Callback Response After a Door Repair

Purpose

A callback is a measurement of your last visit, taken by the customer. Handled well it costs a return trip and buys a door that stays fixed. Handled badly it is a tech re-adjusting whatever he can reach, no record of what changed, and a third visit.

This SOP guarantees two things: a callback naming a safety symptom gets a safety response before a scheduling response, and the visit is diagnosed against the shop's own as-left values rather than against memory. The event it prevents is the classic second failure - a door that reverses at random gets its close force raised on the callback, which removes the entrapment protection the first visit verified, turning a nuisance into an injury mechanism.

Scope

Covers any customer contact reporting that a door or operator the shop worked on is not behaving, from the phone call through diagnosis, correction, re-verification and classification, on residential and light-commercial sectional doors.

Does not cover a first-time call on a door the shop has not touched, which is ordinary dispatch. Does not cover the repair procedures themselves - each has its own SOP and this one hands off. Does not cover an entrapment event where someone was hurt or a vehicle was struck, which is an incident report and a manager on site, not a callback ticket.

Roles and responsibilities

What counts as a callback, and the four outcomes

Treat any complaint about the same door within 30 days of the shop's last visit as a callback until the record shows otherwise. Thirty days is a shop default, not a warranty term - tune it to your workmanship terms, but pick a number so nobody argues it on the phone.

Every callback closes in one of four classifications, each with an evidence line:

• Workmanship. Something the shop set, fitted or torqued is not as it should be. Includes any callback where no as-left value exists to compare, because an unverifiable visit defaults to the shop.
• Adjacent failure. A part next to the repair failed on its own schedule - the second cable after the first was replaced alone, or the mating spring.
• Unrelated. A different fault on the same door, with its own cue.
• Usage or environment. A vehicle bumped a track, a bike was stored across the photo eye.

Classification is not about who pays; it is the only input the weekly review has. A shop closing 200 repair tickets a quarter and seeing 14 return inside 30 days is running at 7 percent of repair tickets, against a 5 percent starting point worth holding to, and the classification split is what tells you whether that is training, parts, or a habit of replacing one of a pair.

Procedure

1. Run the safety screen before scheduling anything. Ask five things in this order: has the door closed on anything or anyone, does it reverse or stop on its own, will it stay closed, is there a bang or a new noise, and can they see a gap in a spring or a slack cable. Acceptance: five answers written on the ticket before a time is offered. Stop rule: any yes means the door is not used until a tech attends - tell the customer plainly to leave it where it is, not to run the opener, and not to pull the release with the door up. Hazard: that instruction is the control, so give it on this call and repeat it in a text, because a customer who keeps cycling a reversing door is repeating the event you are preventing.

2. Pull the original ticket and read the as-left values before dispatch. Acceptance: the tech leaves with the previous visit's balance drift, limit description, force method, reversal result and photo-eye results in hand. Wrong looks like dispatching on "we were there last month." Stop rule: if the previous ticket carries no values, flag it now so the tech arrives expecting to build a baseline rather than compare to one. Hazard: none at this step, it is done at a desk.

3. Secure the door before diagnosing anything. On arrival, get the vehicle out and people clear, unplug the operator and pocket the plug under the cord-and-plug exemption at 29 CFR 1910.147(a)(2)(iii)(A). Acceptance: opening empty, operator unplugged, door DOWN before hands go near track, cables or springs. Wrong looks like running the door six times in front of the customer to "see what it does." Stop rule: a door with a visible spring gap, a slack or off-drum cable, or a section out of the track is not cycled at all, and diagnosis converts to the relevant repair SOP. Hazard: the reported fault is by definition unpredictable motion, so nobody stands in the opening and the customer watches from beside the walk-through door.

4. Reproduce the complaint once, under control, in the customer's words. Ask them to describe it, then reproduce it deliberately with everyone clear. Acceptance: the symptom written in observable terms - where in travel, which direction, every cycle or intermittent. Wrong looks like accepting "it's just acting up." Stop rule: a symptom you cannot reproduce is not diagnosed by adjustment; the door is left as found, the customer notes the time and conditions of the next occurrence, and the ticket stays open. Hazard: the reproduction puts the door in motion in a known-faulty state, so keep the opening clear and stop after one attempt rather than cycling until it repeats.

5. Compare as-found against as-left, value by value. Take the same readings the last visit recorded: balance drift, seal contact across the width, reversal behavior, photo-eye response. Acceptance: a side-by-side line for each value with both numbers present. Wrong looks like a comparison against "it was fine when I left." Stop rule: where the previous ticket recorded a word instead of a value, stop comparing and record the callback as unverifiable, which defaults it to workmanship; then build a fresh baseline before touching anything. Hazard: hand-cycling with the release pulled puts the door on your hands, so keep fingers on the section rail and out of the joints.

