There are several Paternosters in operation where I work (HBC Europe HQ in Cologne), and they are very cleverly designed in subtle ways to work around the likelihood of an execution.
"World's first 3D people mover" might be a more appropriate --- and enticing --- title, because "rope-free" can also refer to hydraulic elevators (which have been around for over a century.)
As far as I can see, the design only allows 2D movement. If you wanted the lift to move toward or away from the camera, you would need either some kind of 2nd attachment point on a different side.
Or a vertical segment could have a curve too. Imagine rotating by 45 degrees as you go up ten floors -- suddenly now you're in a completely different plane.
We could also check whether "The Willy Wonka elevator" arouses more curiosity: It's a feature of the Charlie and the Chocolate Factory book (and in the movie from Tim Burton).
3D Railway Pod? Seems like it follow tracks and switch lines like a train only it can also go up and down. It's a step close to the Great Glass Elevator.
Getting closer to Turbolifts. Canonically they form a network throughout the ship, use induction motors, can also go sideways, have multiple cabins in the shafts, switch routes etc.
No inertial dampeners of course.
No problem either - do it like on freefall towers or on other park attractions, where eddy current brakes, fully passive, are employed as safety mechanism.
All elevators have multiple safety brakes working on their guide rails (this is the famous invention of Otis); so do these. Besides batteries lasting longer (two hours, they say) than the median European power outage.
Plus hopefully battery banks (which you'll likely want on the elevators anyway to be able to store energy from regenerative braking) to power the car long enough to come to a rest where the passengers can get out in the event of a power outage.
There's going to be several layers of safeguards, like there are with current elevators.
The cable/belt mechanism of elevators has remained largely unchanged for most of it's history because of its simplicity in design, manufacturing, and maintenance. This new system is absolutely beautiful, but I agree, it looks expensive. Elevators have to work 24/7 with little maintenance, it will be very interesting to see this in the real world.
Looks to be about.... 10? times the cost of a normal elevator if I had to guess. I bet the maintenance costs are fun as well. As such, doubt you will see these except in niche buildings (things like the porsche tower, etc..)
The tallest skyscrapers have issues with elevators because the cables can only be so long before they snap under their own weight, so you need elevator changing floors, which take up space and lower throughput. Either you find flexible materials with a higher tensile-strength to weight ratio than steel or you forego cables entirely.
Classic elevators also need a lot of floor space to achieve good throughput. This system is basically like a paternoster on steroids; it can rise much higher, move much faster, use arbitrary circuits and has none of the safety issues. Even if this system only saves, say, 40 % floor space, it'd still be a win for many projects.
So many negative comments for a Saturday. This is a marvel of engineering and deserves a celebration. Humbled by our ability to re-invent more and more. Well done to all involved.
I hope HN could avoid mindless cheering and positivity related to technology matters. Negative or positive comments that only express private feeling should be discouraged. They reduce the value of the discussion.
Interesting technological discussion means criticizing.
I feel like they didn't really do much to motivate this.
Some of the comments here about increasing throughput provide clearer motivation - they do show this super briefly in the video with elevators operating in a cycle, but the benefits didn't really click.
This might be a little bit premature. I'd want to see broad commercial success before celebrating it as a marvel of engineering.
Systems like this have been invented repeatedly throughout the decades, but they never succeeded for one reason or another. The problem isn't in coming up with the concept (or, now, making animated 3D renderings), it's in making it actually work well.
This, incidentally, is why so many of us are so ecstatic about SpaceX -- because they've made something amazing that actually works.
Eh... Rube Goldberg Machines are often celebrated as being of interesting design, and they are created with the expressed purpose of being overly complicated. There is room in the world for very clever engineering, very complicated engineering, very elegant engineering, and everything in between.
I hate to be so cynical, but for how often the TK elevator in our building is out of service (these things have been around for 100 years and still break down more than a Fiat), I don't have the confidence necessary to step foot in this thing.
