First, let me confess that it took me a while to realize that my arithmetic was completely wrong in coming up with the 6.3% of maximum thrust number. Oops.
While I have uncertainty both as to the desired fuel at landing (not zero) and as to the empty mass ratio of this stage, I now consider it likely that you are correct in suspecting that at the moment of landing the minimum thrust would provide more than 1G, and thus preclude hovering.
This does not, however, say anything about how hard an impact the landing need be. Rather it defines a requirement to control the final ignition timing (well above surface) and throttling profile through final descent. While fuel economy would push away from this, the "safest" choice would compute the final ignition altitude to be one for which throttling in the middle of the allowed range would give zero vertical speed at barge-deck (or landing pad) altitude. Then the actual throttling throughout the final descent would have half the range above and half the range below the planned nominal with which to adjust to conditions such as differences in the vertical wind component form planned, vertical rate consequences of horizontal adjustments, etc.
After all, airliners don't hover in landing, and given throttle response times, can't choose to at the last moment either, yet we expect fairly gentle landings and usually get them.
Unexpected throttle response leading to crashes has a precedent in the airliner business. I believe no fewer than four 727s were lost in crashes for which a primary element of similarity was that too-late throttle increase (coupled with much longer throttle response lag than the pilot's instincts expected) failed to taper off a high sink rate near the ground before impact. The 727 had been specifically designed to shorten the time duration of short flights by facilitating steep descents. Part of the fix was to back off on doing steep descents near the ground. Another part was to adjust the flight parameters to assure the engines were spooled up higher during final approach, thus shortening throttle response time. Perhaps the biggest part of the fix was just the passage of time--we don't have flocks of pilots who just transitioned from flying DC-4s and DC-6s flying jets anymore.
@Mikey : it not the choice of control set being an issue, it is the function of the ignition chamber within the rocket using non-hypergolic fuels. Sadly it is the case that below a certain threshold combustion will pop & fart & expire ( needing re-lighting ) quite regularly, and in that mode the behaviour is ridiculous for serious use. In detail this depends on the chemistry* of the chosen fuels.
Hypergolic fuels ignite on contact without a source of heat. These are great for intermittent use. Alas their specific impulse ( a measure of bang per mass of fuel ) is rather low. The non-hypergolic ones need the flame to exist from previous combustion ( or enough heat in the chamber walls etc ) to sustain reaction without re-light. So for the Merlin 1D the flame literally goes out under 70% throttle. I'm not aware of any designs that have the equivalent of a pilot light.
@archae86 : (a) I'd luv to know the detailed burn plan/algorithm for terminal descent ! I might poke around and see what might be publicly available. Don't hold your breath ! :-)
(b) Steep approaches are a hoot! But that is quite 'sporty' and finesse is the order of the day. To keep the engines spooled up but have the plane still descending you have to add drag. An very inefficient mode for sure, hence quite fuel expensive. Pop the flaps down full, raise the angle of attack and one can come on in. This requires much faith in the landing gear, and for that matter would prejudice the lifetime of even the best.
(c) Also : the 2 m/s^2 nett upwards at takeoff I had deduced as follows.
total thrust of Falcon 9 ~ 6000kN
total mass at liftoff ~ 500,000kg
upward acceleration = 6000/500 = 12
nett upward = 12 - 9.8 = 2.2 m.s^2
.... which at a glance is about right if you look at the initial separation from the pad on the videos.
@robl : that's what I have deduced. It's hard to have a rocket design that gives a payload 7km/sec to get into orbit AND also may softly, softly descend. But Elon is clever for sure. That's why he has nine engines per Falcon. Not only for redundancy but to later obtain 70% of 1/9th of full takeoff thrust in order to land. Again I'm not saying Elon has made an error, I'm saying he has made a choice.
However that suggests a useful variation : put an engine in the centre spot that is throttleable at a lower power range. But then you have see what effect that has on the boost back manoeuvre .... etc .... and so the consequences flow. Because you want the one basic design to do many things.
Cheers, Mike.
* Essentially there is a minimum temperature to get them to go. Catalysts like tri-ethyl borane lower that level. The hypergolic fuels don't need another molecule to assist.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
At touchdown, the thrust to weight ratio of the vehicle was greater than one, proving a key landing algorithm for Falcon 9.
.... while that is with a different mass craft and fuel available, it shows the hover-slam is the intended mode. Though I think that may mean the Grasshopper in testing must be more fueled/massed up than an empty Falcon Nine 1st stage post launch .... ???
