How Elevators Actually Work

Every tall building in existence depends on elevators. Without them, structures taller than five or six storeys are impractical for most occupants. Elisha Otis did not invent the lifting platform, but when he demonstrated his safety elevator in New York in 1854, he introduced the one mechanism that made tall buildings viable: a device that automatically arrested a fall if the hoist rope broke. Before that, a suspended platform was a calculated risk that few would take above a modest height.

The fundamental architecture of a modern traction elevator has changed remarkably little since the early twentieth century: a car counterbalanced by a heavy weight, connected by ropes over a driven sheave, powered by an electric motor that handles only the net imbalance. What has changed completely is the motor drive technology, the rope materials, the control electronics, and the safety systems. A 1935 Otis installation and a 2024 KONE MonoSpace share a physical principle separated by ninety years of engineering refinement.

This article covers how traction elevators work from the rope and sheave through the motor drive, the safety subsystems, the door mechanisms, the dispatching algorithms, and the energy recovery systems. Hydraulic and pneumatic elevators appear as well. European safety standards (EN 81) are referenced throughout because they define what every lift sold or installed in the EU must do and why.

The Counterweight and the Traction Equation

A traction elevator does not hoist its car against gravity. It counterbalances it. A counterweight, connected to the car by the same set of ropes passing over the drive sheave, carries a mass equal to the empty car weight plus approximately 40 to 50 percent of the rated load capacity. At half load, the system is nearly balanced: the motor provides just enough torque to overcome friction and inertia. At full load, the motor accelerates the net weight advantage of the car side. At no load, the counterweight side is heavier, and the motor must hold back a descent.

For a lift rated at 1,000 kg with a 630 kg empty car and a counterweight of 1,130 kg (car weight plus 50 percent of rated load), a fully loaded car ascending requires the motor to work against roughly 500 kg net imbalance. That same car descending at full load regenerates energy into the drive system because the motor is being driven by gravity rather than fighting it. The counterweight arrangement cuts the motor rating required for a given load by roughly half compared to a direct hoist.

Traction, in the mechanical sense, refers to friction between the ropes and the grooved drive sheave. If the ropes slip, the lift cannot move. The limiting relationship is the capstan equation:

T₁ / T₂ = e^(μθ)

T₁ is the tension on the heavier side, T₂ is the tension on the lighter side, μ is the friction coefficient between rope and groove, and θ is the rope wrap angle in radians. For a typical arrangement with a half-wrap (θ = π) and a friction coefficient of 0.1 in an undercut groove, the maximum allowable tension ratio is e^(0.1π) ≈ 1.37. The car side cannot weigh more than 37 percent more than the counterweight side before the rope slides on the sheave.

Groove geometry determines μ. A plain semicircular groove gives high contact area but low unit pressure and therefore low friction. A V-groove raises unit pressure and friction significantly but accelerates rope wear. The undercut groove, a semicircular profile with a narrow slot machined at the base, is the engineering compromise used in almost every modern traction elevator: the rope cannot seat fully in the bottom, increasing unit contact pressure without the aggressive rope wear of a pure V-groove. The undercut subtended angle is typically between 90 and 105 degrees, and the chosen geometry appears in the lift's traction calculation documentation submitted for EN 81-20 compliance.

Roping Arrangements

Most passenger lifts use 1:1 roping: ropes connect directly from the car crosshead to the counterweight via the drive sheave. Car speed equals rope speed, and the motor torque directly opposes the load imbalance.

Some installations, particularly heavy-duty freight lifts and certain hydraulic-adjacent configurations, use 2:1 roping. The rope starts from a fixed dead-end hitch, wraps under a pulley on the car, rises to the drive sheave, descends to a pulley on the counterweight, and terminates at another dead-end hitch. This gives a 2:1 mechanical advantage: the car moves at half rope speed, and the motor needs to exert half the force. The tradeoff is that the motor runs at twice the speed to achieve the same car velocity, and the rope is twice as long for the same shaft height.

Wire Ropes and Flat Belts

Traditional traction lifts use circular-cross-section wire ropes made from multiple strands of high-carbon steel wire wound helically around a central core. A rope for a mid-rise passenger lift might be 13 mm in diameter, constructed as 8 strands of 19 wires each, wrapped around a fibre or steel core. EN 81-20 requires a minimum of three ropes per car, each independently rated to support the load.

Wire rope life depends heavily on sheave diameter. Bending a rope over a small sheave fatigues the steel wires at the bend point. EN 81-20 mandates a minimum sheave-to-rope diameter ratio of 40 for passenger lifts: a 13 mm rope requires a sheave at least 520 mm in diameter. In a typical installation, ropes are inspected regularly for broken wires, diameter reduction from internal corrosion, and surface defects. EN 81-20 defines rejection criteria: for example, in any rope section equal to one lay length, more than a prescribed number of broken wires on the outer strands requires immediate replacement. In a continuously operating commercial building, ropes typically last 10 to 20 years.