Note that a TK elevator doesn't mean it's actually still maintained by TK. For example in my previous building there was an Otis elevator but at some point the maintenance stickers on it became TK's, probably because they offered a cheaper maintenance contract.
TLDR: the manufacturer can't do much about the owner not maintaining their product properly.
Came to post a similar comment. Of course it's anecdotal, but in our old building we had a +20 years old Schindler's lift (yes that sounds very close to Schindler's List, the joke is never far away) and in the 5 years I was there I never saw them failing. In a new building no single month passes by without one of the TK elevators needing servicing, at worst multiple times a week. And if they're in service there are still always chances they forget the floor entered or one of the special programs for limited access areas stops working. Definitely software problems but might also be hardware related sometimes. Like the doors sliding open a couple of mm then closing again, until the technician shows up once again. I know that doesn't mean all their products are like that and maybe human error is at fault here, but it ceryainly does not leave the same good impression as the Schindler's ones.
Weird you say that. Fiat engines are so reliable most other constructors use a FIRE or a common rail derivative in their compact segment. Here there are uncountable 90s and 00s punto panda and the likes, still marching after few couple hundred thousand kilometers. Drove one myself to 250000, with minimal issues and regular maintenance. Now driving a "famously unreliable" Alfa and just passed 100kkm with zero issues yet. Never have been left on the road by them, and the alfa drove me all the way from milano to dublin and back, after multiple track days.
My dad was is an elevator repair man, has been for the last 33 years. When I asked him what he thought about his job, he said "just like any other, it has its ups and downs"
But after seeing this, now we are going back and forth on the future of the industry!
They have to have a good plan for cases when things break and electricity is not available. Firemen and maintenance crews must be able to rescue people from the elevators stuck in positions with minimum training and tools.
The real problem I see with this, is the complexity of the components required to make anything seen in this video is orders of magnitudes more that vertical shaft versions -- thus the cost of an elevator will me massive.
What are the seismic implications - the tolerances on the XY intersections are surely tight.
How the hell would you inspect some of the complex systems.
You would need a sensor car that just roams the tracks measuring tolerance levels - make it a service car that repairmen can be in which has no walls. You'd be more cost effective putting all your large sensor objects in a car as opposed to thousands of them throughout the system
What happens when a rotating intersection piece fails and its the only route a car can take/the piece fails with a car on that portion of track - there is no door/escape/access to the passengers in that scenario.
Also - the building would require extra height to accommodate horizontal passages if the intersection between XY cant happen at ingress/egress points. If they DO happen at such points - what happens when E1 wants to go right and E2 wants to go left. one car need to give way, then get back on track to go... routing conflicts could occur frequently - like the skyway connectors between towers where that Sim takes the car from tower1 to tower2 but he gets out of the middle of a row of doors....
And as mentioned, elevators are already not that cheap - I don think all but the Nakatomi Plaza and Ono Sendai HQ buildings could afford these yet...
Actually - There are buildings with primary structural elements on there external (HSBC Building in Hong Kong, for example) -- So I could see this as an external bolt-on layer to the side of a building - as opposed to a network of tunnels and shafts throughout - but then that eliminates a majority of the horizontal movement, unless it wraps onto two sides of a building...
Accelerating and decelerating in line with gravity is a lot more forgiving on the occupants than lateral accelerations. Seems like this would put a big limit on horizontal movement.
Depends on whether or not the car can 'tilt' while it is moving sideways (I know they don't show it doing so). If it can then any amount of acceleration can be presented as 'up' and 'down' force to the occupants much like a plane.
Less than you'd think, if you design it right. You only need to gimbal the car and make it bottom-heavy. Still more complex and expensive than not doing it, but no computers or actuators are required.
When combined with the acceleration they are indistinguishable. I got a chance to ride in the full motion simulator at NASA Ames and it used this 'trick' to recenter itself after it had made a left or right motion. Inside you felt nothing, like the system was unmoving but outside it was moving the flight pod back to the center of motion so that it had maximum reach for the next maneuver.
