Roller and Hinge Replacement Service

Purpose

Rollers and hinges read as the easiest hardware on the door, and one position on that door is the most dangerous fastener a residential tech ever puts a wrench on. The bottom bracket carries the lift cable, so its fasteners are under the full counterbalance load whenever the door is down, and people have lost hands taking one off to change a roller. This SOP separates the two jobs that look identical from a ladder: the eight rollers that come out with the door clamped and the springs untouched, and the two that do not come out until the springs are unwound. It also holds the hinge numbering, because a hinge in the wrong joint puts a roller at the wrong depth in the track curve and the door binds every cycle afterwards.

Scope

Covers replacement of rollers, centre hinges and end hinges on a residential sectional door, including the assessment that decides which hardware is worth replacing.

Does not cover the spring work that a bottom-bracket roller requires, which is garagedoor-torsion-spring-replacement, or a section whose end stile is too damaged to hold a hinge, which routes to garagedoor-panel-replacement-on-a-sectional-door. A door that is off its track gets garagedoor-door-off-track-recovery first; rollers replaced on a derailed door are rollers you will replace twice.

Where each roller sits, and why the bottom one is different

On a four-section door, each side carries five rollers, so ten in total: one in the bottom bracket, one in each of the three end hinges at the section joints, and one in the top fixture. Eight of the ten are ordinary hardware. Two are not.

   top section
  --- top fixture ------- roller, no cable
   third section
  --- joint 3 ----------- end hinge no. 3, deepest stem offset
   second section
  --- joint 2 ----------- end hinge no. 2
   bottom section
  --- joint 1 ----------- end hinge no. 1, shallowest offset
  --- bottom bracket ---- roller AND the lift cable ends here

Hinges are numbered from the bottom up, and the number sets how far the roller stem is offset from the door face. That offset is what lets each section follow the curve where the vertical track turns into the horizontal. Put a number 1 hinge in the number 3 joint and that section rides too shallow through the curve, which shows up as a bind at the same height on every cycle and eventually as a roller stem walking out of the track.

Roles and responsibilities

Procedure

1. Sort the work into intermediate hardware and bottom-bracket hardware before you start. Walk both sides and mark which rollers are being changed. Acceptance is a written split: the eight intermediate and top-fixture rollers on one line, the two bottom-bracket rollers on another. The hard rule is that no fastener on a bottom bracket is loosened while the lift cables carry tension, which on a door in the closed position is all of it. A tech who is not qualified to unwind torsion springs does not open a bottom bracket at all, and the correct answer to a worn bottom roller is to book it with the spring work rather than to try it.

2. Assess every roller and hinge before quoting, and write what you found. Spin each roller by hand and feel for stem play in the hinge sleeve, look at the race for rust bleed or a flat spot, and check each hinge for a worn knuckle, a cracked leaf and elongated fastener holes in the end stile. Acceptance is a per-position note for all ten roller positions and every hinge. The failure cue is a blanket "rollers worn" on the ticket: it hides the one damaged hinge that is the actual cause of the noise, and the customer pays for ten rollers and still has the noise.

3. Select rollers off the ones you removed, not off a catalogue default. Match the stem length, the wheel diameter and the race style. Residential hardware commonly runs a 2 in wheel on a stem around 4 in, in 7, 10 or 13 ball races, with nylon wheels running quieter than steel and sealed races lasting longer in a coastal or salted-road garage. Acceptance is a stem length measured against the removed roller and a wheel that sits in the track channel without side slop. A stem even slightly short is the failure cue, because the wheel rides at the edge of the channel and climbs out of the curve under load.

4. Clamp the door down and open only one joint at a time. With the door DOWN, clamp locking pliers into both vertical tracks above the bottom roller so the door cannot be raised, and unplug the operator with the plug kept in your control per the cord-and-plug exemption at 29 CFR 1910.147(a)(2)(iii)(A). Acceptance is two clamps and one joint open at any moment. Two adjacent joints open at once is the failure cue and it is how a section drops: with both its hinges off, a section is held by nothing but friction and the roller in the track.

5. Replace intermediate rollers and hinges joint by joint, keeping the numbers. Work one side of one joint at a time. Take the hinge off, pull the old roller, fit the new one, and put back a hinge of the SAME number, since the number sets the stem offset. Acceptance is each joint closed and fastened before the next is opened, and every hinge number matching the joint it sits in. Keep fingers clear of the knuckle as the leaves come together, because a section settling half an inch onto a knuckle will take a fingertip, and wear cut-resistant gloves against the section skin edges.