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Interesting graphic, Mike. I had no idea the boostback burn looked like that.
6 m/sec maximum touchdown is not what most people would call soft. In our units, that is 19.7 feet/sec. US navy carrier aircraft have been required to tolerate up to 22 feet/sec for decades, but their normal touchdown (most would call it impact) is less than half that.
Bottom line, that graphic says the landing legs need to be nearly as sturdy as the famously sturdy naval aircraft landing gear.
I'll bet someone will be trying to get the team to commit to a significantly lower maximum landing sink rate in order to allow weight to be taken out of the landing gear and the structure to which it transfers the load.
Yeah it's a boost-up-and-back burn in reality, so that pegs back the tangential component to yield a purely radial total vector ( wrt the rotating Earth, so that's still non-inertial but what the heck ). There is much subtlety in this topic. :-)
Certainly 6m/sec is a substantial task for the landing gear. Mind you if they are separately fit-able then a shorter lifetime wrt the whole frame is still OK.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike, my confusion comes/came from the use of "hover slam". The two videos in the link you provided along with the graphic are of two very early prototypes which I had not seen and are titled "hover/slam". If by "hover/slam" the descent from space will provide a similar landing in the last seconds before touchdown then we have a winner. The last two failed barge landings are more descriptive of a "hover/slam" or "hover ricochet" landing. 6 meters/sec seems high as archae86 pointed out. Its hard to say but in all of the prototype videos it does not seem like ground contact occurred at 6 meters a second but at a slower rate. How much deformity, if any, in the 1st stage integrity is acceptable?
In thinking about a maximum 6 meters/second, I thought about "bounce", then about the need for shock absorbers. So I was comforted that on following Mike's link below the graphic I found myself reading of "landing legs with hydraulic dampers".
All this time I thought Mike made up the "hover/slam" terminology. Instead 'tis faithful transmission.
I'm still worried about it staying erect on the barge until the welders show up with either sea state or wind higher than gentle, but my impression of their ability to land on land at low crosswind has gone back up.
Mike, my confusion comes/came from the use of "hover slam". The two videos in the link you provided along with the graphic are of two very early prototypes which I had not seen and are titled "hover/slam". If by "hover/slam" the descent from space will provide a similar landing in the last seconds before touchdown then we have a winner. The last two failed barge landings are more descriptive of a "hover/slam" or "hover ricochet" landing. 6 meters/sec seems high as archae86 pointed out. Its hard to say but in all of the prototype videos it does not seem like ground contact occurred at 6 meters a second but at a slower rate. How much deformity, if any, in the 1st stage integrity is acceptable?
Fair point. A hover slam is basically the reverse timeline of takeoff in terms of velocities. Just as a rocket on the pad just barely lifts and then accelerates, so on return the 1st stage will severely decelerate ( upward marked as positive axis here ) and then vertically stop just barely above the pad. Perhaps slam-hover gives a better description, where the slam refers to the heavy deceleration ( aka the door-slammer of drag racing ) and hopefully not the contact with the barge ! Thus the hover is the aim of just getting to zero vertical velocity at zero vertical height. But that is an ideal so we accept a 'residual'. As discussed 6 m/sec is a decent residual .... :-)
Interestingly the most fuel efficient modes are the high acceleration ones in terms of point A to point B transit. Hover is quite inefficient obviously because you are chucking fuel away but going no-where. Thus in terms of going from ground to orbit or vice versa fuel concerns dictate harsh gees. For what it's worth the F9 User Manual specifies the limits that a payload must be rated for.
For that matter the word 'throttle' needs some discussion to. These days we think of our cars go pedal as the car's throttle. Which it is. But in history 'throttle' was used in the sense of 'choke' ie. reduce engine performance by it's application. But nowadays we use this in converse of the original sense thus 'hitting the throttle' means go faster not slower. Thus the old gag of a worse car than one which won't start is the one that won't stop ! :-0
Quote:
All this time I thought Mike made up the "hover/slam" terminology. Instead 'tis faithful transmission.
Tut tut ... :-) :-)
Imagine parking a car in slam-hover style. Slide right in under full brake lock, smoke billowing, long skid marks on the asphalt .... stopping 10mm short of the parking bay margin. Not a procedure likely to attract insurance coverage.
Cheers, Mike.
( edit )
Quote:
Imagine parking a car slam-hover ....