I'm sure the TK folks could manage it. The trick is that the 'tilt' of the car on the linear motor would add an additional degree of freedom to deal with. And yet it would be expensive but they seem to have already crossed that bridge :-)
True, but I guess that enough value will lie in just being able to put a whole fleet of cabins into a small number of one way shafts. Horizontal travel with people inside is a gimmick, but a very illustrative one. I expect early adopters to overuse it for showing off and later regret it quite a bit, for the reasons you stated.
Actually they say that horizontal movement is critical. Because in the morning when all elevators go up, they need another shaft to return them down and make a loop.
Like buses and trams put a limit on horizontal movement? ;) Honestly though, as long as the change of direction is somehow indicated in advance, and the acceleration isn't too abrupt, there shouldn't be anything limiting the horizontal travel distance.
My ideal elevator would accelerate at a pleasant 4 m/s^2 and have a jerk not exceeding 4 m/s^3. It would have doors that close instantly when the doorway is clear, and start right away.
That would be able to take me up a 100 yard high building in 11 seconds.
Why can't real elevators be like that? Why must the doors move slowly, and wait for 5 seconds before even starting moving? Why does the elevator reach a sluggish top speed, even with very little load? Why when arriving at the destination floor does it take a further 2 or 3 seconds for the door to open?
The door closing mechanism is designed with limited power and there's an optical sensor to prevent it from even trying to close on someone. Why wouldn't that still be the case on a fast closing door?
More tellingly: fast closing doors are absolutely a thing, so clearly the safety issues are not insurmountable, but I only see them in more modern/expensive locations. I suspect it's a price grading move and if you wait another century fast-closing doors will be everywhere.
Yeah but the faster they go the more inertia the whole thing has to handle for those safety mechanism, right? And the added machinery weight impacts on the elevator loading as well.
Yes, but there are ways of dealing with inertia. Like making the door lighter and putting padding on the contact surfaces. It's not rocket science.
Also, most of the improvement would come from replacing the worst offenders (5-10 second closing times) rather than from shaving 1 second times to 1/2 a second, and there's tons of precedent for the former from the faster-closing 10% or so of existing installations.
A sensor to detect a clear doorway is straightforward.
Doors that have sensitive rubber edges that can detect hitting something and engage a lock within milliseconds are also trivial.
The two methods combined mean that you have safety for soft human fleshy bits incase either malfunctions, yet the door can still be moving at 15 mph or more and close in 100 milliseconds.
Don't forget the part where they disable the door close button and add a 10 second wait time!
My favorite bit, though, is where they lock the escape hatch from the outside and provide no "alarm will sound" override mechanism. That way if the elevator dies you'll be stuck there for hours until they can get the fire department on the scene.
Forcing people to be stuck waiting for hours is usually preferable to having untrained and untethered people climbing around active elevator shafts N floors up.
If the elevator ascends too fast, riders' ears will pop, and some people, especially those with colds, will experience ear pain. Some high speed elevators (Tapei 101, at least) pressurize the elevator cab so they can control the rate of air pressure change. But that runs the cost way up.
Doors have to close slowly for safety. There's a limit on how much kinetic energy (7 foot-pounds, from memory) can be in a closing door. Opening can be much faster, and on high-speed elevators, it usually is.
At the destination floor, there's often a delay for precise leveling. Position feedback for leveling is just limit switches until you get to the really fancy high-speed elevators, so it works by shifting to slow speed and inching until the switch trips. Fancier elevators used analog optical systems and inductive systems that gave an analog signal of leveling error. Here's a modern elevator position sensor, with a bar code strip running the height of the shaft.[1]
Precision position and speed control of large motors was really hard in the 20th century.
