-
#120
by
ShaunML09
on 20 Dec, 2019 23:00
-
Officially concerned about recovery now. There is clearly a bigger problem than earlier discussed. No confirmation of the additional burns and this statement from the mission commander:
We are working several options to recover Spacecraft 3 at one of our Western U.S. landing locations within the next 48 hours and are approaching the next few days with guarded optimism of a successful recovery. Our Landing and Recovery teams are executing their pre-planned contingency deployment to White Sands Missile Range and will be in position by later this evening to make a fully supported recovery.
https://starlinerupdates.com/
-
#121
by
meekGee
on 20 Dec, 2019 23:04
-
But you get free return when launching in a lower one.
The most valuable resource during a contingency is time.
If the spacecraft is unhealthy, a fast forced return is a recipe for LOC.
Being stranded in orbit is a very remote second risk in comparison.
Unsupported claims
The spacecraft can support uncontrolled entry. It is dynamical stable. So fast return is not an ingredient in the recipe for LOC.
Therefore, stranded in orbit is a much higher risk by magnitudes
A HEALTHY spacecraft can.
One that's doing weird s**t and you don't know why - I'd rather have an opportunity to figure out what's going on before being forced into EDL.
Is suborbital insertion the default for all missions or was it special for this one?
Wrong. It can passively reenter. Which means an unhealthy spacecraft (no attitude control) can return. Just needs the ability to deploy the chutes.
Passively is one thing...
OTOH a confused avionics suite can decide to deploy the chutes at some dumb time because the timer is wrong, or the barometer is calibrated wrong, or fire thrusters at the wrong time, or who knows what..
Nothing is gained by hurrying up in a contingency. First step is to assess the situation, and you can't do that if you are about to re-enter in 45 minutes and have limited comms because your attitude is unstable.
-
#122
by
Quindar
on 20 Dec, 2019 23:18
-
While I believe there is something to the MET timer causing an issue, I don't understand how a mission timer could be sole reason for the failure to properly insert.
Are you telling me the entire Starliner autopilot just operates on that timer, and once it was clear that no insertion was happening, its RCS thrusters continued to operate as though it was inserting? It doesn't take into account attitude, velocity etc.... and you know, sensors that verify the engines are firing. The explanation doesn't make sense IMO based on what we know right now.
Yep, this has been bugging me too, since we know that Orion and Dragon have triple-redundant computer systems that can withstand being blasted by radiation and still recover.
Here we have an apparently nominal flight where the entire spacecraft loses track of what time it is? Doesn't the software check to see if all clocks are in sync before each mission objective is executed?
Also, the fact there was a failure to communicate for 8 minutes seems excessive. I get temporary drop outs due to coverage, but 8 minutes? Seems like a poor design to do your orbital insertion maneuver where there is no communication link to NASA??? We are not on the far side of the moon here.
Yep, the other thing bugging me. Why is there a need to communicate with Earth at all? We have fairly decent self-driving cars here on Earth, so why can't they have fully autonomous flight control software that would monitor and manage the flight?
And I'm not arguing against being able to take manual control, just that we are technologically advanced enough to have computer systems that can understand when something is going wrong, and to reassess whether the mission should proceed or that input from humans is required. Obviously there is no overall system monitoring fuel consumption vs mission objectives, otherwise this wouldn't have happened.
To summarize, the hardware worked but the software failed. This is a pattern we've seen before with Boeing, and they need to understand that they may have an institutional problem.
My $0.02
As a software and systems guy, this failure and the public explanation don't really make sense to me. If the MET was the single point of failure, then NASA/Boeing standards, reviews, and checks aren't worth more than just toilet paper in terms of usefulness. So, as others like Wayne have pointed out, don't believe the early explanations.
It really is unfathomable that a single, out of band clock could cause a failure like this. There has to be a cascade of failures and failures of the exception handling and redundant verification systems to explain the incorrect thruster firing and off-nominal insertion. And the whole TDRS and sending commands from the ground to fix the issue are really secondary. A manual override cannot be the solution to a known single-point-of-failure built into the system.
As ShaunML09 mentioned, there should be cross-checks with other systems to verify that the spacecraft is where and when it thinks it should be; working blindly from a single clock doesn't pass the smell test.
So we'll have to wait and see what the official root-cause analysis points to. But the story we've been told so far, really can't be the only explanation.
Boeing: Same guys who did the software in 737MAX, seems unintended consequences are haunting these guys.
-
#123
by
obi-wan
on 20 Dec, 2019 23:21
-
But you get free return when launching in a lower one.
The most valuable resource during a contingency is time.
If the spacecraft is unhealthy, a fast forced return is a recipe for LOC.
Being stranded in orbit is a very remote second risk in comparison.
Unsupported claims
The spacecraft can support uncontrolled entry. It is dynamical stable. So fast return is not an ingredient in the recipe for LOC.
Therefore, stranded in orbit is a much higher risk by magnitudes
A HEALTHY spacecraft can.
One that's doing weird s**t and you don't know why - I'd rather have an opportunity to figure out what's going on before being forced into EDL.
Is suborbital insertion the default for all missions or was it special for this one?
Wrong. It can passively reenter. Which means an unhealthy spacecraft (no attitude control) can return. Just needs the ability to deploy the chutes.
You also need to jettison the service module prior to entry interface.
-
#124
by
CorvusCorax
on 21 Dec, 2019 00:41
-
You also need to jettison the service module prior to entry interface.
Not necessarily. When reentering over Russia, you can let the aerodynamic forces do that job for you...
From
http://www.spacesafetymagazine.com/space-disasters/close-calls/soyuz-5-multiply-reentry-failures-endanger-cosmonaut/
Following the retrofire, the service module failed to separate from the reentry module, resulting in the spacecraft reentering the atmosphere hatch-first, the most aerodynamically stable orientation. The resultant aerodynamic and thermal loading caused the hatch to bulge inward and the hatch seal to smolder, spreading fumes into the cockpit.
Piloting the spacecraft became increasingly difficult, as the cabin filled with smoke and Boris Volynov, single astronaut onboard, was pressed outwards from his seat against the restraint harness. Fortunately the struts connecting the service module to the reentry module disconnected due to mechanical and thermal loading, resulting in the separation of the service module. The reentry module was then able to re-orient itself correctly, exposing its thermal protection in the direction of the velocity vector.
Following the attitude correction, the spacecraft continued on a 9G ballistic reentry, deviating from its planned trajectory by 2000km and landing 2km southwest of Kustani. The parachute deployed correctly, but shortly before touchdown at 07:78 GMT, a failure occurred in the soft-landing rockets. Consequently, the reentry module struck the ground much harder than usual, throwing Volynov from his harness across the cabin and breaking several of his teeth. Volynov walked away from the landing site to Kustani, through snowy, -40-degree temperatures, where he was recovered by rescue crews.
Ok, in retrospect, detaching the service module is definitely the preferred method...
-
#125
by
ZachS09
on 21 Dec, 2019 01:15
-
Could this anomaly be the result of human error based on the recent discussions on this thread?
-
#126
by
ChrisWilson68
on 21 Dec, 2019 01:49
-
Could this anomaly be the result of human error based on the recent discussions on this thread?
I don't see how it could be anything but human error.
-
#127
by
SWGlassPit
on 21 Dec, 2019 01:52
-
Could this anomaly be the result of human error based on the recent discussions on this thread?
I don't see how it could be anything but human error.
We simply don't have enough information to make that kind of judgment call right now. Let's not get ahead of ourselves
-
#128
by
ChrisWilson68
on 21 Dec, 2019 01:53
-
I think NASA won't require Boeing to do another uncrewed test flight to the ISS before putting people on board, and I don't think Boeing will do one if NASA doesn't force them to because it would cost so much. And I think that's a mistake.
Having an uncrewed flight go through all the steps of docking with the ISS, staying there for a while, then leaving and returning to Earth would catch a bunch of kinds of failure modes that wouldn't be caught by this cut-short test flight, and those failure modes might kill the crew.
-
#129
by
su27k
on 21 Dec, 2019 03:33
-
But you get free return when launching in a lower one.
The most valuable resource during a contingency is time.
If the spacecraft is unhealthy, a fast forced return is a recipe for LOC.
Being stranded in orbit is a very remote second risk in comparison.
Unsupported claims
The spacecraft can support uncontrolled entry. It is dynamical stable. So fast return is not an ingredient in the recipe for LOC.
Therefore, stranded in orbit is a much higher risk by magnitudes
That's what I thought too, but this news article seems to imply otherwise:
https://www.washingtonpost.com/news/business/wp/2018/07/24/feature/nasa-trained-boeing-employed-chris-ferguson-hopes-to-make-history-as-a-company-astronaut/The shuttle had wings, like a plane. His new spacecraft was a thimble-shaped capsule, far more difficult to control. In this particular test, he’d be facing a worst-case scenario: Every computer of the autonomous spacecraft would be out, meaning he’d have to fly it manually, hitting the atmosphere at Mach 25, or 25 times the speed of sound, then, somehow, bring it down for a soft landing. Two of the four NASA astronauts who had attempted it had failed, losing control of the spacecraft so that it tumbled, and Ferguson was eager to get in some extra practice.
“I don’t need to embarrass myself or you,” he had said to his trainer.
“Practice is good,” she said.
He was struggling with what he called the “pitch worm.” If the spacecraft pitched too much, it would start a tumble almost impossible to recover from. He was also having a hard time seeing the screen in front of him because it was so close.
“I have to tilt my head up to see through the right part of my glasses,” he complained.
Seems to me some sort of active control is needed for a successful re-entry.
-
#130
by
AnnK
on 21 Dec, 2019 06:12
-
Shows again and again nothing beats a properly trained crew. They would of flown the correct mission. Engineers think their computers can fly the air or spacecraft better and are wrong. The Max is a great example of this taking control away from the crew and causing crashes. Am sad to say Boeing steps in it again. Looks like Space X will be the first to restore crewed flight to the USA.
-
#131
by
Lars-J
on 21 Dec, 2019 06:41
-
Shows again and again nothing beats a properly trained crew. They would of flown the correct mission. Engineers think their computers can fly the air or spacecraft better and are wrong.
The VAST majority of spacecraft are unmanned and computer controlled. They work because they have to.
Relying on computers is not Boeing’s problem.
-
#132
by
Hankelow8
on 21 Dec, 2019 06:55
-
What I find interesting is whereas Boeing finds it necessary to empty the abort engine fuel tanks on entry into orbit, which would be jettisoned before reentry, SpaceX does not feel the need for this. As well as not emptying the fuel tanks the Dragon capsule retains fuel onboard not only in orbit but through reentry and touchdown.
-
#133
by
TheRadicalModerate
on 21 Dec, 2019 07:55
-
One that's doing weird s**t and you don't know why - I'd rather have an opportunity to figure out what's going on before being forced into EDL.
Is suborbital insertion the default for all missions or was it special for this one?
Wrong. It can passively reenter. Which means an unhealthy spacecraft (no attitude control) can return. Just needs the ability to deploy the chutes.
Even if the vehicle is dynamically stable, over a pretty wide range of angles, you still have to use RCS to turn it around for reentry, don't you? And if you've got the RCS to do that, then you've got the RCS to put in 60-90 m/s of delta-v for deorbit.
-
#134
by
kendalla59
on 21 Dec, 2019 08:06
-
What I don't understand is why the RCS would need to use so much fuel when the spacecraft was idle? Apparently the software thought that highly accurate orientation was necessary, even though the insertion burn was not really happening. So if there is no burn, why wouldn't the system orient the spacecraft and then simply stop firing, or at least fire the thrusters very rarely? That seems like a big issue also, the software isn't designed to simply fine-tune the orientation and then just quit in the absence of other forces pushing it out of position. I've worked with motion control systems, and I've seen them oscillate a lot when they are poorly designed or when the filtering parameters haven't been tuned correctly. I've also seen companies short-change the software development aspect of their business because they have no idea how much more complex software is than mechanical systems and passive electronics. I've known managers who really think that squeezing the software budget is the smart way to control costs.
-
#135
by
TheRadicalModerate
on 21 Dec, 2019 08:14
-
I thought the insertion burn was relatively small. Why do they have to be suborbital to burn that amount of propellant?
