Elevator rail clip and fishplate bracket design guide

A rail clip is a 200 g bracket carrying a guide rail that may weigh 20 kg per metre, and it does the job with friction rather than with a key. When it works, nobody sees it. When it does not, the car leans, the guide shoes wear on one side and the lift starts to hum at the top floors. The same is true of the fishplate bracket at a rail joint: the joint is set once by a mechanic on a ladder, and the bracket has to hold that setting for the next twenty years.

Both parts are simple to draw and easy to get wrong, and the recurring mistakes in elevator rail clip and fishplate bracket design are consistent: a slot too short to shim, a clamp arm too narrow for the rail foot, a bend radius under one material thickness, and a powder coat over the face that is supposed to grip. ISO 7465 fixes the rail section — the T-profile is standardised, the number in the designation T75, T89, T114, T127 tracks the rail width — but it does not tell anyone how much clamp force or how much clearance the fixing needs.

This guide covers elevator rail clip and fishplate bracket design at the drawing stage: what the rail foot dictates, how the slot and shim stack are sized, what preload a bolt actually delivers, how the shaft environment changes the finish, and the inspection that catches a bracket before it reaches a shaft. It is written from the programmes we run at Ganesh for TKE and Eros Elevators, where clips, brackets, sills and frames are made in the same plant as the car.

What the rail foot dictates before anything else

Every clip dimension starts at the rail foot, not at the wall. Take the foot width and foot thickness from the rail drawing or the mill certificate for the shipment, because a 5 mm bracket designed around a nominal foot that arrives 0.4 mm thicker will not seat — it will sit on the edge and rock.

Two rules come out of that:

The clamp arm should cover at least 40% of the foot width. A clip arm shorter than that puts the clamp load on the edge of the foot, and edge contact is a line load. Under guide shoe forces the foot rotates slightly in the clamp, the rail loses plumb and the mechanic blames the wall.

Leave 3 mm of clearance to the rail head. The head is the running face. A clip that touches it becomes a second guide shoe with no lubrication.

The other dimension that gets drawn wrong is bolt access. A clip whose bolt head sits 15 mm from the rail web cannot be tightened with a socket, and the installer will use an open-end wrench at half the specified torque. Draw 25 mm of clear space around the bolt head, and check it against the rail section rather than against the bracket alone.

The limits worth putting in the drawing's tolerance block

The table below is the set we check a rail clip and fishplate bracket against before a drawing goes to quotation. Values are for 3–8 mm structural plate, which is the band these parts live in.

Two rows carry most of the argument.

Hole to bend. For 5 mm plate on an R5 bend, 2t + R is 15 mm. Draw the slot at 10 mm and the bend pulls the slot into an hourglass, the bolt shank bears on the corner and the clamp loses about a third of its contact area. On a fishplate bracket in 8 mm the same rule gives 20 mm, which is often wider than the 15 mm web land the designer had in mind — that is a signal to move the bend, not to reduce the radius.

Rail expansion. Steel expands at about 12 × 10⁻⁶ per kelvin. A 100 m rail run in a shaft that swings from 15 °C to 35 °C moves 24 mm. If the clip clamps the foot vertically as well as laterally, that movement goes into the bracket and the wall. Design the clip to clamp sideways and allow the rail to slide: a vertical slot, a plain washer under the head, and a torque that produces clamp load without locking the joint solid.

How much clamp force a clip actually needs

Guide shoe loads are modest. On a passenger car the horizontal force at a guide shoe is in the low kilonewtons even under buffer impact, and most of the time it is a few hundred newtons of load imbalance. The clip does not need to be strong in the way a structural bracket is strong; it needs to stay put.

Work the friction number. An M12 grade 8.8 bolt tightened dry to 80 N·m develops roughly 40 kN of preload. With a coefficient of friction of 0.15 between a galvanized clip and a galvanized rail foot, one bolt gives about 6 kN of slip resistance. Two bolts per bracket give 12 kN. Against a shoe load in the low kilonewtons, that is a factor of four or more — but only if the joint is clean, dry and free of paint.

That last condition is where installations fail. Powder at 80 µm in a clamping joint does three things: it reduces the real contact area, it lowers the friction coefficient, and it creeps under sustained load. A coated joint at 80 N·m behaves like a bare joint at perhaps 50 N·m, and the rail creeps a millimetre a year until the mechanic is called back.

Three design consequences:

• Mask the clamping faces. A mask on the contact face costs a few seconds per part and protects the joint.
• Specify the washer. A plain washer spreads the clamp load; a wedge-lock washer holds it against vibration from the machine. Thread-locker is a poor substitute on a joint that has to be adjusted on site.
• Do not galvanize a joint face and then grind it. Grinding removes the zinc and leaves bare steel in the dampest part of the assembly.

For the wall side, the bracket is welded to an embedded plate or a wall anchor. On 5 mm plate, a 4 mm fillet at 100–120 mm pitch is enough; weld it with a back-step sequence from the middle outwards, because a bracket welded from one end pulls its free end 1–2 mm out of position and the slot then has to absorb an error it was never sized for. The same distortion logic applies to welded frames, and it is covered in more depth in controlling weld distortion in frames.

Choosing the clip type

Three families cover almost every shaft.