I love that position sensor. It doesn't say how the code on the strip is determined, but I'd like to imagine it's some kind of gold code so the sensor can find the position even if 99% or more of the code is occluded with dust or dirt.
what problem does this solve to justify the crazy expense, complexity and maintenance?
the reason elevators exist is because people cannot easily climb 100+ vertical ft, especially while carrying stuff.
on the other hand, humans are well adapted for horizontal travel - 15mi per day without breaking a sweat. and when they're not, there are movable walkways like in airports.
this whole thing is a solution looking for a problem, IMO. we don't yet live in Matrix-style human incubators with miles of horizontal travel in buildings that are also miles high - it's usually either/or, and not because we've been lacking these revolutionary elevators.
If those elevators worked they would actually solve a big engineering problem: Very High Buildings can only have a limited number of elevator shafts. But it takes a significant time to use them which makes it problematic if there are a lot of people who want to go. But if elevators were able to travel horizontally that is a first step towards having multiple elevators in one shaft. There are still other problems that need to be solved in order to do this but if it works you can easily transport a lot of people in skyscrapers by having one "upwards" shaft and one "down" shaft which the elevator capsule can change between by moving horizontally.
you can have dedicated up and down shafts where the cars move horizontally only at the top and bottom, like a ferris wheel. without having them move horizontally at any floor. you just need a way to disconnect the rope (like ski lifts). or you can have towing cars that have only 50 ft of rope and operate on rails (like ships that get towed through Panama locks)
> what problem does this solve to justify the crazy expense, complexity and maintenance?
Elevators have a natural height limit, due to the weight of the rope, which is why taller buildings need sky lobbies. You also can't (really) increase the capacity of a single elevator shaft with a traditional rope-based system.
It solves the problem that you can only have on cabin (or two) in an elevator shaft. That's very inefficient in super-tall buildings. With horizontal travel elevator cabins can go out of their way and enable much better usage of elevator shafts. Further you become independent of the total length of the elevator shaft. So you drop another limiting factor to super-tall buildings.
Elevators share a similar problem to rockets. Only instead of the rocket tyranny, it's the elevator shaft tyranny. At a certain point, making a building taller can subtract from the square footage because of the extra shafts required.
With a linear motor system, more elevators can be added without adding shafts. Part of making that possible is allowing elevators to change shafts.
the limiting factor is that shafts havr to accommodate bidirectional movement and using a rope only allows 1 per shaft.
you can solve this by having the cars on rails and having unidirectional shafts where the cars only travel horizontally at top and bottom. this would allow muliple elevators per shaft.
I think this solves a number of problems. Foremost the horizontal movement between buildings. China has a lot of these towers and typically there's skyways to walk across. An elevator that moves horizontally could make it easier to move from 1st of one building to 30th of another. Another point made in the video is that the elevators can move around one another. Solving the issue we have now of a single shaft moving one car to multiple, busy destinations (i.e. the morning rush) allowing multiple cars to use the same shaft to work around one another. All in all I think this is aiming to solve the two greatest problems: time, and cost.
This will be how we build towers twice as tall as existing megatowers.
As beautifully explained by xkcd[0], "If your building has lots of floors, you need lots of different elevators, since there would be so many people trying to come and go the same time. If you make a building too tall, the whole thing gets taken up by elevators and there's no space for regular rooms."
But if one elevator shaft can have multiple independent cars, suddenly you've got a huge capacity multiplier without the space costs.
The horizontal motion part is cute, but I don't think it will matter nearly as much.
The car motion is not independent and it's hard to use the capacity well. Consider the morning rush in a big office tower when everyone starts on the ground floor. You can only use the bottom car to move these people.
You need a sideways dimension of movement to avoid this problem.
Oh, definitely, I agree. You need at least 2 (probably 3 or 4 for extra redundancy) shafts. But you need them to be usually one-way shafts. So everyone going up gets in at the same door, and the empty cars can fly down the other side to be ready to take more people up.
