SpaceX Falcon 9 v1.1 - SpX-7/CRS-7 DRAGON - Discussion Thread 1

[FinalFrontier]

Here is the LOXcam from CRS-4 as an example:



Three COPVs visible, likely a fourth outside the camera's field of view. Trying to find footage from other flights, but my google-fu is weak today.

Perfect. Really appreciate this.

Now imagine if the contents of one or more of these was suddenly dumped into the LOX tank. Either COPV failure or helium valve failure. This is my thinking right now.

Now this camera was looking down on the tank from the region of the upper tank dome. The COPV's are mounted near the top of the tank. To determine if one ruptured, and that was the failure, I would recommend reviewing the zoomed in slow-mo gif's and comparing the region where we first see the tank walls fail, with where each COPV is mounted/ or best guess.

[butters]

The fact that they're having so much difficulty figuring this out from telemetry suggests that the either the vehicle was operating normally right up to the moment of a very rapid event (which couldn't be captured in telemetry) or that a faulty sensor erroneously indicated normal operation (causing the failure and masking the cause in the telemetry). Another possibility is that they just don't have a telemetry channel which would record whatever went wrong.

[pbreed]

Ok assume the IDA came loose.
Assume it pushed the electronics stack into the top of the Lox dome.
A pressurized tank is amazingly strong when exposed to outside forces that do not tear the skin.

So the simple static case...

IDA sitting on electronics stack   500Kg IDA, Electronics  stack 30cm square 3G acceleration pressure on the tank dome would be 22psi, probably under the tank static pressure.

Now if the IDA falls 1 meter and then stops in 30cm using a 30cm square electronics stack pushed into the lox dome I calculate 99psi pressure on the dome, probably more than enough to tear the dome if the electronics boxes being pushed in have sharp edges.

[meekGee]

You'd expect to see a structural failure like that at Max Q. This was much later than Max Q when structural loads were much lower.

That's the intuitive version...

I think we might safely dispense with Elon's "counter-intuitive" comment since it now seems they have no clear idea yet what happened.

I buy the interpretation that at that point pressure should have been dropping, and hence a rise in pressure is "counter-intuitive" - which allows all sorts of scenarios.

But the new information paints a very clear contradiction:
- The break up was slow
- There was no (or little) useful information before telemetry got corrupted.

From this we understand that the breakup:
- affected the avionics before the avionics had time to send a full data frame
- did no affect the first stage.

This eliminates almost all of the other proposals.  (vents, valves, etc) and leaves only a few standing:
- Helium tank completely bursting, if such an event can destroy the LOX tank and avionics above it quickly enough, and yet not completely blow up the LOX tank.  (not impossible)
- Loose IDA
- Failure that starts at the avionics. (literal computer melt-down...  followed by bad commands to the pressurization system)

[Kabloona]

Now this camera was looking down on the tank from the region of the upper tank dome. The COPV's are mounted near the top of the tank. To determine if one ruptured, and that was the failure, I would recommend reviewing the zoomed in slow-mo gif's and comparing the region where we first see the tank walls fail, with where each COPV is mounted/ or best guess.

Except the first signs of LOX cloud seem to come from the far side of the vehicle, so you can't see the source of the cloud.

[punder]

Now this camera was looking down on the tank from the region of the upper tank dome. The COPV's are mounted near the top of the tank. To determine if one ruptured, and that was the failure, I would recommend reviewing the zoomed in slow-mo gif's and comparing the region where we first see the tank walls fail, with where each COPV is mounted/ or best guess.

Except the first signs of LOX cloud seem to come from the far side of the vehicle, so you can't see the source of the cloud.

That at least narrows it down by half.

[pbreed]

That calculation is just predicated on the forces necessary to stop the IDA fall vs the assumes area of the electroncis stack pushing into the dome.  This assumes the opposing forces on the inside of the tank are provided by gas pressure or liquid contact.  If the opposing pressure is provided by liquid contack the "spike" pressure could be way more than 99psi as tank top would stop deforming and come to an abrupt halt. If the distance traveled goes from 30cm of crush to 10 cm of crush the pressures would go up about 3x.

[butters]

I'm doubting the loose IDA theory because I think they would have seen unexpected vibrations and acceleration changes in the telemetry prior to catastrophic vehicle damage. If if was loose IDA they'd have figured that out by now.

[pbreed]

If it suddenly detached it would travel 1m in 260 msec.  It would hit the avionics stack sending telemetry first.
So unless the IDA detach was detected by dragon telemetry you would never see it.

Assuming the whole stack still weights 100K lbs the acceleration change while it was in free fall would be 1% or so...

