Dragon v2 Kremlinology from Helodriver's unveiling pics

[Helodriver]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control. I think many people are overestimating the amount of "flying" this vehicle's crew is going to have to do in off nominal situations based on the paradigm set by other early generation vehicles. I highly doubt its going to have the ability to be manually flown to orbit with the hand controller while attached to the booster, and I'd expect that should an abort be called for, a maximally redundant system is going to fire the Superdracos long before an astronaut is even lifting their arm off the rest to reach for a button. Surely there will be overrides and options, as their should, but being a true fly by computer vehicle, every control input will really be voting with the computer about what the vehicle should do.

The thought that a crewmember is going to have to drill down into multiple pages of submenus or type in fiddly alphanumeric sequences to get something critically important done under high vibration or acceleration environments reflects poor system engineering that even a relative spaceflight newcomer like SpaceX is bound to avoid. A big red hand size button popping up that blinks "ABORT" is far more likely.

[pagheca]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control.

[...]

Yep. That looks as a much more reasonable explanation to my doubts, to be honest. Thanks for the insight.

[ChrisWilson68]

Oh ye of little faith; hast thou not viewed process control from a modern software engineering paradigm?
Go now an contemplate the appropriate button for the blue screen of woe and have faith in the systems engineers.

do you mean something like this, except you don't have two hours and half to remain "dead in the air" while an astro-tug comes and collects you?

Well then it's a good thing SpaceX uses Linux and Qnix, not Windows, on its spacecraft.

[cscott]

If I have a mechanical "3D" switch i can grab it and push on it even in a very stressful environment, feel the movement and the vibration of the mechanism. If this is just an area of a screen the feedback is not the same.

If there is smoke in the cabin, I can still grab the last button in the row, count the third left and press it.

When I type something on my iphone... android,  the shape of the key do not help to "reset" the position of my hand and the fingers may shift continously If I do not watch the keyboard.

If there's visibility-impairing smoke in the cabin, it's already likely a Very Bad Day.

When I've read the Columbia accident report and the Apollo 13 reports, etc, I'm usually struck by how *much* time the astronauts have.  Most actual incidents do not demand "hair trigger twitch" reactions -- those elements which do require superhuman reaction times (eg, LAS) have already been computerized to take the (slow) human out of the loop.

IMO the main role of an astronaut-in-the-loop is in careful deliberative analysis, improvisation, and response.

But I'm not an expert in control design, so I don't really feel qualified to continue this particular thread.

[ChrisWilson68]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control. I think many people are overestimating the amount of "flying" this vehicle's crew is going to have to do in off nominal situations based on the paradigm set by other early generation vehicles. I highly doubt its going to have the ability to be manually flown to orbit with the hand controller while attached to the booster, and I'd expect that should an abort be called for, a maximally redundant system is going to fire the Superdracos long before an astronaut is even lifting their arm off the rest to reach for a button. Surely there will be overrides and options, as their should, but being a true fly by computer vehicle, every control input will really be voting with the computer about what the vehicle should do.

The thought that a crewmember is going to have to drill down into multiple pages of submenus or type in fiddly alphanumeric sequences to get something critically important done under high vibration or acceleration environments reflects poor system engineering that even a relative spaceflight newcomer like SpaceX is bound to avoid. A big red hand size button popping up that blinks "ABORT" is far more likely.

Absolutely.  When there are engines firing, there's a very limited number of things to be done in an emergency that a person could decide to do manually.

Remember when the engines are firing: on ascent, when the big Merlin F9 engines are firing and in the last moments before touchdown.  In the former case, about the only decision a person could make is abort or not abort.  In the later case, there might be an option to abort to parachutes or to redirect to a different landing site, but that's about it -- you're going to be on the ground one way or another in a very short time.

At all other times, there won't be any engines firing, hence no vibrations and acceleration to worry about.

[Helodriver]

If I have a mechanical "3D" switch i can grab it and push on it even in a very stressful environment, feel the movement and the vibration of the mechanism. If this is just an area of a screen the feedback is not the same.

If there is smoke in the cabin, I can still grab the last button in the row, count the third left and press it.

When I type something on my iphone... android,  the shape of the key do not help to "reset" the position of my hand and the fingers may shift continously If I do not watch the keyboard.

If there's visibility-impairing smoke in the cabin, it's already likely a Very Bad Day.

When I've read the Columbia accident report and the Apollo 13 reports, etc, I'm usually struck by how *much* time the astronauts have.  Most actual incidents do not demand "hair trigger twitch" reactions -- those elements which do require superhuman reaction times (eg, LAS) have already been computerized to take the (slow) human out of the loop.

IMO the main role of an astronaut-in-the-loop is in careful deliberative analysis, improvisation, and response.

But I'm not an expert in control design, so I don't really feel qualified to continue this particular thread.

In my military flight training we were taught to make time during an emergency. Very few actually required an immediate knee jerk response, less than half a dozen had steps that needed rote memorization, no matter what aircraft it was. Instructors taught techniques like  "step 1, wind the clock"  and "no fast hands in the cockpit" to break the tendency toward quick nonthinking response that may exacerbate an already deteriorating situation. (shutting down the good engine when the bad one is actually on fire) There are always a few seconds to maintain aircraft control, take appropriate actions, and refer to checklists. Making time, coupled with simulator training and craft familiarity and assisted by automated systems and ground control will be the way SpaceX capsule flyers will work through problems, no matter what configuration the screens or sticks have. Little to none of them are going to be solved by any sort of split second snap judgments or heroics on the controls.

[MP99]

You worked on OLPC[2]? That's AWESOME.

Yup.  @jg is hanging around these parts, he's an OLPC alumnus as well.

Will the touchscreen monitors survived and be touchable during vacuum environment at very low temp if they had to open the hatch to go EVA from the Dragon V2 in the distant future?

No physical reason why they cannot be.  Assuming they are LED (not LCD), it's all just solid state transistors and some transparent plastics.  (For our part, we only rated our machines down to "walking to school in outer Mongolia" temperatures.)

cscott, do you think that obviously you can reproduce any fault that requires split seconds intervention and save the mission by pressing one of the 30 (I counted them) buttons available in the V2 main console without having to push the "abort" one (if exists) at the first real malfunction in a critical situation?

I'm not sure what point you're trying to make.  Finding the "right" switch in the maze of of a 787 -vs- drilling down on a touchscreen right in front of you?  Comparing the severity of "touchscreen failure" in a Dragon v2 -vs- in a F-35?  Or are you just upset that the computer is likely to be better at reacting to and flying failure modes than a human pilot?

There's also something called flight simulations and training that I believe is used extensively in commercial aeronautics and just perhaps might be transferrable to aerospace.
Cheers

Don't forget that Garrett Reisman said at the Q&A that he'd already flown a Dragon simulator (and that the approach to ISS was surprisingly similar to Shuttle).

Cheers, Martin

[pagheca]

Absolutely.  When there are engines firing, there's a very limited number of things to be done in an emergency that a person could decide to do manually.

I can imagine endless discussions between engineers about discarding a-priori a "Try SCE on aux" option at Boeing and Spacex. But I guess time are mature and the risk of an incorrect or late response by the astronaut or ground control in emergency situation is larger today than the risk of a s/w or h/w failures.

On the other side, I think that this would apply to aviations too, but psychological and social issues still deny the possibility of applying this to the civil aviation market. For example, even not considering pilot errors, if one considers "suicide" a "human s/w error" corresponding a little bit to the blue screen of death, there have been already a couple of accidents like that.

DISCLAIMER: I like to fly and I like the aviation environment, where I worked for a couple of decades. If you are a pilot, even an astronaut, please don't take it personally. It's just a "cold" opinion based on what I think are facts.

[ThereIWas3]

Sometimes we forget, with our modern computers, just how primitive the Apollo spacecraft were.
The on-board computer served as an autopilot, and did the orbital mechanics math, and not a
whole lot else.  It simply did not have the spare capacity for such things.  Remember, on Apollo 11,
when Buzz Aldrin left the ascent radar turned on while they were descending to the moon, the
computer was overloaded by inputs coming in from both radars at once - that is how marginal
it was.

So everything else was manual - hence the miles of wiring and hundreds of switches.

Today, with minature sensors and actuators, multi-access bus protocols, and processors literally
2,000 times faster, we have the luxury of being able to automate just about everything.  There is
no need any more for a manual "SCE to AUX" switch, because the computer(s) are programmed
to make such decisions by themselves, and directly control the necessary subsystems.

The same trend is visible in airliners.  The cockpit of an Airbus A380 is simpler than the cockpit of
a Boeing 747 from 20 years ago.  And the monster A380 and the much smaller A320 have such
similar cockpit layouts that much of the 'type' training a pilot gets is transferrable.  Starting an auxiliary
power unit on an airliner today is as easy as pressing one button on the overhead - the computer
does all the temperature and RPM monitoring to safely bring it up to speed, or shut it down if it does
a 'hot start' or some other failure.  Pilots used to do that by hand, and had all the necessary knobs and readouts to do it.

The main engine "Full Authority Digital Engine Control" (FADEC) on an airliner takes full responsibility
for the safe operation of an engine - the pilot would have a hard time damaging the engine even if he wanted to.

[Prober]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control. I think many people are overestimating the amount of "flying" this vehicle's crew is going to have to do in off nominal situations based on the paradigm set by other early generation vehicles. I highly doubt its going to have the ability to be manually flown to orbit with the hand controller while attached to the booster, and I'd expect that should an abort be called for, a maximally redundant system is going to fire the Superdracos long before an astronaut is even lifting their arm off the rest to reach for a button. Surely there will be overrides and options, as their should, but being a true fly by computer vehicle, every control input will really be voting with the computer about what the vehicle should do.

The thought that a crewmember is going to have to drill down into multiple pages of submenus or type in fiddly alphanumeric sequences to get something critically important done under high vibration or acceleration environments reflects poor system engineering that even a relative spaceflight newcomer like SpaceX is bound to avoid. A big red hand size button popping up that blinks "ABORT" is far more likely.

Absolutely.  When there are engines firing, there's a very limited number of things to be done in an emergency that a person could decide to do manually.

it's called training

[JasonAW3]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control. I think many people are overestimating the amount of "flying" this vehicle's crew is going to have to do in off nominal situations based on the paradigm set by other early generation vehicles. I highly doubt its going to have the ability to be manually flown to orbit with the hand controller while attached to the booster, and I'd expect that should an abort be called for, a maximally redundant system is going to fire the Superdracos long before an astronaut is even lifting their arm off the rest to reach for a button. Surely there will be overrides and options, as their should, but being a true fly by computer vehicle, every control input will really be voting with the computer about what the vehicle should do.

The thought that a crewmember is going to have to drill down into multiple pages of submenus or type in fiddly alphanumeric sequences to get something critically important done under high vibration or acceleration environments reflects poor system engineering that even a relative spaceflight newcomer like SpaceX is bound to avoid. A big red hand size button popping up that blinks "ABORT" is far more likely.

Absolutely.  When there are engines firing, there's a very limited number of things to be done in an emergency that a person could decide to do manually.

it's called training

I could go both ways on this argument.

Although the Challenger crew were exceedingly well trained, there was nothing that they could do to save themselves as it happend so fast.  Were there thremal and vibratory sensorsinplace either on the ET or the SRB, the onboard computers MIGHT have aborted them free of the ET and SRB's before the explosion.

On the other hand, the system on the Columbia realised that it was in trouble and did everything it could to save itself and the crew.  Had there been impact sensors along the leading edge of the wings, (either internally or disposable on reentry, they MIGHT have known about the damage soon enough to abort to a landing before making orbit, or been able to hold out long enough for a rescue mission (Yes, a consumables mission would have HAD to be launched to get the crew down safely).

Overall it has to be a balance of tech and training, as sometimes, things happen that no system is designed to compensate for.

[pagheca]

Although the Challenger crew were exceedingly well trained, there was nothing that they could do to save themselves as it happend so fast.  Were there thremal and vibratory sensorsinplace either on the ET or the SRB, the onboard computers MIGHT have aborted them free of the ET and SRB's before the explosion.

I guess someone more experienced than me will reply you about that, but I do not think this is really true. First, the STS was not designed to allow an early separation of the SRBs before they were exhausted. Second, as far as I know it was too early in the flight even for an RTLS abort mode with no MEs and SRBs: the OV would have anyway crashed into the sea, developing forces killing the crew.