6. Route the two bottom-bracket rollers through the counterbalance procedure. These come out only with the springs fully unwound, the turn count recorded, the shaft locked and the cable lifted off the bottom bracket. Acceptance is a shaft that turns freely by hand before a single bottom-bracket fastener moves. A wound spring is stored mechanical energy under 29 CFR 1910.147 and no clamp on the track and no unplugged operator controls it. Wrong looks like a tech reasoning that the door is down and resting on the slab so the cable must be slack: it is not, the spring is at maximum torque in that position, and the bracket departs under load the moment the last fastener clears.

7. Treat elongated fastener holes as a section problem, not a fastener problem. If a hinge leaf has been working against its holes and the holes in the end stile are oval, the answer is not a longer or fatter screw. Acceptance is holes that are round and tight on a new fastener of the original size. Elongated holes get a manufacturer-approved backing plate, a hinge relocated to sound steel where the pattern allows it, or a section replacement, and that decision goes to the service manager. Driving a bigger screw into a torn stile buys a few weeks and then tears a bigger hole, and the section that fails is the one carrying a roller.

8. Verify by manual cycle, then powered cycle, then balance. With clamps off and the trolley still disconnected, run a full manual cycle and listen. Acceptance is no bind at any point, no roller chatter through the curve, every wheel visibly turning rather than skidding, and every hinge number matching its joint on a final walk. A wheel that skids rather than rolls means it is binding in the channel and gets diagnosed now. Then reconnect, run three powered cycles, and finish with the balance test: the door holds within a couple of inches at mid-travel over ten seconds.

The record this produces

A hardware ticket that names positions, because "replaced rollers" tells the next tech nothing about what is still original.

• Fields: the per-position condition note from step 2 for all ten roller positions and every hinge, which positions were replaced and which were left, roller stem length and race type installed, hinge numbers installed by joint, whether the bottom-bracket pair was done or deferred, any elongated-hole finding and how it was routed, and the manual cycle and balance results.
• Where it lands: the door's equipment record, with the bottom-bracket line flagged as open if it was deferred so it appears on the next dispatch.
• Who reads it later: the tech doing the spring work, who needs to know two rollers are waiting on that visit, the service manager tracking a noise complaint that survived a roller change, and the warranty administrator when a stile repair is claimed.

Worked pass: noisy 16 by 7 door, four sections, one routed finding

Complaint was rumbling and a squeal through the curve. Ten roller positions, three joints per side.

Step 1: split written as eight intermediate and top-fixture rollers today, two bottom-bracket rollers deferred, because the lead is not spring-qualified and the springs were sound and not otherwise being touched.

Step 2: per-position notes came back with all eight intermediate and top-fixture rollers flat-spotted and rust-bled, both bottom-bracket rollers gritty but still turning, one right-side number 2 end hinge with a worn knuckle and visibly oval fastener holes in the end stile, and everything else sound. The squeal traced to that hinge rather than to the rollers.

Step 3: stem length measured off a removed roller at just over 4 in, wheels 2 in, replaced with sealed 13 ball nylon since the house sits two blocks from salt water.

Step 4: door down, both tracks clamped, operator unplugged with the plug pocketed, one joint open at a time throughout.

Step 5: eight rollers changed, hinge numbers checked back into their own joints on the closing walk.

Step 7 failed and took its stop rule. The oval holes at the right-side number 2 position would not hold a new fastener of the original size, and the tech's first move was to reach for a longer self-drilling screw. The step forbids it. The finding went to the service manager, who took the manufacturer-approved backing plate option rather than a section replacement, since the stile was torn only at that one hinge and the rest of the section was sound. Plate fitted, fastener of original size seated round and tight. Had the bigger screw gone in, the stile would have torn further and the failure would have been at a roller-carrying hinge with the door mid-travel.

Step 6 deferred as planned. Bottom-bracket line left open on the record so it dispatches with the spring visit.

Step 8: manual cycle clean, no chatter through the curve, every wheel turning rather than skidding, all three hinge numbers per side matching their joints. Three powered cycles clean. Balance held at mid-travel with about an inch of drift over ten seconds. Ticket closed with the squeal attributed to the number 2 hinge, not to the rollers, and the two open bottom-bracket positions flagged.

References

• 29 CFR 1910.147, control of hazardous energy, covering the wound torsion spring as stored mechanical energy in step 6 and carrying the cord-and-plug exemption at (a)(2)(iii)(A) used in step 4.
• The door manufacturer's published hardware data for the series, which governs hinge numbering, fastener size and whether a backing plate is an approved repair for a torn end stile.
• ANSI/DASMA 102, specifications for sectional doors, an industry consensus document reaching you through the manufacturer's product data and your contract rather than through building code.
• See related: garagedoor-torsion-spring-replacement for the bottom-bracket pair, garagedoor-panel-replacement-on-a-sectional-door for a stile beyond repair, and garagedoor-door-off-track-recovery.