Flat belts replaced wire rope in mid-rise machine-room-less lifts starting in the late 1990s. KONE introduced the EcoDisc flat belt in 1996 with the first MonoSpace installation. Otis followed with the Gen2 belt in 2000. Schindler uses flat ropes in the 3300 and 5500 product families.

A flat elevator belt is not a rubber belt. The structural elements are thin high-tensile steel cords (12 to 24 cords, depending on rated load) embedded in a polyurethane matrix. The finished belt is roughly 30 mm wide and 3 mm thick. The polyurethane provides the traction surface against the sheave and protects the steel cords from moisture and abrasion.

The key advantage is sheave size. Because the belt is thin, the sheave radius at the bend point is much smaller: a 100 mm sheave works with a 3 mm flat belt without exceeding fatigue limits, compared to the 520 mm minimum for a 13 mm wire rope. A smaller sheave allows a smaller, slower-turning motor and enables the compact machine design that makes machine-room-less lifts practical. The flat contact geometry between belt and sheave also increases the effective friction coefficient, reducing the number of belts required per car.

KONE's UltraRope, released in 2013, substitutes carbon fibre cords for steel. Carbon fibre has roughly one-fifth the density of steel at similar tensile strength. In a 500 m shaft, conventional steel rope running in a 2:1 system weighs approximately 18,000 kg. UltraRope weighs approximately 3,000 kg over the same installation. The reduced rope weight changes the counterweight mass requirements and makes economically viable the double-deck and sky-lobby configurations found in megatall buildings. In European terms, the planned 300 m Copenhagen Tower and similar future projects in Frankfurt and Milan would benefit from reduced-mass rope systems to reach higher speeds without motor oversizing.

The Drive System: Ward-Leonard to VVVF

Early electric elevators controlled speed with rheostat banks in series with a DC motor armature. Switching resistors in or out of the circuit changed motor current and therefore torque. The control was coarse, and resistors dissipated significant energy as heat at every start and stop.

The Ward-Leonard system improved this by replacing the rheostats with a motor-generator set. A constant-speed AC induction motor drove a DC generator whose output voltage was controlled by varying the generator field excitation. Varying field current varied the generator voltage smoothly and continuously, giving precise DC motor speed control with high efficiency through the operating range. Ward-Leonard systems were standard in premium installations from approximately 1905 through the 1970s. Their drawbacks were significant capital cost, continuous no-load losses from the running motor-generator, and large space requirements.

Variable Voltage Variable Frequency (VVVF) drives displaced the Ward-Leonard system as silicon power electronics became reliable and affordable through the 1980s and 1990s. A VVVF drive rectifies the incoming three-phase AC supply (400 V at 50 Hz in European installations) to DC, stores energy in a DC link capacitor bank, then inverts it back to three-phase AC at whatever frequency and voltage the motor controller commands. The inversion stage uses IGBTs (Insulated Gate Bipolar Transistors) switching at 8 to 16 kHz, with sinusoidal pulse-width modulation to synthesise a near-sinusoidal output waveform.

An induction motor's synchronous speed is proportional to supply frequency:

n_s = 60f / p

Where f is frequency in Hz and p is the number of pole pairs. At 50 Hz with a 2-pole-pair motor, synchronous speed is 1,500 rpm. Driving the motor at 10 Hz reduces synchronous speed to 300 rpm. The VVVF drive scales voltage proportionally with frequency (maintaining a roughly constant V/Hz ratio) to keep the motor's air-gap flux density at the design value. Too low a flux causes under-torque; too high saturates the iron core.

Modern elevator installations increasingly use permanent-magnet synchronous motors (PMSMs) rather than induction motors. PMSMs eliminate rotor copper losses (the rotor carries permanent magnets rather than windings), improving part-load efficiency. They also allow direct drive at low shaft speeds without a gearbox: the motor can turn at 50 to 150 rpm directly connected to the sheave, eliminating the worm or helical gearbox that traditional high-speed motors required. Removing the gearbox eliminates a major source of audible noise, vibration, gear oil maintenance, and mechanical efficiency loss. The gearless PMSM with VVVF drive is now the standard in all new mid-rise and high-rise European installations.

Velocity Profile and Ride Comfort

Passengers are more sensitive to jerk (rate of change of acceleration) than to acceleration itself. A sudden change in acceleration causes a lurch; a gradual change goes nearly unnoticed. Elevator controllers define an S-curve velocity profile to bound jerk:

• Jerk increases from zero to a maximum (typically 1 to 2 m/s³) during initial acceleration onset.
• Acceleration builds at constant jerk to the target rate (typically 0.8 to 1.2 m/s²).
• Jerk is reversed to zero, establishing constant acceleration.
• Constant acceleration continues until close to target speed.
• Jerk is applied to reduce acceleration back to zero as the car reaches target speed.
• Constant speed travel.
• Mirror sequence for deceleration to stop.

