CRS 4 staging:

[nadreck]

So is he saying that losing grid fin authority was "the" cause of this off-nominal landing? Or a contributing factor?
I mean, if they got that close and this quickly determined the only thing keeping them from success is more Hydraulic fluid, which they already have planned, then that is quite remarkable.

This is probably just the first issue identified.

I agree with your point JBF, not in a pessimistic way, but the fact is that whether or not the landing would have been a success with more hydraulics, this is not the only bit of data that they will uncover. Part of their work in determining cause of failure will result in some amount of 'overdetermination' of the failure and probably cleaning up other issues that could be marginal contributors to failure.

Had the stage survived the landing perfectly there would still have been a fault analysis of the landing to find any elements that exceeded planned tolerances.  Because there is less data (some parts will have been too damaged for accurate analysis) many possible causes will be examined as well as certain causes.

[Herb Schaltegger]

I would have thought they'd have used an electro-hydraulic actuator for the grid fins instead. Strange tradeoffs at work there, I'm sure.

So you're positing and electrically-powered hydraulic pump? That requires heavy batteries, cabling, connectors, etc. and has a much smaller body of flight history than a standard hydraulic system. Further, an open-loop hydraulic system has the (minor) advantage that as fluid and pressurant is consumed and lost overboard, stage mass goes down.

[randomly]

They already have the energy in the high pressure helium tanks to use.

[rcoppola]

So if they "Already" have 50% more fluid planned, that means they are and/or continue to be ridiculously fast at iterative design or more likely in this case, already knew through modeling they'd need more margin. Or both.

[Herb Schaltegger]

So if they "Already" have 50% more fluid planned, that means they are and/or continue to be ridiculously fast at iterative design or more likely in this case, already knew through modeling they'd need more margin. Or both.

They probably already scarred the interstage design with brackets or at least mounts and connection points into the system for another fluid reservoir and possibly an additional pressurant tank.

[OxCartMark]

I would have thought they'd have used an electro-hydraulic actuator for the grid fins instead. Strange tradeoffs at work there, I'm sure.

Ditto

Second thought (probably a poor one due to misunderstanding supersonic / hypersonic flow (and being at the wrong end of a long cylinder)) how about powering the grid fins with something like the ram air turbines that airliners use? You'd lose control authority as things slow down near the barge but by then the fins aren't(?) needed.  Probably a longer term solution for the future if the rest of the initial attempts and investigations into landing and recovery continue to go well.