6. Classify the callback with the evidence line attached. Pick one of the four and write the observation that decides it. Acceptance: one classification, one evidence sentence naming a value or a physical cue. Wrong looks like "customer error" with nothing behind it. Stop rule: where the tech's classification contradicts the previous ticket's record, it goes to the manager rather than being settled on site. Hazard: none at this step, written down away from the opening.

7. Correct it, then verify the protection you disturbed still works. Hand off to the relevant repair SOP, and when energy goes back in, re-latch the trolley, plug in standing beside the head, and re-run the contact reversal plus three photo-eye tests per garagedoor-limit-and-force-verification-after-a-repair. Acceptance: contact, reversal, full return to open, and immediate reversal on all three beam breaks, recorded as observed results. Both tests are the entrapment protection itself: UL 325 lists a residential head with an inherent contact reversal plus a secondary means such as photo eyes, and 16 CFR Part 1211 requires both, so a pass on one and a fail on the other is a fail. Stop rule: force is never raised to make a symptom go away - a door needing more force to complete travel has a mechanical fault, and if it cannot pass both tests today it is red-tagged, released to manual and unplugged. Hazard: the reversal test drives the door onto an obstruction on purpose, so the opening is empty and no hand, foot or tool is ever the object.

8. Close it in writing and feed it to the weekly review. Acceptance: classification, evidence line, what was changed, the new as-left values, and the customer told what was found. Wrong looks like a ticket reading "adjusted, working now." Stop rule: a callback does not close while the door is on manual - that ticket stays open until the door is back in service. Hazard: confirm before leaving that the customer knows whether the door is usable, because an ambiguous handover is how a red-tagged door gets run overnight.

The record this produces

• Fields: original ticket number and date; the five safety-screen answers; complaint in the customer's words; symptom in observable terms; as-found and as-left values side by side for balance drift, seal contact, reversal and photo eyes; classification with its evidence line; parts and adjustments made; disposition and any red-tag.
• Where it lands: the door's equipment record, linked to the original ticket so the pair reads as one history, plus a photo of whatever the evidence line names.
• Who reads it later: the manager running the weekly review, who needs the classification split rather than the count, and the next tech on that door.

Worked pass: 9 ft by 7 ft steel single, 11 days after a drum and cable job, one step fails

Step 1: customer reports the door "stops before the floor and goes back up." Safety screen: nothing struck, it does reverse on its own, it will not stay closed, no bang, no spring gap seen, cables look tight. Two answers trip the screen, so the door is taken out of use on the phone and a same-day slot is given.

Step 2: original ticket pulled. It records the drums re-set and both cables replaced, and under the verification line it says "limits set, tested OK." No values.

Step 3: vehicle out, homeowner beside the walk-through door, operator unplugged and plug pocketed. Step 4: reproduced once. From full open the door ran down normally, contacted nothing, stopped roughly a hand's width off the slab and returned to full open. Same behavior every cycle.

Step 5 failed and took its stop rule. Nothing to compare against: "tested OK" is a word, not a value, so the callback was recorded as unverifiable, classified workmanship, and a baseline built from scratch. Fresh measurements with the door hand-lowered: bottom seal gap 1.75 in at the right jamb and 0.25 in at the left, so 1.75 minus 0.25 leaves 1.5 in of difference across the width. The shop's gate is that more than 0.5 in of difference across the width is a drum or cable-set problem rather than a limit problem, and 1.5 in is three times that gate. Balance was good at about 1 in of drift over 10 seconds at mid travel.

Step 6: classified workmanship, evidence line "1.5 in of gap difference across the width, left drum set one wrap short of its mate." A limit adjustment would have hidden it: dropping the down limit to close the high side would have driven the low side into the slab and taught the operator to read the floor as an obstruction.

Step 7: drums re-set to matched wraps under garagedoor-drum-and-shaft-replacement-standard, seal contact re-checked at 0.1 in of difference across the width, inside the 0.5 in gate. Trolley re-latched and hand-checked, power restored from beside the head. Contact reversal: contact, reversal, full return to open. Photo eyes: three closes, beam broken from outside the opening at three points, immediate reversal each time.

Step 8: closed with the classification, the evidence line, both gap figures, and the as-left values written as numbers this time. The weekly review picked it up as the second unverifiable ticket that month from the same tech, which is a record-keeping fix rather than a skills one.

References

• 29 CFR 1910.147(a)(2)(iii)(A), the cord-and-plug exemption relied on in step 3.
• 16 CFR Part 1211 and the UL 325 listing it incorporates, exercised in step 7 and binding manufacturers of operators made on or after January 1, 1993; the reversal timing and permitted test object for a given head sit in that unit's manual.
• See related: garagedoor-limit-and-force-verification-after-a-repair and garagedoor-drum-and-shaft-replacement-standard, both called from this procedure.