SpaceX Dragon 2 Updates and Discussion - Thread 3

[joek]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Yes, but that particular example might be generalized to... What factors might be critical input to a decision which automation is unaware, or would be difficult to make automation aware?

For example, you have injured crew and a couple options... fast and hard or slow and soft?  Automation is unlikely to be able to make such decisions in a reasonable manner for the foreseeable future.  Humans will make such decisions (on-board crew, ground, or both); automation will execute the details of the decision.
 
That is why there will always be humans-in-the-loop at some level.  Less hands-on than traditional stick-and-rudder pilots, but more than passive meat-sack cargo.

[woods170]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

[JamesH65]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

[woods170]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

[Alexphysics]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

Crew health is checked at MCC-X, right? That means the engineers at mission control would tell the crew if the best option is to abort or not. If the crew member with the health issue can wait for orbital insertion and then go through a normal reentry, IMO that would be better than going through high g's either during the abort or during reentry.

[mn]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

I find it hard to believe that SpaceX did not want a manual abort capability, there's Hubris (aka Tesla Model X) and then there's Hubris^2

Computers are blind, deaf and dumb, they know only what was anticipated by the engineers as possible failure modes and correctly instrumented and programmed to handle. They have 0 situational awareness (actually 'understanding' what is going on). Humans are certainly much slower than computers and they can (and often do) make mistakes, but they still have many advantages over computers that will not go away for a very long time.

I can see an argument about 'manual control' ability, but to argue that they don't need at least a button that says 'abort' (and even that still relies on the computers to execute the abort based on where they are in the flight profile, etc) is absolutely mind boggling.

[chrisking0997]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

Crew health is checked at MCC-X, right? That means the engineers at mission control would tell the crew if the best option is to abort or not. If the crew member with the health issue can wait for orbital insertion and then go through a normal reentry, IMO that would be better than going through high g's either during the abort or during reentry.

I think this line of thinking is well past the point of reason.  What possible health issue that pops up during launch could somehow be rectified by an abort?  its not like once the chutes pop another crew member can unbuckle and assist the person with the issue.  At the very earliest assistance could only be given once the craft is safely back on earth, and even then could be delayed by several minutes due to procedures, hindrance by suits/capsule environment, lack of required medical equipment, etc. 

Trying to negate tiny one-off problems like this is going to result in over engineering the system to the point that no one is going anywhere.

[RDMM2081]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

Crew health is checked at MCC-X, right? That means the engineers at mission control would tell the crew if the best option is to abort or not. If the crew member with the health issue can wait for orbital insertion and then go through a normal reentry, IMO that would be better than going through high g's either during the abort or during reentry.

I think this line of thinking is well past the point of reason.  What possible health issue that pops up during launch could somehow be rectified by an abort?  its not like once the chutes pop another crew member can unbuckle and assist the person with the issue.  At the very earliest assistance could only be given once the craft is safely back on earth, and even then could be delayed by several minutes due to procedures, hindrance by suits/capsule environment, lack of required medical equipment, etc. 

Trying to negate tiny one-off problems like this is going to result in over engineering the system to the point that no one is going anywhere.

I agree it is a stretch at this point.  I was merely trying to come up with reasons for the existence for a manual abort switch/lever/button.  I am well past my instinct that there should be manual "flight controls" and just trying to poke the corners of the envelope to (dis)prove the need for some who seem to insist in favor of these manual controls.

[Alexphysics]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

Crew health is checked at MCC-X, right? That means the engineers at mission control would tell the crew if the best option is to abort or not. If the crew member with the health issue can wait for orbital insertion and then go through a normal reentry, IMO that would be better than going through high g's either during the abort or during reentry.

I think this line of thinking is well past the point of reason.  What possible health issue that pops up during launch could somehow be rectified by an abort?  its not like once the chutes pop another crew member can unbuckle and assist the person with the issue.  At the very earliest assistance could only be given once the craft is safely back on earth, and even then could be delayed by several minutes due to procedures, hindrance by suits/capsule environment, lack of required medical equipment, etc. 

Trying to negate tiny one-off problems like this is going to result in over engineering the system to the point that no one is going anywhere.

Eh, eh, eh, don't get me wrong, I'm pro automation, I was talking only about the Crew Dragon, take that clear, that capsule has manual abort capability so I was just talking about that special and single case and not in general. Also I don't know too much about what are the procedures they're going to do and things like that so it was more like a question than anything. In general, as I said, I think we should push further with automation. That is all.

[ncb1397]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Who says the manual abort is triggered by alarms? Could be a video feed on the booster, audio cues, sense of orientation, etc. If the avionics are installed wrong, the booster could be failing because of the avionics and think it is doing the right thing as in the 2013 Proton failure when all redundant angular velocity sensors were installed wrong. The avionics aren't as smart as you think they are. Probably have the intelligence equivalent to a tape worm. They can't interpret all sensor data available like video/audio feeds.