No. I absolutely refuse to look on YouTube, but I bet any money that some twit has done that.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Compare the Merlin 1D to the range of SuperDraco performance mentioned here. The stats there indicate throttling of a single SuperDraco from 20% to 100% ie. ~ 1300 N to 6800 N
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike, First, let me
)
Mike,
First, let me confess that it took me a while to realize that my arithmetic was completely wrong in coming up with the 6.3% of maximum thrust number. Oops.
While I have uncertainty both as to the desired fuel at landing (not zero) and as to the empty mass ratio of this stage, I now consider it likely that you are correct in suspecting that at the moment of landing the minimum thrust would provide more than 1G, and thus preclude hovering.
This does not, however, say anything about how hard an impact the landing need be. Rather it defines a requirement to control the final ignition timing (well above surface) and throttling profile through final descent. While fuel economy would push away from this, the "safest" choice would compute the final ignition altitude to be one for which throttling in the middle of the allowed range would give zero vertical speed at barge-deck (or landing pad) altitude. Then the actual throttling throughout the final descent would have half the range above and half the range below the planned nominal with which to adjust to conditions such as differences in the vertical wind component form planned, vertical rate consequences of horizontal adjustments, etc.
After all, airliners don't hover in landing, and given throttle response times, can't choose to at the last moment either, yet we expect fairly gentle landings and usually get them.
Unexpected throttle response leading to crashes has a precedent in the airliner business. I believe no fewer than four 727s were lost in crashes for which a primary element of similarity was that too-late throttle increase (coupled with much longer throttle response lag than the pilot's instincts expected) failed to taper off a high sink rate near the ground before impact. The 727 had been specifically designed to shorten the time duration of short flights by facilitating steep descents. Part of the fix was to back off on doing steep descents near the ground. Another part was to adjust the flight parameters to assure the engines were spooled up higher during final approach, thus shortening throttle response time. Perhaps the biggest part of the fix was just the passage of time--we don't have flocks of pilots who just transitioned from flying DC-4s and DC-6s flying jets anymore.
@Mikey : it not the choice of
)
@Mikey : it not the choice of control set being an issue, it is the function of the ignition chamber within the rocket using non-hypergolic fuels. Sadly it is the case that below a certain threshold combustion will pop & fart & expire ( needing re-lighting ) quite regularly, and in that mode the behaviour is ridiculous for serious use. In detail this depends on the chemistry* of the chosen fuels.
Hypergolic fuels ignite on contact without a source of heat. These are great for intermittent use. Alas their specific impulse ( a measure of bang per mass of fuel ) is rather low. The non-hypergolic ones need the flame to exist from previous combustion ( or enough heat in the chamber walls etc ) to sustain reaction without re-light. So for the Merlin 1D the flame literally goes out under 70% throttle. I'm not aware of any designs that have the equivalent of a pilot light.
@archae86 : (a) I'd luv to know the detailed burn plan/algorithm for terminal descent ! I might poke around and see what might be publicly available. Don't hold your breath ! :-)
(b) Steep approaches are a hoot! But that is quite 'sporty' and finesse is the order of the day. To keep the engines spooled up but have the plane still descending you have to add drag. An very inefficient mode for sure, hence quite fuel expensive. Pop the flaps down full, raise the angle of attack and one can come on in. This requires much faith in the landing gear, and for that matter would prejudice the lifetime of even the best.
(c) Also : the 2 m/s^2 nett upwards at takeoff I had deduced as follows.
total thrust of Falcon 9 ~ 6000kN
total mass at liftoff ~ 500,000kg
upward acceleration = 6000/500 = 12
nett upward = 12 - 9.8 = 2.2 m.s^2
.... which at a glance is about right if you look at the initial separation from the pad on the videos.
@robl : that's what I have deduced. It's hard to have a rocket design that gives a payload 7km/sec to get into orbit AND also may softly, softly descend. But Elon is clever for sure. That's why he has nine engines per Falcon. Not only for redundancy but to later obtain 70% of 1/9th of full takeoff thrust in order to land. Again I'm not saying Elon has made an error, I'm saying he has made a choice.
However that suggests a useful variation : put an engine in the centre spot that is throttleable at a lower power range. But then you have see what effect that has on the boost back manoeuvre .... etc .... and so the consequences flow. Because you want the one basic design to do many things.
Cheers, Mike.
* Essentially there is a minimum temperature to get them to go. Catalysts like tri-ethyl borane lower that level. The hypergolic fuels don't need another molecule to assist.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Actually here is quite a good
)
Actually here is quite a good graphic :
Plus if you read this SpaceX press release ( Grasshopper tests ) and note the comment :
.... while that is with a different mass craft and fuel available, it shows the hover-slam is the intended mode. Though I think that may mean the Grasshopper in testing must be more fueled/massed up than an empty Falcon Nine 1st stage post launch .... ???