Without the phasing delta-v, you'd need:
33 m/s to go from the initial 72 x 181 to 181 x 181
67 m/s to go from 181 x 181 to 181 x 410
65 m/s to go from 181 x 410 to 408 x 410
Throw in another 35 m/s for phasing corrections and other rendezvous/docking jiggery-pokery and you're at a nice round 200 m/s.
I assume that they're using RCS, and not the abort thrusters, so... Isp=210? With a 13 t vehicle, that would eat up 1.2 t of prop. That's not chickenfeed.
Two things:
1) I wonder if having the prop in the SM instead of in the CM gave them trouble with the torque and moment limits for the IDA and NDS.
2) If you were dead-set on justifying your existence by competing on down-mass, you'd want to find a way to burn off every last kg of prop you could. I wonder if Boeing feels that they need to justify their higher price by providing more value on down-mass, and that led them to make what seems to be shaping up to be a dumb call.
-
#136
by
kendalla59
on 21 Dec, 2019 08:36
-
Shows again and again nothing beats a properly trained crew. They would of flown the correct mission. Engineers think their computers can fly the air or spacecraft better and are wrong. The Max is a great example of this taking control away from the crew and causing crashes. Am sad to say Boeing steps in it again. Looks like Space X will be the first to restore crewed flight to the USA.
A cheap toy quad-copter drone has a computerized flight control system costing a few dollars that allows it to be flown by young children. Without the computer, even a very skilled and experienced human operator would crash the aircraft more often than not. I'd venture to say that the problem with Starliner (and Max) is a lack of hyper-focused attention to the details of the computer control system, not the lack of highly trained human operators.
-
#137
by
GWR64
on 21 Dec, 2019 08:58
-
I thought the insertion burn was relatively small. Why do they have to be suborbital to burn that amount of propellant?
Without the phasing delta-v, you'd need:
33 m/s to go from the initial 72 x 181 to 181 x 181
67 m/s to go from 181 x 181 to 181 x 410
65 m/s to go from 181 x 410 to 408 x 410
Throw in another 35 m/s for phasing corrections and other rendezvous/docking jiggery-pokery and you're at a nice round 200 m/s.
I assume that they're using RCS, and not the abort thrusters, so... Isp=210? With a 13 t vehicle, that would eat up 1.2 t of prop. That's not chickenfeed.
Two things:
1) I wonder if having the prop in the SM instead of in the CM gave them trouble with the torque and moment limits for the IDA and NDS.
2) If you were dead-set on justifying your existence by competing on down-mass, you'd want to find a way to burn off every last kg of prop you could. I wonder if Boeing feels that they need to justify their higher price by providing more value on down-mass, and that led them to make what seems to be shaping up to be a dumb call.
The Starliner also offers the possibility for Space Station reboost.
I suspect Starliner is pulling the ISS and therefore always docking on IDA-2.
So there should normally be fuel reserves.
https://www.rocket.com/space/human-exploration/cst-100-starliner
-
#138
by
ChrisWilson68
on 21 Dec, 2019 14:56
-
Shows again and again nothing beats a properly trained crew. They would of flown the correct mission. Engineers think their computers can fly the air or spacecraft better and are wrong. The Max is a great example of this taking control away from the crew and causing crashes. Am sad to say Boeing steps in it again. Looks like Space X will be the first to restore crewed flight to the USA.
I know what you mean -- I once had this apple and it was bad. So all oranges throughout the world must be better than all apples until the end of time.
-
#139
by
Jim
on 21 Dec, 2019 15:06
-
One that's doing weird s**t and you don't know why - I'd rather have an opportunity to figure out what's going on before being forced into EDL.
Is suborbital insertion the default for all missions or was it special for this one?
Wrong. It can passively reenter. Which means an unhealthy spacecraft (no attitude control) can return. Just needs the ability to deploy the chutes.
Even if the vehicle is dynamically stable, over a pretty wide range of angles, you still have to use RCS to turn it around for reentry, don't you? And if you've got the RCS to do that, then you've got the RCS to put in 60-90 m/s of delta-v for deorbit.
No, it is self righting. No, deorbit needed. It is suborbital