Unless there is a huge problem with abort happy astronauts (which for the record, there isn't), there is no logical reason to not have basic safety features like manual abort switches.

[woods170]

What about a case where a crew member was having a health issue as a result of the stress of launch?  Could this be a case to cause the commander to initiate a manual abort that the computer would not otherwise initiate?

Sensors and computers will know all about this and react much faster than any sort of human intervention. Human in loop is SLOW....

The abort routines are NOT monitoring crew health. They are monitoring launch vehicle health.

Crew health is checked at MCC-X, right? That means the engineers at mission control would tell the crew if the best option is to abort or not. If the crew member with the health issue can wait for orbital insertion and then go through a normal reentry, IMO that would be better than going through high g's either during the abort or during reentry.

I think this line of thinking is well past the point of reason.  What possible health issue that pops up during launch could somehow be rectified by an abort?  its not like once the chutes pop another crew member can unbuckle and assist the person with the issue.  At the very earliest assistance could only be given once the craft is safely back on earth, and even then could be delayed by several minutes due to procedures, hindrance by suits/capsule environment, lack of required medical equipment, etc. 

Trying to negate tiny one-off problems like this is going to result in over engineering the system to the point that no one is going anywhere.

I agree it is a stretch at this point.  I was merely trying to come up with reasons for the existence for a manual abort switch/lever/button.  I am well past my instinct that there should be manual "flight controls" and just trying to poke the corners of the envelope to (dis)prove the need for some who seem to insist in favor of these manual controls.

Emphasis mine.

Neither the original design for Crew Dragon, nor the original design for Starliner, had manual abort capabilities. Manual abort capabilities were added because NASA asked the CCP providers to do so.

[woods170]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

I find it hard to believe that SpaceX did not want a manual abort capability, there's Hubris (aka Tesla Model X) and then there's Hubris^2

Computers are blind, deaf and dumb, they know only what was anticipated by the engineers as possible failure modes and correctly instrumented and programmed to handle. They have 0 situational awareness (actually 'understanding' what is going on). Humans are certainly much slower than computers and they can (and often do) make mistakes, but they still have many advantages over computers that will not go away for a very long time.

I can see an argument about 'manual control' ability, but to argue that they don't need at least a button that says 'abort' (and even that still relies on the computers to execute the abort based on where they are in the flight profile, etc) is absolutely mind boggling.

Emphasis mine.

Is it?

Here is news for you. Soyuz doesn't carry a manual abort capability for ascent either. But NASA is regularly flying its astros on Soyuz.

The only reason why NASA requested the CCP providers to add a manual abort capability (for ascent) is because Apollo had one.

In both situations, where a manned spacecraft had to perform an ascent abort, it was done by the computer, way quicker than a human would have done.

Having a manual abort capability only serves to provide the crew with a false sense of safety IMO.

[litton4]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

I find it hard to believe that SpaceX did not want a manual abort capability, there's Hubris (aka Tesla Model X) and then there's Hubris^2

Computers are blind, deaf and dumb, they know only what was anticipated by the engineers as possible failure modes and correctly instrumented and programmed to handle. They have 0 situational awareness (actually 'understanding' what is going on). Humans are certainly much slower than computers and they can (and often do) make mistakes, but they still have many advantages over computers that will not go away for a very long time.

I can see an argument about 'manual control' ability, but to argue that they don't need at least a button that says 'abort' (and even that still relies on the computers to execute the abort based on where they are in the flight profile, etc) is absolutely mind boggling.

Emphasis mine.

Is it?

Here is news for you. Soyuz doesn't carry a manual abort capability for ascent either. But NASA is regularly flying its astros on Soyuz.

The only reason why NASA requested the CCP providers to add a manual abort capability (for ascent) is because Apollo had one.

In both situations, where a manned spacecraft had to perform an ascent abort, it was done by the computer, way quicker than a human would have done.

Having a manual abort capability only serves to provide the crew with a false sense of safety IMO.

Are you saying it's like those "close door" buttons in a lift (elevator) - ie it doesn't actually do anything, apart from providing the occupant with something to do, so they feel involved?

[jacqmans]

SpaceX rehearses helicopter landing and patient loading on its recovery ship, GO Searcher, practicing how the aircraft will pick up astronauts and fly them to a nearby hospital in the unlikely event of a medical emergency. The company outfitted the ship with a medical treatment facility and a helipad in the center of the vessel. When astronauts splash down into the ocean after their journey to the International Space Station on SpaceX’s Crew Dragon spacecraft, NASA and SpaceX doctors will work together to evaluate the crew onboard the vessel. Should astronauts need to be airlifted to a hospital, the helicopter also will pick up paramedics and doctors from the ship who will care for the astronauts in-flight.

Photo credit: SpaceX

[jacqmans]

[Lar]

Are you saying it's like those "close door" buttons in a lift (elevator) - ie it doesn't actually do anything, apart from providing the occupant with something to do, so they feel involved?

Pressing an active close door button almost never does anything bad (elevator doors recoil).
Pressing an active manual abort is very likely to do a bad thing except in very specific circumstances. Training can handle this.

Pressing an inactive close door button almost never does anything bad (other than annoying people who are tired of talking about whether close door buttons are real or not).
Pressing an inactive manual abort almost never does anything bad, except in the rare case when the software didn't make the abort decision but it actually is needful. Then it's really bad that it is not real.

I'm not sure I'd advocate for the button being a fake.

[born01930]

There will be events no one has planned for. Look at Apollo 12 or 13 (yes 13 was not a launch problem, but it certainly wasn't simulated) I can't recall which Apollo had the bad pogo which was remedied by the center engine failing.
In some cases you may want an auto abort override...you just never know. Not enough credit is given for the quick thinking of pilots. Armstrong in the LEM simulator is a good example a search of various test pilot anomalies can provide even more. There are examples of autopilots crashing planes (Air France). It's better to have it and never need it than to need it and not have it.

[programmerdan]

Not sure I understand all the handwringing; I'd suspect the majority of aborts would be telemetry, and if the auto-abort didn't catch it, humans are going to be in the loop back at control... with the telemetry, eyes on vehicle, and the like. The humans in the spaceship will have far less information about the total situation then either the auto systems or the humans back in control.

Adding the manual abort has positives and negatives (gives direct abort control to folks with a very narrow perspective on the total situation, but also gives them abort control if remote control is lost / fails) -- the trades make sense, and overall seems net positive to have added it, but there were humans involved at every level of every abort system (designers, engineers, programmers who built the autosystems, control w/ telemetry and comms...) with or without a manual abort button.

If I were flying on either crew vessel, I'd take far more comfort in knowing the depth of testing of the actual escape engine and flight dynamics testing, the parachute and other recapture systems, then the presence or absence of a button.

[mn]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

I find it hard to believe that SpaceX did not want a manual abort capability, there's Hubris (aka Tesla Model X) and then there's Hubris^2

Computers are blind, deaf and dumb, they know only what was anticipated by the engineers as possible failure modes and correctly instrumented and programmed to handle. They have 0 situational awareness (actually 'understanding' what is going on). Humans are certainly much slower than computers and they can (and often do) make mistakes, but they still have many advantages over computers that will not go away for a very long time.

I can see an argument about 'manual control' ability, but to argue that they don't need at least a button that says 'abort' (and even that still relies on the computers to execute the abort based on where they are in the flight profile, etc) is absolutely mind boggling.

Emphasis mine.

Is it?

Here is news for you. Soyuz doesn't carry a manual abort capability for ascent either. But NASA is regularly flying its astros on Soyuz.

The only reason why NASA requested the CCP providers to add a manual abort capability (for ascent) is because Apollo had one.

In both situations, where a manned spacecraft had to perform an ascent abort, it was done by the computer, way quicker than a human would have done.

Having a manual abort capability only serves to provide the crew with a false sense of safety IMO.

Very surprising, but if Soyuz doesn't have it I hereby humbly eat my words.

[envy887]

On Crew Dragon just about the only thing the crew can do to intervene in the computer-controlled ascent is to use the manual abort capability. Which will end the flight.

The flight avionics are so sophisticated that by the time the human brain can process the alarms it is hearing and decide to execute a manual abort the avionics will have already identified the fault, decided an abort is required, blown the hold-down bolts and have ignited the abort engines. The Crew Dragon will already be away from the Falcon 9 and under abort acceleration before the Commander can even form the words "abort".

Correct. But NASA insisted on having a manual abort capability nevertheless. And SpaceX obliged.

I find it hard to believe that SpaceX did not want a manual abort capability, there's Hubris (aka Tesla Model X) and then there's Hubris^2

Computers are blind, deaf and dumb, they know only what was anticipated by the engineers as possible failure modes and correctly instrumented and programmed to handle. They have 0 situational awareness (actually 'understanding' what is going on). Humans are certainly much slower than computers and they can (and often do) make mistakes, but they still have many advantages over computers that will not go away for a very long time.

I can see an argument about 'manual control' ability, but to argue that they don't need at least a button that says 'abort' (and even that still relies on the computers to execute the abort based on where they are in the flight profile, etc) is absolutely mind boggling.

Emphasis mine.

Is it?

Here is news for you. Soyuz doesn't carry a manual abort capability for ascent either. But NASA is regularly flying its astros on Soyuz.

The only reason why NASA requested the CCP providers to add a manual abort capability (for ascent) is because Apollo had one.

In both situations, where a manned spacecraft had to perform an ascent abort, it was done by the computer, way quicker than a human would have done.

Having a manual abort capability only serves to provide the crew with a false sense of safety IMO.

Very surprising, but if Soyuz doesn't have it I hereby humbly eat my words.

It has a manual abort option. However, it is not controlled by the crew in the vehicle, but by ground operators.