On the other side, there were about 13-20 sec between the SRB internal pressure started to drop - first detectable sign of a malfunction from the existing on-board sensors - and the start of the destruction of the overall vehicle integrity. So, a differently designed vehicle could have actually automatically triggered an abort mode, but not the STS.

The point is that Dragon, to go back on topic, is designed from the start to allow an abort mode at any time during the flight. This, joint with the improvement in s/w and h/w in the latest years, allows to design secure automated abort modes procedure not available in the past.

p.s. I'm curious to see how many will pick-up wrong information in what I wrote above in the next minute...

[Jim]

Were there thremal and vibratory sensorsinplace either on the ET or the SRB, the onboard computers MIGHT have aborted them free of the ET and SRB's before the explosion.

No, nothing short of breakwires or burnthrough wires could have detected the issue and then there still was nothing that could be done.  The SRB's could not have been jettisoned while burning and the Orbiter could not leave the ET with either SRB's or SSME's burning.

[Jim]

On the other side, there were about 13-20 sec between the SRB internal pressure started to drop - first detectable sign of a malfunction from the existing on-board sensors -

Not enough to know that it was a problem.

[pagheca]

On the other side, there were about 13-20 sec between the SRB internal pressure started to drop - first detectable sign of a malfunction from the existing on-board sensors -

Not enough to know that it was a problem.

absolutely. I just got that to define the maximum theoretical time from when the on-board sensors started to detect a problem and the time things were definitely compromised. Anyway, thanks for this remark.

[obi-wan]

I realize that your point is even more valid at higher pressure but where did you get the value of "8.3 psi cabin pressure"?
That's not the interior pressure of the ISS, which is (a fairly heavy for some of us) 14.7 PSI.

There's a lot of discussion in the Exploration community on the ideal cabin atmosphere for BEO - an 8.3 psi 32% O2 atmosphere lets you denitrogenate just about as fast as you can put on and check out a 4.3 suit. I was mostly trying to give a most favorable case, but it probably also reflects my hope they look for opportunities to take Dragon beyond LEO as soon as practical...

[ChrisWilson68]

The cargo Dragon doesn't have any buttons, screens, controllers or displays inside and has managed to be recovered from anomalies just fine, thanks to automation and ground control. I think many people are overestimating the amount of "flying" this vehicle's crew is going to have to do in off nominal situations based on the paradigm set by other early generation vehicles. I highly doubt its going to have the ability to be manually flown to orbit with the hand controller while attached to the booster, and I'd expect that should an abort be called for, a maximally redundant system is going to fire the Superdracos long before an astronaut is even lifting their arm off the rest to reach for a button. Surely there will be overrides and options, as their should, but being a true fly by computer vehicle, every control input will really be voting with the computer about what the vehicle should do.

The thought that a crewmember is going to have to drill down into multiple pages of submenus or type in fiddly alphanumeric sequences to get something critically important done under high vibration or acceleration environments reflects poor system engineering that even a relative spaceflight newcomer like SpaceX is bound to avoid. A big red hand size button popping up that blinks "ABORT" is far more likely.

Absolutely.  When there are engines firing, there's a very limited number of things to be done in an emergency that a person could decide to do manually.

it's called training

All the training in the world doesn't change the fact that there's a very limited number of useful things a person could do in an emergency while the engines are firing.  Any particular contingency that you could train for, the computer can be programmed to react to automatically.

[oiorionsbelt]

Of course that's not the final hatch mechanism or seal. Or do you think that the final capsule will have a neat car-like chrome handle to open it up? I drink the "SpaceX koolaid" more than most, but the gullibility in this thread is pretty staggering.

The hatch handle and knobs were polished on a whim and may or may not be in the final prep checklist. 

[meekGee]

Of course that's not the final hatch mechanism or seal. Or do you think that the final capsule will have a neat car-like chrome handle to open it up? I drink the "SpaceX koolaid" more than most, but the gullibility in this thread is pretty staggering.

The hatch handle and knobs were polished on a whim and may or may not be in the final prep checklist. 

...  and the seats, and the screens, and the motors ... 
Turns out, sometimes the simplest thing is simply to listen to the Kremlin.

[JBF]

For everyone who was complaining about armrests take a look at this shuttle video. http://forum.nasaspaceflight.com/index.php?topic=34958.msg1208450;boardseen#new No armrests.