Bent plate clip

A single 4–5 mm blank, laser cut, bent into a Z or a U, bolted through a slot. Cheapest per piece, no tooling, and adjustable in two axes. It is the right answer up to about 2.5 m/s and for machine-room and low-rise installations. Its weakness is torsional stiffness: an arm that is longer than it is wide will twist under load, so keep the arm as short as the foot allows and add a gusset rather than a longer lever.

Bolted clamp with a separate pressure plate

Two parts, a body and a plate that bears on the foot. The pressure plate can be hardened and ground, and it can be replaced when the contact face wears. Use it where the rail is heavy — T114 and T127 sections on goods lifts — or where the clip is adjusted often during commissioning.

Rubber-padded clip

A clip with a bonded elastomer pad between the clamp and the foot. It cuts structure-borne noise from the rail into the wall and it accommodates a foot that is not perfectly parallel. The pad has to be specified by hardness and compression set, not by colour; a pad that takes a permanent set at 30% compression stops isolating after a few months.

The decision rule: bent plate below about 5 kN of slip demand, bolted clamp above it, padded where the shaft is attached to occupied space. Above that, high-rise and high-speed installations usually move to a forged or cast clamp, which we do not make — see the limits section below.

Coating a bracket that lives in a damp shaft

Lift shafts are humid, poorly ventilated and often the coldest part of a building. Condensation on a pit-level bracket is normal. The finish has to survive that, plus handling, plus a bolt being tightened on top of it.

We run two routes. The usual one is shot blasting in the 16 ft × 16 ft chamber, seven-tank pretreatment, then 60–80 µm of electrostatic powder in the 18 ft × 14 ft booth, with a DFT reading per batch and a cross-hatch adhesion check. That is a good finish for brackets inside a dry shaft and for anything visible.

The other route, hot-dip galvanizing to IS 4759 or ISO 1461 at 45–55 µm on 5 mm plate, is not an in-house process for us. Where a drawing calls for it, we supply the part shot blasted and free of weld spatter so it can be galvanized without a second pickling, and we ask for the coating thickness on the certificate rather than on the drawing alone.

Three finish notes that change the part:

• Every cut edge must be deburred. A laser edge holds a thin film at the burr root, and that is where rust starts. On a galvanized part the zinc has to cover the edge too; a sheared edge with a 0.3 mm roll-over takes a thinner zinc layer than a clean one.
• Weld spatter is a coating defect. It holds powder off the surface and it survives shot blasting. Clean it before blasting.
• Threaded parts are masked. A powder-coated weld nut is a nut that has to be chased with a tap on site.

The powder side of this is covered in why powder coating fails on outdoor enclosures, and the pretreatment detail in seven-tank pretreatment.

What Ganesh runs for rail clips and fishplate brackets

Clips and brackets start on the 12 kW fibre laser, which leaves an edge on 3–8 mm plate that needs no drilling or dressing. Forming runs on the 5-axis 200 T brake; its 4,450 mm bed is more than a 300 mm clip needs, but it keeps one radius setting across a batch of clips and a long fishplate bracket at the same time.

Welded brackets go to the manual MIG stations unless the volume justifies a fixture in one of the three CLOOS cells. The tool room builds the checking fixtures, so a bracket is checked against a rail gauge rather than against a caliper: slot position, clamp arm height, bend angle and hole pattern in one setup, with the rail section represented in the fixture.

For lift programmes we issue first article inspection reports with the dimensional results, the material certificate and the coating record, and we can supply PPAP documentation where the OEM's supplier quality team asks for it. The welded assembly tolerances article covers how the welded bracket's tolerances are set and inspected, minimum bend radius by material gives the radius table for the plate grades used here, and the rest of the part families we make for shafts sit in the elevator and lift components hub.

What the lift drawing has to settle

• Rail designation — T75-3/B, T89/B, T114/B or T127-2/B — plus the foot width and thickness from the rail drawing.
• Clamp face and masked area, with a note that no coating is permitted in the friction joint.
• Slot schedule with length, width and position, and the shim range the slot must accept.
• Bend table with inner radii, and the deburr requirement on the bend line.
• Bolt specification: size, grade, finish, washer type and torque, plus the thread condition assumed for that torque.
• Weld symbols and sequence, with the datum face and the hole pattern that must be held after welding.
• Coating system with DFT, the standard for galvanizing, and the salt-spray or adhesion evidence required.
• Inspection datum — which face and which hole define the bracket in the fixture.

The RFQ checklist lists the remaining fields, and the elevator parts page shows the bracket and frame families we build. If the drawing is ready, send the revision with the rail designation through Request a quote; the clamp geometry is the first thing our engineering team checks, because it is the cheapest place to change it.

What we cannot make, and what that means for the design

We do not forge or cast rail clips. High-rise and high-speed installations sometimes specify a forged clamp or a cast bracket with a machined seat, and those come from a different supply chain. We can make the mounting bracket, not the clamp body.

We do not hot-dip galvanize in-house. We can supply parts ready for galvanizing and we can specify the coating, but the zinc bath is an external process and it adds a lead time that has to be planned rather than assumed.

We do not drill or machine the guide rail itself. Rail holes come from the rail manufacturer, and a fishplate bracket has to be designed to that pattern. If the rail arrives undrilled, the joint is a rail problem before it is a bracket problem.

And we do not align rails in a shaft. We make the bracket to hold an alignment that a lift mechanic sets; if the slot is right and the wall is 8 mm out, no bracket will fix it.