This is exciting possibility, but I would expect the builders will try to minimize the volume taken by shaft. Corridors are less expensive to build and maintain than these high-tech shafts.
I'd be interested in why they decided to make the 'track' as complex as they did. The transition nodes are an amazing mechanism but it seems also a likely failure point.
205 comments
[ 2.6 ms ] story [ 191 ms ] thread-the Paternoster (http://www.audioguideportal.de/assets/Foto/thumbnail520x397/...)
The post title is just clickbait for a ThyssenKrupp commercial, plus it only shows animations of these new elevators. Yawn.
https://www.youtube.com/watch?v=E7QlAsxJP-g#t=58s
Also, they could have a completely circular horizontal segment to move people through an upper level of a round building - like a stadium...
And back on-topic, googling to find the details instead of just a flashy video, here's Tom Scott on MULTI: https://www.youtube.com/watch?v=kdTsbFS4xmI
not sure there's anything "revolutionary" here.
There's going to be several layers of safeguards, like there are with current elevators.
Interesting technological discussion means criticizing.
Some of the comments here about increasing throughput provide clearer motivation - they do show this super briefly in the video with elevators operating in a cycle, but the benefits didn't really click.
Systems like this have been invented repeatedly throughout the decades, but they never succeeded for one reason or another. The problem isn't in coming up with the concept (or, now, making animated 3D renderings), it's in making it actually work well.
This, incidentally, is why so many of us are so ecstatic about SpaceX -- because they've made something amazing that actually works.
And it has it's first production project in the OVG East Side Tower Berlin (though not yet finished building)
TLDR: the manufacturer can't do much about the owner not maintaining their product properly.
But after seeing this, now we are going back and forth on the future of the industry!
Wonder what tech they use for traction while moving vertically.
What are the seismic implications - the tolerances on the XY intersections are surely tight.
How the hell would you inspect some of the complex systems.
You would need a sensor car that just roams the tracks measuring tolerance levels - make it a service car that repairmen can be in which has no walls. You'd be more cost effective putting all your large sensor objects in a car as opposed to thousands of them throughout the system
What happens when a rotating intersection piece fails and its the only route a car can take/the piece fails with a car on that portion of track - there is no door/escape/access to the passengers in that scenario.
Also - the building would require extra height to accommodate horizontal passages if the intersection between XY cant happen at ingress/egress points. If they DO happen at such points - what happens when E1 wants to go right and E2 wants to go left. one car need to give way, then get back on track to go... routing conflicts could occur frequently - like the skyway connectors between towers where that Sim takes the car from tower1 to tower2 but he gets out of the middle of a row of doors....
And as mentioned, elevators are already not that cheap - I don think all but the Nakatomi Plaza and Ono Sendai HQ buildings could afford these yet...
Actually - There are buildings with primary structural elements on there external (HSBC Building in Hong Kong, for example) -- So I could see this as an external bolt-on layer to the side of a building - as opposed to a network of tunnels and shafts throughout - but then that eliminates a majority of the horizontal movement, unless it wraps onto two sides of a building...
I'm sure the TK folks could manage it. The trick is that the 'tilt' of the car on the linear motor would add an additional degree of freedom to deal with. And yet it would be expensive but they seem to have already crossed that bridge :-)
That would be able to take me up a 100 yard high building in 11 seconds.
Why can't real elevators be like that? Why must the doors move slowly, and wait for 5 seconds before even starting moving? Why does the elevator reach a sluggish top speed, even with very little load? Why when arriving at the destination floor does it take a further 2 or 3 seconds for the door to open?
1- no one has bothered.
2- it's been tried, something terrible happened and now they don't try it anymore.
3- it's been tried, but the extra cost doesn't cover the extra benefit. No one wants to pay for it.
All that said, I was at a hotel once that had, to my mind, perfect acceleration like you describe. No one else cared, but I was blown away by it.
