h a l f b a k e r y
Fewer ducks than estimates indicate.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
See first link. Ignore the dashed circle.
Start with an ordinary connected pair of hardware items, the bolt and nut. If you hold the nut stationary and turn the bolt, then the screw also moves partly through the hole in the nut. The "threads" outside the bolt and
inside the nut have a special shape that also is known as a "helix".
Let us take at least four helices, and possibly eight, and arrange them at the corners of an elevator cab. (Consider four LONG helices, each occupying the whole length of a vertical corner-edge of the cab.) One electric motor is arranged to be able to turn all of them at the same time, and at the same speed. Each helix meshes with an edge-corner of the shaft in a special way. Instead of high-friction metal-sliding-on-metal, as inside a nut-and-bolt, there are bearings. Again, see the first link. That dashed circle is about to become relevant. The portrayed contact points are a bit sketchy; there is no reason why the "lip" of the helix cannot be bent to make perfectly flat contact with the bearings. Also, it is perfectly possible (although not necessarily essential) that there be two sets of bearings in each shaft-corner, something like this:
In that top-view ASCII sketch, the shaft corner is at the upper-left. A bearing strut extends from the left side, a second bearing strut extends from the upper side, and the helix engages bearings on both struts. The dashed circle represents one bearing of the second group, and is there mostly to show how its location along the wall of the shaft is offset relative to the locations of the bearings of the other group. Still, it may not be essential to have two sets of bearings in each shaft corner; one set extending diagonally from the corner may be sufficient.
OK, since the bearings are firmly mounted to the walls of the elevator shaft, it follows that when the motor turns the helices, the elevator cab needs no cables to move itself up or down a shaft. (Please forego any "designed for screwing" jokes.) Obviously multiple cabs could occupy a single shaft. All they need are electric power, which can be supplied via "trolley" (or "streetcar") connections to power cables along the shaft wall. Each elevator cab motor should actually be a motor-generator, so that when any cab descends in a shaft, power can be generated to lift some other elevator cab.
Regarding elevator speed: This will depend on the spacing between the bearings and the rotation rate of the helices. One rotation of a helix will move the cab one bearing-spacing-unit. So, if the spacing is 0.1 meter (10cm), and the motor spins the helices at 3000RPM, then the cab moves up or down the shaft at 300 meters/minute. This is a fairly high speed for elevators, and probably completely adequate for any building.
Any torsional effects due to the rotating helices, that might cause a cab to try to escape from its bearing supports, is partly balanced by arranging the helices in counterrotating pairs. Also, just to be completely safe, the cab could copy the scheme used by roller coaster cars, and roll along a guide rail set into one shaft wall.
There should be a kind of electric DE-clutch that normally doesn't engage the elevator cab motor drive-mechanism. In case of power failure, though, it DOES automatically clutch (it uses a small amount of power to stay de-clutched), and thus locks the cab into its position in the shaft. However, any people inside the cab should have some means of manually loosening the clutch, so they can gradually descend to a floor where they can exit the cab.
Because this kind of elevator cab is a completely independent entity with respect to ordinary cabled elevators, it is possible to arrange an overall system that moves cabs between shafts. A single shaft can be completely devoted to UP elevators, and another shaft can be completely devoted to DOWN elevators. Spacing between the cabs should be such that ON THE AVERAGE, a cab that stops for passengers will start moving again before the next cab reaches that floor. Computer modelling can assist in determining that.
Portrayed at the link is a grouping of four elevator shafts (any even number can work). At the bottommost and topmost levels in the buliding (likely a subbasement and the roof), a kind of carosel exists. The large circle in the sketch represents a section of rotating FLOOR, and within the carosel, a 1-storey-high section of all four elevator shafts can rotate. So, two cabs arrive (either coming up to the top level or down to the bottom level), then while the doors open to let passengers exit/enter, the carosel rotates 90 degrees, and then the cabs can be on their way in their new shafts, while two more arrive in the other shafts. A modification of the carosel can allow removal/addition of cabs in the system, depending of need-of-repair or passenger requirements.
