SpaceX Falcon 9/Dragon CRS-2 SpX-2 MISSION GENERAL DISCUSSION

[Robotbeat]

Root cause still unknown, but wouldn't be the first time check valves implicated in spacecraft propulsion anomalies. Loss of Mars Observer in '93 was attributed to probable hypergol leakage past check valves during long cruise to Mars.

Check-valves are kind of a curse-word among rocket engineers, from what I hear.

[R7]

2. CO2 -- since most of He production goes from natural gas, some processing schemes give He with substantial CO2 content. However, it is difficult to imagine that CO2 would form a solid ice blocking valve.

Well the He is released from very high pressure reservoir, it expands, does work, cools, no? But does it cool enough to freeze CO2, dunno. Moisture more likely would freeze.

[sdsds]

So is the result of a successful "co-elliptic" burn having the two orbits be co-planar ellipses? How close did they come to that? Does deviation or lack thereof indicate still-troubled or now-healthy thrusters?

[Norm38]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

[john smith 19]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw ... That's way too much
Also, listening to Elon's answers, I got impression that it is not a software problem. Well, then it could be **bad** helium:
In general, industrial high-pressure He could have 3 type of contaminants - well, may be more, but from my experience I know 3.
1. Rotary pump oil contamination in He was very common problem in our lab, but it was LONG time ago. I don't think they use this kind of pumps to compress He anymore.
2. CO2 -- since most of He production goes from natural gas, some processing schemes give He with substantial CO2 content. However, it is difficult to imagine that CO2 would form a solid ice blocking valve.
3. Dust -- some He bottles had very fine dust blocking our bleeding valves dead.

So, there are couple things SpaceX could do before the Dragon's return - check their He stock tanks for dust / organic contaminants; and - check the other Dragon's He pressure bottles for manufacturing dust inside.

I know it's obvious but I would have expected some kind of fine filter in the outlet pipes from the GHe tanks (possible with a heater wrapped round them to melt any frozen water/CO2 in it.

Of course they should be backup to a good filtering (an LN2 cold trap should catch this stuff on the ground and a warm, bone dry GN2 purge should keep the pipes clear till launch.

On the upside they've got them working and they are on a section that's going to be returned to Earth.

This remains one of the benefits of reusable systems.

[Kabloona]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

Not likely a single point failure, because Elon said one of the four NTO (oxidizer) tanks had full pressure initially. His mention of check valve(s) as possible culprits suggests that each branch of He plumbing to the ox tanks has its own check valve, which would be typical. In that scenario, three of the four check valves would have stuck closed...or frozen condensate in three of the four He lines...

[kirghizstan]

Assuming that it is three seperate failures you could guess it was the he if all tanks are filled from the same source

[LouScheffer]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

[Kabloona]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed. It would be like a fork with 4 tines, and a check valve at the end of each tine. Reasons being, you don't want a single stuck check valve to disable all 4 branches, and you don't want the He to tanks farther down the line passing through multiple check valves, when each branch needs only one.

[krytek]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed. It would be like a fork with 4 tines, and a check valve at the end of each tine. Reason being, you don't want a single stuck check valve to disable all 4 branches.

Then how did they pressure hammer it?

[Kabloona]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed. It would be like a fork with 4 tines, and a check valve at the end of each tine. Reason being, you don't want a single stuck check valve to disable all 4 branches.

Than how did they pressure hammer it?

AIUI, they cycled the main He isolation valve(s), which would be upstream of the check valves, on the "handle" of the "fork".

[starsilk]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed.

that's not how it should be plumbed.

[starsilk]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed.

that's not how it should be plumbed.

hmm. I was joking.. but perhaps there's a reason to plumb it that way - for example, if they need to pressurize each tank one after another, instead of all at the same time? if that was a requirement, it's easier to see why that configuration could have been picked..

[Kabloona]

If 3 of 4 He valves got blocked or stuck - this does not look like manufacturing flaw

Agreed, which is what I first thought. So to be clear, 3 out of four independent systems failed?  It wasn't a single point, one valve, that took out three thrusters?

How about if the system looks like

Gas  -> ----- -> ----- -> ----- -> ------
             T1       T2        T3       T4

where each "->" is a check valve?  Then if the second one sticks, T1 works and T2-4 have low pressure.

That's not how it would be plumbed.

that's not how it should be plumbed.

hmm. I was joking.. but perhaps there's a reason to plumb it that way - for example, if they need to pressurize each tank one after another, instead of all at the same time? if that was a requirement, it's easier to see why that configuration could have been picked..

Well, we're just arm waving now...but since we are, there's just no reason to plumb it that way. There's no reason i can think of to require *sequential* pressurization of the tanks, and if there was a reason, you wouldn't do it by stacking n+1 check valves on adjacent branches. You're just reducing overall system reliability for no benefit.

[kirghizstan]

I have no understanding of orbital mechanics so can someone please explain to me why Dragon uses near instant launch windows but can free drift for hours while fixing issues then can simply recompute the course to station?

[Galactic Penguin SST]

I have no understanding of orbital mechanics so can someone please explain to me why Dragon uses near instant launch windows but can free drift for hours while fixing issues then can simply recompute the course to station?

Because the tight launch window is for launching into the target orbital plane.

[dcporter]

I have no understanding of orbital mechanics so can someone please explain to me why Dragon uses near instant launch windows but can free drift for hours while fixing issues then can simply recompute the course to station?

Because the tight launch window is for launching into the target orbital plane.

To elaborate: Once you get off the rock and into the correct orbital plane, you can take your time playing all kinds of cheap-ish catchup games. Doing crazy things within an orbital plane isn't too bad; changing orbital planes is disproportionally expensive.

If you're not familiar with orbital planes, then this part ought to blow your mind a bit: Picture the globe, then put a sheet of paper through it at any angle you like. Now rotate the globe – day/night – but hold the piece of paper still. The ISS (and everything else in orbit) stays in orbit around the globe within that one sheet of unmoving paper. All the crazy orbit lines you see on maps on the front wall of NASA HQ are only crazy because they're loops in a plane drawn over a rotating globe. Crazy stuff. But the important thing is that changing your orbit from one piece of paper to another takes a lot of fuel, but moving around within the same one is comparatively cheap. As long as you end up launched into the right piece of paper (in this case, the same one as the ISS), then the rest is cake.

Disclaimer: Everything I know about orbital mechanics I learned from this forum, so if I missed something please someone chime in!

[IRobot]

I have no understanding of orbital mechanics so can someone please explain to me why Dragon uses near instant launch windows but can free drift for hours while fixing issues then can simply recompute the course to station?

Play some Kerbal Space Program and you will "get it" after a couple of launches

[Galactic Penguin SST]

I have no understanding of orbital mechanics so can someone please explain to me why Dragon uses near instant launch windows but can free drift for hours while fixing issues then can simply recompute the course to station?

Play some Kerbal Space Program and you will "get it" after a couple of launches

Or even better the Orbiter simulator: http://orbit.medphys.ucl.ac.uk/

[jcm]

Looks like a second burn to raise apogee at around 0023 UTC Mar 2, orbit is now 321 x 404 km per Space-Track