More tellingly: fast closing doors are absolutely a thing, so clearly the safety issues are not insurmountable, but I only see them in more modern/expensive locations. I suspect it's a price grading move and if you wait another century fast-closing doors will be everywhere.
Also, most of the improvement would come from replacing the worst offenders (5-10 second closing times) rather than from shaving 1 second times to 1/2 a second, and there's tons of precedent for the former from the faster-closing 10% or so of existing installations.
Doors that have sensitive rubber edges that can detect hitting something and engage a lock within milliseconds are also trivial.
The two methods combined mean that you have safety for soft human fleshy bits incase either malfunctions, yet the door can still be moving at 15 mph or more and close in 100 milliseconds.
My favorite bit, though, is where they lock the escape hatch from the outside and provide no "alarm will sound" override mechanism. That way if the elevator dies you'll be stuck there for hours until they can get the fire department on the scene.
So my guess is that the slowness has to do with safety.
Doors have to close slowly for safety. There's a limit on how much kinetic energy (7 foot-pounds, from memory) can be in a closing door. Opening can be much faster, and on high-speed elevators, it usually is.
At the destination floor, there's often a delay for precise leveling. Position feedback for leveling is just limit switches until you get to the really fancy high-speed elevators, so it works by shifting to slow speed and inching until the switch trips. Fancier elevators used analog optical systems and inductive systems that gave an analog signal of leveling error. Here's a modern elevator position sensor, with a bar code strip running the height of the shaft.[1]
Precision position and speed control of large motors was really hard in the 20th century.
[1] https://www.cedes.com/documents/produkte/Broschueren/108783e...
the reason elevators exist is because people cannot easily climb 100+ vertical ft, especially while carrying stuff.
on the other hand, humans are well adapted for horizontal travel - 15mi per day without breaking a sweat. and when they're not, there are movable walkways like in airports.
this whole thing is a solution looking for a problem, IMO. we don't yet live in Matrix-style human incubators with miles of horizontal travel in buildings that are also miles high - it's usually either/or, and not because we've been lacking these revolutionary elevators.
with a lot of cars in uni-directional shafts, is this really an issue?
> all cars must wait each time a car reaches a floor
maybe solved by a 2-lane shaft with a passing lane. very doable with rails or towing cars.
Elevators have a natural height limit, due to the weight of the rope, which is why taller buildings need sky lobbies. You also can't (really) increase the capacity of a single elevator shaft with a traditional rope-based system.
With a linear motor system, more elevators can be added without adding shafts. Part of making that possible is allowing elevators to change shafts.
you can solve this by having the cars on rails and having unidirectional shafts where the cars only travel horizontally at top and bottom. this would allow muliple elevators per shaft.
As beautifully explained by xkcd[0], "If your building has lots of floors, you need lots of different elevators, since there would be so many people trying to come and go the same time. If you make a building too tall, the whole thing gets taken up by elevators and there's no space for regular rooms."
But if one elevator shaft can have multiple independent cars, suddenly you've got a huge capacity multiplier without the space costs.
The horizontal motion part is cute, but I don't think it will matter nearly as much.
[0]https://what-if.xkcd.com/94/
You need a sideways dimension of movement to avoid this problem.
you mean like the elevator system being discussed here?
> But if one elevator shaft can have multiple independent cars, suddenly you've got a huge capacity multiplier without the space costs.
> The horizontal motion part is cute, but I don't think it will matter nearly as much.
My point is: 1. Stacked cars are not independent. 2. Horizontal motion matters a lot.
I'd be interested in why they decided to make the 'track' as complex as they did. The transition nodes are an amazing mechanism but it seems also a likely failure point.
Looks like it will be crazily expensive.
https://www.youtube.com/watch?v=kdTsbFS4xmI
I. Set the stage by showing a trend.
II. Describe two premises and derive a logical conclusion from them.
III. Show a product that not only is the objective answer to the conclusion above but also fits perfectly into the initial narrative.