The general principles of this Idea are portrayed. [Vernon, Oct 04 2004]
GOOD bearings are required
I originally thought of these while trying to fit that bill. I still think they can be made to work, although a design-tweak may be needed. [Vernon, Oct 04 2004]
(?) unpredictable irreversability
[Fussass, Oct 04 2004, last modified Oct 21 2004]
No helices required
...and some of the much more interesting things you can do once the cab is self-motivated. [DrCurry, Oct 04 2004, last modified Oct 21 2004]
Surf around at the site for a cross section, they go through a query, so there is no link. [kbecker, Oct 04 2004, last modified Oct 21 2004]
This one does not use a pantograph power connection. [Vernon, Oct 04 2004, last modified Oct 21 2004]
A variety of trolleys
Some have "half pantograph" power connectors; some have full pantograph connectors. Pantographs are, I think, more common in Europe than in USA. [Vernon, Oct 04 2004, last modified Mar 25 2015]
I once read that this type of gearing can easily have a 1000:1 reduction ratio. If mounted with a high-speed electric motor on the elevator cab, and if the driven and slowly-moving gear has external teeth that mesh with a rack in the elevator shaft, it may be that the weight of the elevator may be unable to cause the interior planetary gears, attached to the drive motor, to rotate. [Vernon, Mar 25 2015]
elevator with its own propulsion [notexactly, Mar 25 2015]
The point of this idea seems to be to implement a paternoster in a new way. [notexactly, Mar 25 2015]
Variation of gear differential
This variant has the planetary gears on the outside. The main trick is that the two gears meshed by the planetary gears have different numbers of teeth (often just a difference of one tooth). That's why, if one of those gears is held still, the other one rotates slowly. [Vernon, Mar 25 2015]
Another way of implementing a climbing elevator. I think it would be easier than the screw mechanism. [notexactly, Mar 25 2015]
Alternative to planetary gears for gearing down by large ratios. Also search for 'harmonic drive' and 'strain wave gear' on YouTube. [notexactly, Mar 25 2015]
This is what you really want, with a few added features. And yes, it can do vertical. [MechE, Mar 26 2015]
what does forego mean?
[pashute, Nov 29 2019]
||An annotation nearly worthy of a Vernon idea..
||1) Instead of 'screw' or 'helix' you might say 'worm gear'. You're talking about a rack-and-worm system.
||2) The key feature of a worm gear is that the thread is usually fine enough (the 'spacing' is small) and there's enough friction that while the worm-gear-based device can move easily enough under its own power, it's virtually impossible for it to slip or be moved by an external force. However, you're talking about a relatively coarse thread spacing, and putting roller (or similar) bearings in there. Thus, it would appear that the elevator car would readily fall if the motor/brake/clutch/etc assembly broke away from the worm gear assembly. While it wouldn't be a fast free-fall, it would still be a dangerous possibility (eg, when someone is getting on or off, and gets their limb caught through the door-way as the car is descending).
||3) The upshot of your flavour of worm gear is that it doesn't appear to be significantly different to a rack-and-pinion system - ie, a rack on the wall to mesh with and plain toothed rotary gears on the corners of the car. The worm gear has an advantage of multiple holding points along the height of the car - an equivalent pinion system would need many (powered) gears to provide such redundancy. However, there is the disadvantage of the more distributed set of moving parts - your idea has bearings over the whole of the lift shaft that would need maintaining. A conventional (cabled) lift has only the cab and the motor unit to maintain. The same is true for a rack-and-pinion system (which would be required for the second part of your idea, as a conventional elevator car isn't a 'completely independent entity').
||4) This design suffers in efficiency w.r.t. cabled elevators. It must carry the heavy motor in the lift instead of it being kept stationary. There is no counterbalance; thus the elevator must be capable of lifting the car's weight (including motor) + passengers' weight. A conventional design only needs to be able to lift the passengers' weight. This adds up to using more power, and requiring a larger motor.
Regenerative braking, while good, might not be that helpful in this environment. During low-use periods, often only a single car may be moving; and during peak demand periods, much of the traffic will be going the same direction (all up or all down). I suspect that regenerative braking will mostly be useful for moving cars that are "in the way" of other cars in your loop design.
||5) I don't like the carousel system: It seems like a strong candidate for failure; KISS has not been applied. A simple carousel would end up rotating the cars by 90 degrees, requiring cars to have opening doors on all 4 sides. A better carousel would be yet more delicate (teacup fairground rides require some TLC to run smoothly). I have no idea how the power connection is dealt with while a car is in a carousel. Is it unplugged and then plugged back in to a 'trolley' connection on a wall? Or do these trolleys perform some cunning twirling of their own (how)? Either way, major points of possible failure.
||[benjamin] thank you; your comments are indeed mostly worthy.
||1) Semantics -- but not quite. Pretty much everyone knows what a screw is, and I suspect more people know what a helix is (does "double helix" ring a bell?) than a worm gear. Also, because of the normal use of a worm gear, it tends to have a specialized ("wasp-waisted") helical shape that allows it to make contact along the curvature of a more-ordinary gear. That shape is not perfectly suited for a "rack". Finally, "helix" is the name used by mathematicians for that shape, irrespective of application. That is, a helix is ALWAYS that shape. Worms are not always gears....