The VVVF drive executes this profile via a closed-loop speed controller with a position encoder (typically a magnetic incremental encoder on the motor shaft or sheave) providing feedback. The controller compares the demanded velocity profile to actual motor speed and adjusts the output frequency accordingly at the drive's update rate (typically 1 to 4 kHz).

Rated speeds vary by building height and purpose. European residential lifts typically run at 0.63 to 1 m/s. Commercial mid-rise installs run at 1 to 2.5 m/s. High-rise office buildings such as Tour La Défense in Paris or the Commerzbank Tower in Frankfurt use express shuttles at 5 to 6 m/s. Purpose-built observation or sky-lobby transfers can reach 8 to 10 m/s. The 8 m/s Hitachi lifts in the Shanghai Tower (not European, noted for context) required pressurised cabins because the air compression audible effects at that speed are uncomfortable without pressure equalisation.

Safety Systems

The safety architecture of a modern lift is built around the principle that any credible single failure must leave the car in a controlled state. Multiple independent mechanisms handle different failure modes, and the electrical safety circuit ties them together so that any open contact prevents motor power from being applied.

The Speed Governor

The speed governor is a centrifugal mechanical device, typically mounted in the machine room or at the top of the shaft, connected to the car by a closed-loop governor rope that passes over a tension sheave in the pit. As the car moves, the governor rope drives the governor sheave, rotating it at a speed proportional to car velocity.

At approximately 115 to 125 percent of rated speed (the first trip threshold defined by EN 81-20), centrifugal flyweights inside the governor pivot outward against a spring and trip an electrical contact. This contact is wired into the safety circuit; opening it cuts drive power and releases the electromagnetic brake. If the car continues to accelerate despite the brake engaging (because the brake has failed or a rope has snapped), the flyweights reach a second threshold at roughly 140 percent of rated speed and mechanically grip the governor rope, preventing it from moving.

The governor rope is connected to the progressive safety gear actuating linkage on the car frame. When the rope is arrested while the car continues to descend, the relative motion between rope and car lifts the safety gear actuating lever, deploying the safety gear.

Progressive Safety Gear

The progressive safety gear consists of two wedge-shaped jaws, one on each side of the car frame, bearing against the lift guide rails. Under normal operation, springs hold the jaws slightly clear of the rails. When the safety actuating lever lifts, a cam drives the jaws outward into contact with the rail. As the car continues to descend, friction between jaw and rail develops a self-energising force: the car's downward motion drives the wedge further into contact, increasing clamping force.

The cam profile shapes the deceleration. EN 81-20 requires the retardation during safety gear engagement to remain between 0.2g and 1g for progressive devices. Too low and the car travels too far before stopping; too high and passengers sustain injury. The rail surface condition matters: wet or corroded rails change the friction coefficient and can shift the actual deceleration outside the designed range, which is why guide rail condition is a maintenance item.

Instantaneous safety gears, which engage abruptly, are only permitted on lifts with rated speeds below 0.63 m/s in EN 81-20. Above that speed, the stopping shock would be injurious.

Electromagnetic Brake

The drive sheave is equipped with a spring-applied, electrically-released electromagnetic brake. In the de-energised state, compressed springs force brake shoes or pads against a drum or disc on the motor shaft. When the motor control system energises the brake coil, the electromagnetic force overcomes the springs and retracts the pads, freeing the shaft to rotate. Power loss for any reason causes the springs to immediately re-engage.

EN 81-20 requires the brake to hold 125 percent of the rated load stationary on the sheave with the drive motor de-energised. Two independent mechanical paths (two electromagnets, two sets of springs, two pairs of brake pads acting on the same drum) are required so that a single brake component failure still provides sufficient braking. The brake must engage within a prescribed time after the motor control relay opens.

Door Interlocks

The door interlock system prevents car movement unless every landing door and the car gate are closed and mechanically locked. Each landing door carries an electromechanical interlock device mounted at the top of the door frame. The interlock contains two contacts connected in series with the main safety circuit: a door-closed contact and a door-locked contact.

The door-closed contact confirms the door panels are fully in the closed position. The door-locked contact confirms the interlock latch has engaged, mechanically preventing the door from being pulled open from the landing side. Only when both contacts are closed (and the equivalent contacts on every other landing door and the car gate are also closed) does the safety circuit complete, allowing the motor control relay to energise.

Landing doors can be opened from the landing side with a special triangular emergency release key, or from the car side with an internal release mechanism accessible through the car gate. EN 81-20 prohibits landing doors from being openable from the landing without the key when the car is not present, since an open shaft door with no car presents an unguarded fall hazard.

Buffers

Buffers at the pit bottom absorb the kinetic energy of a car or counterweight that reaches the pit while moving. They are passive mechanical devices that require no electrical power.

For rated speeds up to 1 m/s, polyurethane or steel spring buffers (energy-accumulation type) are acceptable. They deflect under load, absorbing kinetic energy elastically, and then rebound. The rebound can be uncomfortable for any occupants if the car hits the buffer, but at 1 m/s the energy involved is limited.

For rated speeds above 1 m/s, oil hydraulic buffers (energy-dissipation type) are mandatory. An oil buffer is a cylinder containing hydraulic oil with a calibrated orifice plate. As the buffer plunger is driven into the cylinder by the descending car, oil is forced through the orifice, converting kinetic energy to heat. The orifice profile is designed to maintain a roughly constant decelerating force throughout the stroke, rather than spiking at initial contact. EN 81-20 specifies minimum buffer stroke as a function of rated speed, derived from absorbing the kinetic energy of the car arriving at 115 percent of rated speed.

Machine-Room-Less Elevators

Until the mid-1990s, every traction elevator required a machine room above or adjacent to the shaft: a dedicated room housing the motor-generator set or VVVF drive, the drive sheave and gearbox, the speed governor, and the control cabinet. Machine rooms typically needed to be at least twice the shaft cross-section in floor area, with specific height requirements, ventilation, and structural provisions for the drive machinery loads.

KONE's MonoSpace, introduced commercially in 1996 in Finland, eliminated the separate machine room. The gearless PMSM drive unit is mounted directly in the shaft headroom, on a beam spanning the shaft walls at the top of the shaft. The flat-belt sheave is compact enough to fit within the permitted overhead clearance without a projecting machine room overhead. The VVVF controller and governor are mounted in a slim cabinet adjacent to the top landing door, accessible for maintenance without a separate room.

The MonoSpace concept spread rapidly because it reduced total building volume devoted to elevator infrastructure, simplified planning, and reduced installation cost. Otis launched the Gen2 system in 2000 using a similar machine-in-headroom arrangement with flat polyurethane-coated steel belts. Schindler introduced the 3300 using a side-mounted motor in the shaft wall. ThyssenKrupp (now TK Elevator) developed the Evolution MRL with the machine mounted in the upper portion of the shaft on the guide rail brackets.

Machine-room-less designs impose some constraints. Maintenance access is more confined than in a dedicated machine room. The machinery is exposed to shaft conditions (temperature, humidity, dust) rather than a controlled room environment. Some national building codes took time to accommodate MRL designs, since earlier codes explicitly required a machine room. EN 81-20 was specifically drafted to accommodate MRL installations without the older prescriptive machine-room requirements of EN 81-1.

Hydraulic Elevators

Hydraulic lifts use a fluid-powered cylinder to push the car rather than a rope-and-sheave system. A pump pressurises hydraulic oil into a cylinder, which extends a piston to lift the car. Releasing oil from the cylinder allows the car to descend under gravity.

The simplest configuration is the direct-acting hydraulic lift: a single-stage cylinder installed vertically in a bored hole below the pit, with the piston base resting at the bottom of the hole and the piston top attached to the car. This configuration limits travel to the practical cylinder stroke, typically 12 to 20 m. Drilling a hole below the existing building foundation is a significant civil engineering constraint.

Telescoping cylinders use two or three nested stages, each extending in sequence, to achieve greater travel from a shorter buried cylinder. The mechanical advantage changes as stages extend, complicating speed control. Indirect hydraulic configurations use a cylinder mounted to the side of the shaft with ropes and pulleys to transmit motion to the car, allowing 2:1 or greater mechanical advantage and keeping the cylinder within the shaft rather than below it.

The pump is typically a fixed-displacement gear or vane pump driven by an AC induction motor running at constant speed. Speed control is achieved by varying the flow through a control valve. This approach wastes energy as heat in the valve during slow speeds or holds (the pump is always delivering full flow, and the excess is bypassed or throttled). Modern hydraulic lifts use variable-speed pump drives to reduce this waste, though they remain less efficient than traction systems with regenerative drives.

Hydraulic lifts have a practical upper speed limit of about 1 m/s and a practical upper travel limit of about 18 to 20 m for direct-acting configurations. They are common in low-rise residential buildings, automotive workshops, and freight applications where the absence of overhead machinery (no machine room above the shaft) is valued and travel distances are modest. The cylinders require environmental precautions to prevent oil contamination of ground water, and EU regulations on hydraulic fluid disposal affect maintenance costs.

Control Systems and Dispatching

The mechanical and electrical systems described so far move the car safely. The control system decides where to move it and when.

Collective Control

The oldest multi-car dispatch strategy is collective control (also called selective collective). Each hall call button registers a request associated with a direction (up or down) and a floor. The controller assigns a car to answer calls in the car's current direction of travel before reversing. All registered calls in that direction are answered in sequence before the car reverses. This is efficient for moderate traffic because a car making a run in one direction serves multiple calls without unnecessary reversals.

Collective control is implemented in even modern simple installations as the baseline strategy. For a single-car installation in a residential building, it is often sufficient. For a bank of six or eight cars in a busy office tower in central Amsterdam or Helsinki, collective control by itself produces unacceptable waiting times during peak periods.

Traffic-Adaptive Zoning and Predictive Dispatch

Multi-car installations in commercial buildings use traffic analysis to optimise car assignments. The controller continuously monitors call registration patterns, tracks car positions and loads, and applies assignment algorithms to minimise average waiting time and handling capacity.

During up-peak (morning arrival at an office building), the traffic pattern is strongly asymmetric: nearly all calls originate at the ground floor or lobby. The controller can assign cars to lobby runs, coordinating departure intervals to avoid multiple cars leaving simultaneously with similar destinations. Some systems apply predictive zoning: dividing the building into floor ranges and statically or dynamically assigning cars to zones based on current demand. A car assigned to floors 20 through 35 ignores calls in the 1 to 19 range even if it passes through them.

Monitoring load uses a load-weighing device (a strain gauge or load cell in the car buffer support or under the car platform). If the car load exceeds a threshold (typically 80 percent of rated load), the controller can bypass further hall calls and send the car to registered car calls only, preventing overloading and unnecessary stops.

Destination Dispatch

Destination dispatch systems replace traditional up/down hall call buttons with keypads or touchscreens on each landing. Instead of pressing "up" and then pressing a floor button inside the car, the passenger enters their destination floor at the landing. The controller groups passengers going to similar floors, assigns them all to one car, and displays the car designation (A, B, C) on the landing panel.

The result is a dramatic reduction in stops per trip. A conventional system might serve ten passengers going to floors 2, 7, 12, 18, and 24 by making five stops in one car. Destination dispatch can assign passengers to two cars, each making two or three stops, and route them so the two groups depart within seconds of each other. Total trip time falls, and building handling capacity increases.

Schindler PORT, KONE Destination Guidance System, and Otis CompassPlus are the major European-market implementations. Destination dispatch is now standard in new high-rise commercial buildings in cities such as London, Frankfurt, Warsaw, and Stockholm. The tradeoff is that visitors unfamiliar with the system require a learning moment, and the landing panels need to be clearly marked. Building management systems can integrate destination dispatch with access control, pre-assigning a car when an employee badges in at the building entrance.

Regenerative Drives and Energy Recovery

Elevator energy consumption in a busy commercial building is significant. In buildings taller than approximately 15 storeys, lifts can represent 5 to 8 percent of total electrical energy consumption. The VVVF drive creates an opportunity to recover some of this energy.

When the elevator car is heavier than the counterweight side and descending, or lighter than the counterweight side and ascending, gravity drives the motor. The motor acts as a generator, producing electrical energy from mechanical input. In a conventional VVVF drive with a diode rectifier front end, this generated energy has nowhere to go: the DC link voltage rises, and the drive must dissipate it through a braking resistor as heat.

A regenerative drive replaces the diode rectifier with an active front end (an IGBT-based rectifier that can operate bidirectionally). When the motor generates power, the active front end inverts this DC power back to AC and feeds it into the building's three-phase supply. The energy is reused by other building loads (lighting, computers, HVAC) rather than wasted as heat.

KONE claims the EcoDisc drive with regenerative front end recovers approximately 60 to 70 percent of the energy that a conventional drive would dissipate. Otis's ReGen Drive and ThyssenKrupp's ACE (AC drive with Energy Recovery) make similar claims. In a high-traffic installation making hundreds of trips per day, the recovered energy is measurable on building electricity meters.

The active front end also provides a benefit beyond energy recovery: it draws nearly sinusoidal current from the supply with a power factor close to unity, compared to the pulsed current drawn by a diode rectifier that introduces harmonic distortion into the building's electrical supply. In European buildings required to comply with EN 61000-3-12 (harmonics limits for equipment above 16 A per phase), this characteristic simplifies grid-connection compliance.

European Standards: EN 81-20 and EN 81-50

The safety of lifts in the European Union is governed primarily by the Machinery Directive (2006/42/EC) and the Lifts Directive (2014/33/EU), with the harmonised standard EN 81 providing the detailed technical requirements. The current main parts are EN 81-20 (safety rules for the construction and installation of lifts carrying persons or goods, electric lifts) and EN 81-50 (design rules, calculations, and examinations for the safety components of electric lifts).

EN 81-20 replaced the older EN 81-1 in 2017. The transition introduced several significant changes: it accommodated machine-room-less designs without the prescriptive machine-room requirements of EN 81-1, updated buffer stroke requirements, introduced requirements for rescue operations from inside the car, and strengthened door interlock testing requirements.

Key EN 81-20 provisions include:

• Pit depth: minimum 500 mm clearance below the car at lowest travel position when at pit floor level, sufficient for a person to stand on a platform without being struck by a descending car.
• Headroom: minimum 1,000 mm clearance above the car roof to the shaft ceiling at highest travel position, sufficient for a person on the car roof during maintenance.
• Load testing: every new installation must be tested at 110 percent of rated load before commissioning.
• Periodic inspection: lifts must undergo periodic statutory inspection at intervals defined by national implementing regulations (in Greece, every two years for passenger lifts in residential buildings; annual for commercial and public buildings).
• Unintended car movement protection: since 2017, all new lifts must include a device to detect and arrest unintended car movement away from a landing while doors are open, even if the speed governor threshold is not reached.

The unintended car movement protection (UCMP) requirement was introduced in response to accidents where worn or improperly adjusted brakes allowed cars to creep while loading or unloading passengers, causing entrapment or shearing injuries. UCMP can be implemented as a dedicated brake on the governor rope, an additional electromechanical brake on the sheave shaft, or a monitoring system that applies the existing brake at a much lower threshold than the governor trip speed.

Pneumatic Elevators

A small but distinct category of lift operates on air pressure rather than hydraulic fluid or ropes. The car is a sealed cylinder that travels inside a fixed tube. Reducing the air pressure above the car allows atmospheric pressure below to push it upward; admitting air above allows gravity to bring it down. A turbine or rotary vane pump manages the pressure differential.

Pneumatic lifts (sold under brand names such as PVE from Spain) require no pit, no machine room, and no cables. They are installed by assembling prefabricated tube sections. The practical limits are severe: current designs support loads up to about 300 kg, travel distances up to roughly 14 m, and shaft diameters typically between 750 and 1,000 mm. The glass tube makes the mechanism visible, which is the main aesthetic selling point. Energy consumption is higher per trip than a balanced traction elevator because the system always works against atmospheric pressure rather than using a counterweight.

Pneumatic lifts appear in residential installations in southern Europe, including Greece, Portugal, and Spain, where low-rise buildings (two to four floors) need accessible transport for mobility-impaired residents without the civil engineering cost of a pit and machine room. They are not a substitute for traction or hydraulic lifts in any application requiring more than two or three passengers or more than three or four floors of travel.

Door Operator Mechanics

The door system is the component passengers interact with most directly, and it is one of the most maintenance-intensive parts of any installation. A centre-opening door (the most common type in European commercial lifts) uses two pairs of panels: two car door panels that drive directly from the door operator motor, and two landing door panels at each floor that are coupled to the car door panels by a mechanical interlock vane and rollers during the brief period when the car is at floor level.

The door operator is an electric motor mounted at the top of the car, driving a belt or lead-screw mechanism that moves the car door panels. The car door panels carry a vane, sometimes called a clutch or coupler, that projects into the space occupied by the landing door rollers when the car is within the door zone (typically 75 to 150 mm above or below floor level). When the car door opens, the vane mechanically engages the landing door rollers and pulls the landing door open simultaneously. When the car door closes, it pushes the landing door shut. The car and landing doors therefore move together without any direct electrical or mechanical connection other than the vane-and-roller coupling.

The door velocity profile is S-curved, analogous to the car velocity profile. The controller accelerates the panels quickly to a maximum opening speed (typically 0.3 to 0.5 m/s for a standard installation), maintains that speed through mid-travel, then decelerates to a slow crawl before the fully-open or fully-closed position. The deceleration before closure is particularly important: the door must be moving slowly enough at closure that any obstruction is detected and the door reverses before exerting significant force.

Safety reversals are handled by two independent systems. The safety edge (or safety shoe) is a mechanical contact strip on the leading edge of the closing door panel. If the edge contacts an obstruction, a microswitch opens and the door immediately reverses. Most modern installations also use a non-contact light curtain: a grid of infrared beams projected across the door opening. Any beam interruption triggers a reversal. The light curtain detects obstructions that a mechanical edge might miss, such as a walking stick positioned at mid-height between panels.

The nudging function addresses the case where the door is repeatedly obstructed (a passenger holding the door for others). After a configurable time (typically 20 to 30 seconds of unsuccessful closure attempts), the controller switches to nudging mode: the door begins closing at reduced speed, the safety edge is bypassed (the door will not reverse on contact, only slow), and an audible signal warns passengers that the door is closing with limited obstruction response. EN 81-20 permits this function specifically because holding a door open in a busy building creates a backlog of calls and delays other users.

Door operator controllers are increasingly encoder-based rather than purely time-based. An encoder on the door motor shaft gives the controller precise knowledge of panel position at every moment. This allows adaptive door timing: if the door encounters increased resistance (a slightly warped landing door sill, for instance), the controller detects the motor current increase, slows the door, and increases motor torque rather than simply failing to close. Position feedback also allows precise dwell at the fully-open position regardless of mechanical variation across different floors.

Guide Rails and the Ride Quality Chain

The car travels on two guide rails, I-section or T-section steel profiles running the full height of the shaft and bolted to the shaft structure at regular intervals. The guide shoes, mounted at each corner of the car frame (and at the top and bottom of the counterweight frame), slide or roll along the rail surfaces.

Older installations use sliding guide shoes: a cast-iron or nylon liner that bears directly against the rail flange. Sliding shoes require lubrication, either from oil-soaked felt pads fed by gravity from a reservoir, or from automatic lubrication units with metered oil dispensing. Without lubrication, rail-to-shoe friction generates noise and causes wear on both the shoe liner and the rail surface. Well-maintained sliding shoes on a properly aligned rail are quiet and smooth, but they do not isolate vibration at all: any rail imperfection transmits directly into the car frame.

Roller guide shoes replace the sliding liner with three polyurethane-coated steel rollers arranged to bear on the three surfaces of the rail flange (two on the sides, one on the face). Rollers do not require lubrication and provide some vibration isolation from the polyurethane tyre material. They are standard in all modern mid-rise and high-rise installations in Europe.

Active roller guide shoes, used in high-speed lifts in tall buildings, add a control system to the roller mounting. Accelerometers on the car frame detect lateral vibration and feed signals to actuators that adjust roller preload or position, actively cancelling the lateral motion before it reaches the car interior. The ThyssenKrupp TWIN system (which runs two cars in the same shaft independently) and high-speed installations in buildings above 200 m in height use active guides as standard.

Rail alignment is the dominant factor in ride quality. Rails must be plumb within tight tolerances: EN 81-20 specifies maximum permitted deviation from vertical, and installers use laser alignment tools to achieve the required straightness before final brackets are tightened. Rail joints (every rail section is typically 5 m long) require precise end-to-end alignment to prevent the guide shoe from encountering a step or gap at the joint. Even a 0.1 mm step in a rail joint is perceptible at 2.5 m/s as a brief vibration.

Thermal expansion is a practical concern in tall shafts. A 100 m steel rail column expands by approximately 1.1 mm per degree Celsius of temperature change. In a building with significant HVAC cycling, a 10-degree temperature variation across the shaft height produces roughly 11 mm of differential expansion between the top and bottom of the rail run, accumulated as spring-type compression at the rail brackets (which allow limited axial float). Rail guides and fish plates at rail joints are designed to accommodate this expansion without buckling or separation.

Modernisation Strategies

A significant fraction of European elevator installations are 25 to 50 years old, and the economic case for complete replacement (tearing out the cab, shaft equipment, rails, controller, and all wiring) is often unfavourable compared to modernisation: selectively replacing the components that limit safety, reliability, or efficiency.

The typical modernisation sequence for a 1980s-era traction elevator in a Greek or Italian apartment building addresses the control system first. The original relay logic controller (a panel of electromechanical relays, with a relay for each logic function) is replaced by a microprocessor controller. The new controller connects to the existing motor, brake, and safety circuit wiring, requiring only interface adaptation for the older signal levels and relay coil voltages. Benefits include faster response to calls, more sophisticated group control, and electronic fault logging that simplifies maintenance diagnosis.

Drive system modernisation replaces the existing motor and its control (which may be the original AC 2-speed motor, a rheostatic controller, or an early thyristor drive) with a VVVF drive and, if the sheave is compatible, either retains the existing induction motor or replaces it with a more efficient model. The VVVF drive eliminates the mechanical shock of switching between fast and slow speeds on an older 2-speed system, dramatically improving ride quality and reducing brake wear.

Replacing ropes is a mandatory maintenance item rather than a modernisation, but it often accompanies a drive modernisation because the increased starting smoothness of a VVVF drive reduces rope shock loads and can extend the service life of new ropes. Some modernisations take the opportunity to convert from wire ropes to flat belts, which requires a new sheave but may fit an existing MRL shaft layout.

Adding UCMP (Unintended Car Movement Protection) is now legally required for older lifts in many EU member states under national implementing regulations of the Lifts Directive. The retrofit options include a governor rope brake, a rail-mounted electromagnetic catch, or a drive-level monitoring relay that applies the brake at a very low speed threshold when the door zone detector indicates the car is at a landing with open doors. The choice depends on shaft geometry and existing equipment.

Energy-saving modernisations include adding a regenerative drive front end to an existing VVVF installation, replacing incandescent or fluorescent cab lighting with LEDs (reducing standby power from 80 to 100 W per cab to 10 to 20 W), and adding automatic standby modes that park a car at a low-traffic landing and switch off cab lighting and ventilation after a period of inactivity.

Lift Acoustics and Vibration

In residential buildings, elevator noise is a recurring complaint between neighbours and building managers. The sources are several, and each requires different mitigation.

Machine noise is the dominant source in older installations with geared drives. Worm gear mesh generates tonal noise at the gear mesh frequency (shaft speed times number of teeth). In a gearless PMSM drive, machine noise shifts to the motor itself: IGBT switching at 8 to 16 kHz creates a high-frequency acoustic emission from the motor windings and core, sometimes described as a whine. Some drives allow the user to raise the switching frequency above 16 kHz (above the nominal upper hearing limit) at the cost of increased switching losses.

Guide rail noise from poorly lubricated sliding shoes or damaged rollers transmits through the car frame and into the building structure at rail bracket attachment points. Structure-borne sound travels efficiently through concrete, and a noisy guide shoe can be audible several floors away from the shaft. Isolation pads (rubber or neoprene) at the rail bracket attachment points reduce but do not eliminate structural transmission.

Rope noise occurs at specific car speeds where the governor rope or suspension ropes develop standing-wave resonances in the shaft. The rope tension and the rope length between deflection points determine the resonant frequencies. At a car speed that excites a rope resonance, low-frequency humming or buzzing becomes audible in the shaft and adjacent spaces. VVVF drive controllers can be programmed with speed skip zones: ranges of car speed that the drive accelerates through quickly without dwelling, preventing sustained rope resonance excitation.

Door noise is particularly noticeable in residential buildings at night. A well-set-up door with properly adjusted panel stops, a clean sill groove, and polyurethane roller guide shoes closes with a soft thud. A maladjusted door may bang closed, or the car door vane may strike the landing door rollers with a metallic clang if the entry to the door zone is too abrupt. Reducing door closing speed or adding rubber bumpers to the final travel position addresses most door banging complaints without affecting function.

Brake noise is generated at each floor stop when the electromagnetic brake engages. In a VVVF system, the brake is applied after the car has come to rest electrically (the drive holds zero speed for a brief moment before the brake engages), which minimises the mechanical shock. If the brake is applied while the car is still moving slightly (a control timing fault), the metallic impact of the brake shoes against the drum or disc is clearly audible. Brake adjustment and timing calibration is a standard part of periodic maintenance precisely because brake noise is an early indicator of misadjustment.

What Passengers Feel Versus What the System Does

A well-maintained traction elevator in a modern commercial building is nearly imperceptible to ride. The S-curve velocity profile limits jerk to 1 to 1.5 m/s³, which most passengers cannot distinguish from standing in a gently accelerating vehicle. The electromagnetic brake engages with a barely audible thud at each stop. Door timing is calibrated to close slowly enough not to contact passengers and quickly enough not to idle the car unnecessarily.

The moments passengers notice are usually maintenance issues: a slightly over-adjusted brake that applies before the car reaches floor level exactly (leaving a small step between car and landing), a worn door operator that causes doors to hesitate or reverse unnecessarily, a governor rope that vibrates at certain speeds generating audible noise in the cabin, or a worn guide shoe that allows the car to rock slightly on the rails at low speed.

The feel of a fast lift at 6 to 8 m/s in a pressurised cabin, such as those in the Shard in London or the Four Frankfurt Tower, is qualitatively different: the pressure equalisation system operates audibly, the acceleration is clearly perceptible, and the deceleration into the upper floors requires the same S-curve control but over a much longer distance than a mid-rise installation.

Ride quality is formally measured by the ISO 18738-1 standard (Lifts, escalators and moving walks: measurement of ride quality), which defines instruments, measurement procedures, and reporting formats for acceleration, jerk, noise, and vibration inside the car. Measurements are taken with calibrated accelerometers at the floor level of the car interior, sampling at 200 Hz or higher, and the results are compared against performance grades defined in the standard. ISO 18738-1 Grade A (the highest) requires peak horizontal vibration below 15 milli-g (0.15 m/s²), peak vertical jerk below 2 m/s³, and A-weighted noise below 55 dB(A) inside the car. Independent inspection bodies use these measurements at acceptance testing and periodic audits to confirm that ride quality has not deteriorated, which may indicate worn guide shoes, rail misalignment, or belt/rope wear.

Remote monitoring has become standard across the major European lift manufacturers. KONE's 24/7 Connected Services, Otis ONE, and ThyssenKrupp's MAX (now TK Elevator MAX) platforms connect the controller via 4G/LTE to a cloud back-end. The controller streams operational data: number of trips per day, door cycle counts, brake application counts, motor temperature, controller fault logs, and load-weighing readings. Predictive maintenance algorithms on the cloud platform identify anomalies: a door that is taking 0.3 seconds longer to close than its baseline, a motor current draw that is 8 percent higher than the installation baseline at the same load, or a brake engagement time that has drifted outside specification. Field engineers are alerted before a fault becomes a breakdown.

This matters in practice. A door that begins failing intermittently is often the precursor to a nuisance call-out: the car is stranded between floors with passengers because the door circuit fault causes the safety relay to open. A predictive alert 3 weeks before that failure, based on door cycle time trend data, allows a planned maintenance visit during off-hours rather than an emergency call-out at 18:30 on a Thursday. In a high-occupancy residential building in Athens or a commercial tower in Warsaw, that difference in response time matters to hundreds of people.

The engineering underneath an elevator ride is substantially more complex than the experience of pushing a button and watching numbers count up. The traction equation, the governor and safety gear, the interlock circuit spanning every door in the shaft, the VVVF drive executing a jerk-limited velocity profile updated at kilohertz rates, the dispatching algorithm coordinating multiple cars across hundreds of trips per day, and the remote monitoring platform watching operational trends all operate invisibly. When every component functions correctly, the ride is indistinguishable from standing still.