[CuddlyRocket]

Assuming there was an explosive decompression of the S2 LOX tank, did this occur due to ongoing rising internal pressure or was there a one-time increase in internal pressure coupled with the decreasing external atmospheric pressure as the vehicle continued to rise?

A failure due to a gradual increase in the differential between internal and external pressure would cause the tank to fail at its weakest point (though this may subsequently propagate). Alternatively, a sudden large increase may well cause the tank to fail at multiple weak points. It would be unlikely that these weak points would be distributed in a cylinderically or otherwise symmetrical fashion; such a symmetrical failure would imply an external force (though this may just be a result of the combination of continuing S1 thrust and atmospheric drag).

Am I correct in thinking that prior to S1 separation the vehicle is under the overall control of the S2 avionics package? If so, does the fact that the S1 seemed to continue operating normally for some time after the anomaly commenced imply that the S2 avionics package and internal communications also survived for that same period? What implications may this have for the cause of the anomaly?

[DigitalMan]

this is a video segment I found that appears to be the right side view

insert html doesn't seem to be working

look at gallery/s9jndDq on imgur

[DecoLV]

As Jim says, range safety is the most important consideration in an event such as this. Besides, don't forget that the first stage landing ability is, so far, an unproven technology. In a catastrophic event like this, there is no room for speculation or experimentation. Vehicle fails, FTS  should be used.

This is how it works today. But I think when this incident is all over the more interesting question (and maybe deserving a public thread of its own) is whether it is time to re-write the rules on FTS generally. Spx S1 may be recoverable in the near future...maybe others to come. Cargo Dragon, with some new avionics and software, could be recoverable as well...if desired. So maybe the day has come when we consider individual components of the LV instead of just blowing up the whole thing.  For example, override command an early stage sep, spacecraft in-flight abort (as internal FCS would) and only detonate the individual FTS for the individual stage. Yes, that is more complicated, but so is everything else these days.

[Jim]

So unless the IDA detach was detected by dragon telemetry you would never see it.

 IDA related data is not going to be part of the launch vehicle telemetry in the first place.  it would be in the Dragon stream if at all.

[Bubbinski]

Has any debris from the failure been picked up yet from the sea? Or the coastline/beach? 

And have salvage ships been hired or sent out by SpaceX?  If they can't get the cause from telemetry I would imagine that would be the next step, to fish out the LOX tank, Dragon trunk, IDA, etc.  How much time would that take?

[Jim]

This is how it works today. But I think when this incident is all over the more interesting question (and maybe deserving a public thread of its own) is whether it is time to re-write the rules on FTS generally.

why?  The point of FTS is for safety and not saving money.   You want to save money, then make sure your vehicle isn't going to have problems that would require FTS.  That is where the focus should be and not trying to make and engineer solutions for a matrix for all the different destruct scenarios for an already rare event.

[meekGee]

I'm doubting the loose IDA theory because I think they would have seen unexpected vibrations and acceleration changes in the telemetry prior to catastrophic vehicle damage. If if was loose IDA they'd have figured that out by now.

That's a valid point.

If you look at the geometry, the "avionics tower" sticks up, either to the trunk or actually intrudes into it.  It would be one of the first things to get hit. Even at 3g, covering the first 1 m takes about 0.25 seconds - more than the hypothetical Helium tank propagation speed.

However, the relative change in acceleration will be 0.5 ton / rocket-weight-at-the-time which was 150-200 tons, so 0.3%.   (Based on data in the newly designed SF101)

How clear this change would be, I'm not sure.

[Joaosg]

NASA, SpaceX have released new # and e-mail for public to report any Falcon 9 debris: 866-392-0035 or [email protected]

James Dean tweet
https://twitter.com/flatoday_jdean/status/615598432863023104

[JasonAW3]

     I'm wondering if it's possible that the cargo in the Trunk may have broke loose and slammed into the upper tank.

     If the Oxidizer is in the upper tank, that could account for the sudden overpressure.

     I find it interesting that the voice over said that the craft was off course just before you saw the "smoke" coming from the upper stage.

[pbreed]

IDA was designed to be released on command.
So an IDA detach could be a wire chafing and grounding the detach command actuator.
It could be some kind of coupled vibration process releasing the attach..
I believe the detach is purely mechanical not pyro....

It could also be some sort of structural failure farther up the chain, ie the IDA mount held, but the thing it mounted to deflected due to dragon loads and buckled etc.... lots of possibilities in the failure chain.

[Jevans]

     I'm wondering if it's possible that the cargo in the Trunk may have broke loose and slammed into the upper tank.

You and everybody else, it seems. Read upthread.