||2) I did not feel it necessary to specify all details of this design; this is the HalfBakery, after all. I do agree that the coarseness of this rack-and-helix system would allow it to start to spin under the weight of an unpowered elevator cab, and that is why I specified the need to clutch it. However, because this is an issue where a power loss MUST be associated with the phrase "fail safe", you can be quite sure that if this scheme is ever implemented, this particular issue will first be resolved to everyone's satisfaction.
||3) A failed rack-and-pinion system IS subject to free-fall, while as you yourself stated, a rack-and-helix system isn't. That is precisely why I chose this design (and not, say, a linear-motor system). Leverage is in our favor, when it comes to clutching the mechanism (I see the word "brake" in a following comment; same thing). Next, I do agree that the bearings pose a maintenance issue, and one thing I didn't say in the main text is that the bearing-mountings should let them be easily replace-able. Also, this is actually the reason why, some years ago, I spent some time thinking about bearing designs, and came up with the tri-layer idea that is also posted here at the HalfBakery (see second link here). If it worked as intended (and I think that a modification of that idea WILL), then such bearings will be so long-lasting as to make this particular maintenance issue a matter of "replace all bearings every X years".
||4) The efficiency problem is one that will not go away in any system designed to allow multiple cabs in a shaft. Please note that the helix-on-rack leverage effect (that helps us clutch the drive mechanism in an emergency) ALSO helps the motor with its job. The particular objection you have with respect to loads versus time-of-day is something that can be handled by some kind of on-site energy storage system. Such a system could also come in handy during a power failure, to ensure every cab in the system can line-up with the first available floor, to let passengers out.
||5) You have jumped to different conclusion than I, with respect to the carosel drawing. Yes, EVERYTHING in that circle rotates around the central axis, including doors. This just means that on every floor, the quad-shaft is surrounded by a walkway, and there is no problem with the doors (just because I didn't label all 8 doors in the sketch doesn't mean they aren't there). Regarding power, this is simple. The carosels have separate power from the main elevator shafts, such that the carosels can feed power to those sections of the four shafts that are part of the carosels. Note that I did try to indicate that the carosels are NOT in constant rotation. Starting with its shafts aligned with the building's shafts, cabs can enter/leave the carosel. Then the entire assembly rotates 90 degrees, and stops, for the next exchange of cabs. In every orientation, the doors are at the outer corners of the quad-shaft. In every stopped orientation, the power-providing strips are aligned with those in the building's shafts. And in every stopped orientation, the torsion-preventing roller coaster trackway also aligns.
||A worm gear set is said to be self-locking, or irreversible when the gear cannot drive the worm. This condition is obtained, if the lead angle of the worm is less than the friction angle, and as a consequence the efficiency for reversed driving is zero. The friction angle for static conditions will vary with such factors as surface finish and lubrication. Based upon the generally accepted value of static coefficient of friction equal to 0.15, the friction angle would be approximately 8 degrees;. However, the friction angle decreases rapidly with the start of motion, also, vibrations from nearby sources quite often upset the static condition of a locked set of gearing a sufficient amount to reduce the friction angle to a point where motion occurs, this is sometime called stiction. These unpredictable factors make it advisable to resort to a brake rather than relying on the self-locking characteristics of the worm gearing.
||We (well, I) have posted links before to the various schemes underway to do this very thing (cable-less elevator cabs). The main advantage of such systems is that cabs can now travel sideways as well as up.
||[DrCurry], people have been wanting to create "Turbo Lifts" ever since the first Star Trek, and possibly even before then. Mechanical details can be so important, sometimes! Personally, I tend to think that sideways shafts would waste space in a large building. Is everyone really THAT lazy, that if a building instead had "moving walkway conveyors", that they wouldn't want to walk between elevator and conveyor? Still, thanks for the link!
||[Vernon] what you are looking for are ball screws (link, not a medieval torture device). Also you should use four stepper motors, one for each shaft. Stepper motors are easy to move syncronously and will hold the shaft when power fails.
||[kbecker], I think I have to disagree about the ball screws, mostly because the helical shaft on which the device rotates can only have some maximum length, before it must be attached to the walls of the elevator shaft. How does the ball screw get past those mounting points?
||Regarding stepper motors, though, that could be good, although I don't know about stepping them at 3000 RPM....
||[benjamin], as an alternative to a full carosel, consider a kind of oscillator. Either way (designs are essentially identical) would allow the cabs to change shafts, but an oscillating "carosel" could use comparatively solid power connections, for the short sections of shafts that the carosel contains, for feeding the cabs that enter those sections.
||A back+forth oscillator design might be better; though I have never been particularly worried about the powering of the cars while within the carousel - rather, my concern was with the switchover from being powered from one shaft (or perhaps quarter-shaft) to the next.
||Ahh! By 'trolley' connections, you're talking about pantograph-type contact that brushes against a (vertical) rail or wire, as in trolleybus. I was thinking of little 'trolleys', as in things with wheels on rails, that rolled up the side of the shaft and that the car plugged in to. My mistake
*hides under a rock*
||[benjamin], yes, that kind of trolley (or tram; I've now edited the main text), although some use an alternative to a pantograph connection. There is a kind of wheel-like trolley-power connection that I've seen, pressed up against the cable by a single springy arm. Placing two contact points at the end of that arm would ensure that at least one was getting power, when switching from shaft to carosel.
||[notexactly], see item 3 of my May 23, 2004 annotation. It
is critical to maximally reduce the possibility of a free-
falling elevator cab. I don't trust the cog-train system
because it isn't used for vertical climbs; it is only used
where steel-wheel-on-steel-rail friction is inadequate for
the slope of the tracks.
||A worm drive can be back-driven if the pitch of the helix
is long enough. You said you wanted it to regenerate
electricity while descending, which means your worm
drive must be back-drivable. Therefore, the worm drive
itself will not function as an emergency brake; it needs a
brake applied to it. That's no different from a rack and
pinion. A rack-and-pinion elevator could also regenerate,
probably more efficiently because there wouldn't be so
many bearings. If you give up regeneration, though, you
can have a non-back-drivable worm drive that can
function as an emergency brake (an always-on-except-
when-driving brake, really).
||I could be mistaken, but the regeneration notion is not
necessarily incompatible with a low-pitch worm drive.
The key is getting the worm started turning. We are
relying mostly on friction and mechanical DISadvantage
to keep the elevator in place after stopping at any floor
(although a safety latch could be engaged to ensure it).
||With respect to a worm driving a gear, once the worm is
started to turn, its momentum will tend to keep it
applying force to turn the gear should encourage the
worm to keep turning. For the elevator and modified
rack, the helix is started to turn by the electric motor
disengaging any safety latch, of
course), and the elevator's weight would be the source
to encourage the worm to keep turning, thereby making
the motor act like a generator, while the elevator
down the shaft.
||Now I do know that the process of USING generated
electricity (such as storing it in a battery pack) causes a
counterforce to exist in the generator, opposing the
force that is trying to generate electricity. This
obviously would make it more difficult for the weight of
the elevator cab to keep encouraging the worm to turn.
Nevertheless, I'm hopeful that a happy medium is
possible, regarding helix pitch, safety, and
||If the helix system is replaced by a high-ratio planetary
differential system, and a more-ordinary rack on the
elevator shaft-wall, then again safety would result from
the difficulty of the cab's weight to force an initial
rotation (this time of the planetary gears).
||Topic change: paternoster. There are major
differences between this Idea and a paternoster. The
shafts are not 100% filled with elevator cabs (same
number as number of floors in building). Any cab can
move independently of any other cab, and can be
stationary at any floor, for passenger entry/exit.
Quoting from the main text: "Spacing between the cabs
should be such that ON THE AVERAGE, a cab that stops
for passengers will start moving again before the next
cab reaches that floor. Computer modelling can assist in
determining that." Obviously there should be safety
features ensuring that no cab collides with another.
||It might be convenient for a stair-well to be located
very near the elevator system, so that passengers can
use the stairs if a cab is forced to stop just short of its
destination-floor, due to another cab still at that floor.
||That makes sense; I concede those points.
||I think that [Vernon] has his mechanism inside-out. Rather
than the moving (male) screws being on the cab and the
partial (female) "nuts" being fixed to the walls, it would be
far better to have the screws running top-to-bottom, fixed
in the elevator shaft, and powered nuts (probably 8; top &
bottom each corner) on the cab. The screws would be under
tension from the top; have the bottom "floating" (splined
shaft into hub with bearings) so they can't compress/buckle
||I thought of something like that when this was bumped the other day, but I
quickly realized a disadvantage of doing that: You're limited in how many
elevators can be in the same shaft by the tensile strength of the screw,
because all of their combined weight must be transferred to its top mounting
point. With the rack of bearings mounted on the wall, the rack can be
attached to the building structure along its whole length. Each tooth of the
rack can even be independently mounted to the structure, and not transfer
forces to adjacent teeth.
||Another disadvantage of the hanging screw is that it makes the top carousel
really difficult to implement: how do you separate the screws in the shafts
from those in the carousel so the carousel can rotate, while still supporting
the screws in the shafts from their tops, while still allowing the elevators to
climb from the shaft section of the screws to the carousel section or vice
||Because the carousel part is WAY overkill. Better to have a
"dead space" above the top floor and below the bottom floor
for an "extra" cab to hide in if necessary.