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Interesting graphic, Mike. I
)
Interesting graphic, Mike. I had no idea the boostback burn looked like that.
6 m/sec maximum touchdown is not what most people would call soft. In our units, that is 19.7 feet/sec. US navy carrier aircraft have been required to tolerate up to 22 feet/sec for decades, but their normal touchdown (most would call it impact) is less than half that.
Bottom line, that graphic says the landing legs need to be nearly as sturdy as the famously sturdy naval aircraft landing gear.
I'll bet someone will be trying to get the team to commit to a significantly lower maximum landing sink rate in order to allow weight to be taken out of the landing gear and the structure to which it transfers the load.
Yeah it's a boost-up-and-back
)
Yeah it's a boost-up-and-back burn in reality, so that pegs back the tangential component to yield a purely radial total vector ( wrt the rotating Earth, so that's still non-inertial but what the heck ). There is much subtlety in this topic. :-)
Certainly 6m/sec is a substantial task for the landing gear. Mind you if they are separately fit-able then a shorter lifetime wrt the whole frame is still OK.
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike, my confusion comes/came
)
Mike, my confusion comes/came from the use of "hover slam". The two videos in the link you provided along with the graphic are of two very early prototypes which I had not seen and are titled "hover/slam". If by "hover/slam" the descent from space will provide a similar landing in the last seconds before touchdown then we have a winner. The last two failed barge landings are more descriptive of a "hover/slam" or "hover ricochet" landing. 6 meters/sec seems high as archae86 pointed out. Its hard to say but in all of the prototype videos it does not seem like ground contact occurred at 6 meters a second but at a slower rate. How much deformity, if any, in the 1st stage integrity is acceptable?
In thinking about a maximum 6
)
In thinking about a maximum 6 meters/second, I thought about "bounce", then about the need for shock absorbers. So I was comforted that on following Mike's link below the graphic I found myself reading of "landing legs with hydraulic dampers".
All this time I thought Mike made up the "hover/slam" terminology. Instead 'tis faithful transmission.
I'm still worried about it staying erect on the barge until the welders show up with either sea state or wind higher than gentle, but my impression of their ability to land on land at low crosswind has gone back up.
RE: Mike, my confusion
)
Fair point. A hover slam is basically the reverse timeline of takeoff in terms of velocities. Just as a rocket on the pad just barely lifts and then accelerates, so on return the 1st stage will severely decelerate ( upward marked as positive axis here ) and then vertically stop just barely above the pad. Perhaps slam-hover gives a better description, where the slam refers to the heavy deceleration ( aka the door-slammer of drag racing ) and hopefully not the contact with the barge ! Thus the hover is the aim of just getting to zero vertical velocity at zero vertical height. But that is an ideal so we accept a 'residual'. As discussed 6 m/sec is a decent residual .... :-)
Interestingly the most fuel efficient modes are the high acceleration ones in terms of point A to point B transit. Hover is quite inefficient obviously because you are chucking fuel away but going no-where. Thus in terms of going from ground to orbit or vice versa fuel concerns dictate harsh gees. For what it's worth the F9 User Manual specifies the limits that a payload must be rated for.
For that matter the word 'throttle' needs some discussion to. These days we think of our cars go pedal as the car's throttle. Which it is. But in history 'throttle' was used in the sense of 'choke' ie. reduce engine performance by it's application. But nowadays we use this in converse of the original sense thus 'hitting the throttle' means go faster not slower. Thus the old gag of a worse car than one which won't start is the one that won't stop ! :-0
Tut tut ... :-) :-)
Imagine parking a car in slam-hover style. Slide right in under full brake lock, smoke billowing, long skid marks on the asphalt .... stopping 10mm short of the parking bay margin. Not a procedure likely to attract insurance coverage.
Cheers, Mike.
( edit )
No. I absolutely refuse to look on YouTube, but I bet any money that some twit has done that.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
I've just noticed something
)
I've just noticed something on that Grasshopper video :
LOL. :-)
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Compare the Merlin 1D to the
)
Compare the Merlin 1D to the range of SuperDraco performance mentioned here. The stats there indicate throttling of a single SuperDraco from 20% to 100% ie. ~ 1300 N to 6800 N
Cheers, Mike.
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal