Cable Off the Drum Correction Standard

Purpose

A lift cable does not leave its drum on its own. It leaves because something put slack in it, and a drum groove cannot hold a slack cable. Re-seating is the easy half of the job and the half that does not last: a shop that only re-wraps is back at that door inside a few weeks, and the second visit is worse because the loose cable has been chewed by the drum flange in the meantime. This SOP forces the slack source to be named with physical evidence before the cable goes back on, and fixes the acceptance as an equal-tension, level-bottom door rather than "it went up and down once."

Scope

Covers a residential torsion-spring sectional door where one or both lift cables have come out of the drum grooves, whether they unwrapped, crossed over themselves, or wedged between drum and bearing plate.

Does not cover a broken or unwound spring, the upstream fault owned by garagedoor-broken-spring-emergency-call-response; if the coil is broken, that SOP runs first and this one after. A door that has left its vertical track is garagedoor-door-off-track-recovery. Extension-spring systems with pulleys and a fixed cable anchor are excluded, because the tension path is different and there is no drum.

Roles and responsibilities

The five sources of slack, and what each leaves behind

Name one of these with evidence in step 3 or the repair is not finished.

Procedure

1. Secure the door and de-energize before touching a cable. Unplug the operator at the ceiling receptacle and keep the plug in your hand or pocket. The cord-and-plug exemption at 29 CFR 1910.147(a)(2)(iii)(A) only holds while the plug is under the exclusive control of the person doing the work. Then clamp the door where it sits: locking pliers into both vertical tracks below the bottom roller if it is up or part-open, above the bottom roller if it is down. Acceptance is two clamps and the plug in your control. A partly-open door on one good cable is a suspended load hanging on one side, so no part of you goes under the bottom rail until the clamps are set.

2. Photograph and record the as-found state before you disturb it. Note which side is off, how many wraps have come out, whether the cable unwrapped off the top of the drum or wedged behind the flange, and whether the door is cocked in the opening. Acceptance is one photo plus four written observations. The failure cue is a tech who re-wraps first and then cannot say which direction the cable came off, because that direction is half the evidence for the cause: a cable that unwrapped upward points to slack from the top of travel, and one wedged behind the flange points to slack at the bottom.

3. Name the slack source with evidence before the cable goes back on. Work the table above against the door in front of you and write the cause on the ticket. Acceptance is one named source with the physical evidence beside it. The failure cue is a cause chosen by elimination or by what is most common - if you cannot point at the mark, the mark is not there and you have the wrong cause. Stop rule: if no evidence supports any of the five, do not re-seat and hand back a warranty. Re-seat, then set the door to manual with the opener disconnected, and book a return to watch it cycle under power.

4. Take the cable tension to zero by raising and clamping the door, not by unwinding the spring. Torsion spring torque is highest with the door down and lowest with it up, because at full open the door weight rides on the horizontal track rollers rather than on the cables. Release the trolley, raise the door by hand to full open, and clamp both vertical tracks below the bottom rollers. Acceptance is both cables visibly slack at the drums with the door held by the clamps, not by you. Never loosen a drum set screw with the door down: that side's cable is carrying its share of the door weight and releasing it drops the door on the diagonal.

5. Lock the shaft before loosening any set screw. The spring still carries residual turns at full open, so a freed drum lets the shaft spin and the good cable pays out. Clamp locking pliers onto the torsion shaft so the handle bears against the header bracket or the wall in the direction the shaft wants to turn. Acceptance is the shaft immobile when you push it by hand. Wrong looks like pliers that slip a few degrees under load: reset them with more bite before going further, and keep your fingers off the drum face and out of the shaft plane, because a slipping shaft lock turns the drum into a rotating pinch point at full spring torque.

6. Re-seat the cable one wrap per groove and re-tension both drums equally. Loosen the affected drum's set screws, feed the cable end down to the bottom bracket, then lay it into the helical grooves working from the bottom bracket upward, one wrap per groove with no crossover. Rotate the drum against the cable until the slack is out, hold, and torque the set screws to the drum manufacturer's figure, turning the drum with a bar or the flat of the hand, never with fingers in the grooves or on the flange and never with a hand between drum and bearing plate, because a shaft lock that slips with the set screws loose closes that gap at full spring torque - a set screw snugged to the shaft and then advanced a stated fraction of a turn, which varies by maker, so read their sheet rather than guessing. Acceptance is no visible slack in either cable and both drums showing the same number of wraps. Two wraps sharing one groove is the failure cue and it will re-fail: unwind and re-lay.

7. Inspect the cable itself before you trust it. Residential lift cables are typically 3/32 or 1/8 in galvanized aircraft cable. Run a gloved hand along the full length and look at the bottom-bracket loop and the drum-end stop. Acceptance is no broken strands, no bird-caging, no flattened crush and a clean swage. Any one of those means replace both cables as a pair, not just the damaged one, because the survivor has the same run history. A cable that was pinched between drum and flange is damaged whether or not you can see the strands, so it goes in the bin.

8. Correct the cause you named in step 3. A slipped down-limit gets reset so the door stops with the bottom seal just compressed and the chain or belt not jumping. A bind gets the track straightened, the roller replaced or the track bolts re-plumbed. A galled shaft flat gets the drum moved to clean shaft or the shaft replaced. Acceptance is the specific defect corrected and re-checked. Skipping this is the failure mode this whole SOP exists to prevent, and it presents as a callback three weeks later with the same cable off the same drum.

9. Cycle under power and measure the bottom rail. Remove the clamps and shaft lock, reconnect the trolley, plug in, then verify entrapment protection before the door returns to service: obstruct the photo eye on a closing door and confirm it reverses at once, then lay a 2x4 flat in the path and confirm it reverses on contact and returns fully open, per the opener instructions and 16 CFR 1211. You moved the down limit and with it the force setting, so these are the only checks that show the protection still works; a door failing either stays out of service whatever the cable looks like. Then run three full cycles. Acceptance is the bottom rail level across the opening within about 1/4 in, both cables taut throughout travel, and no cable jumping the groove at either end. Out of level beyond that means unequal drum tension: clamp, lock, and re-tension the low side. Stand clear of the opening while it runs, and watch the drums from the side rather than from underneath.

The record this produces

A cable ticket that a later reader can use, which means it must contain a cause, not just a repair.

• Fields: side affected, as-found direction of the unwrap, named slack source and its evidence, cable size and whether cables were replaced as a pair, drum number, set-screw torque reference used, bottom-rail level measurement after three cycles, photo-eye and 2x4 contact-reversal results, and the correction made to the cause.
• Where it lands: the door's equipment record with the as-found photo attached, and the cause field copied onto the invoice line so the customer sees what failed.
• Who reads it later: the service manager reviewing repeat cable calls at the same address, the warranty reviewer deciding whether a second cable failure is the same defect, and the next tech, who needs to know the shaft has a galled flat before assuming a set screw was simply loose.

Worked pass: left cable off, single-car door, one wrong cause

Door stuck about two feet open and cocked, left side low. Opener unplugged and plug pocketed, clamps set below both bottom rollers.

Step 2: left cable unwrapped off the top of the drum, roughly three wraps out, right cable seated. Door cocked about an inch and a half low on the left.

Step 3 failed and took its stop rule. The tech's first call was an obstruction, because the customer said a recycling bin sits on that side. The step requires the mark, and the bottom seal carried no crush line and the slab had no bin outline. Wrong cause. Re-worked against the table, the evidence landed on the down limit: the seal was mashed flat along its whole length and the chain showed a bright rub where it had been jumping the sprocket at the end of down travel. The customer then confirmed the door had been thumping at the floor for about a month. Had the tech stopped at the bin, the cable would have gone back on and come off again within a few weeks.

Steps 4 and 5: door raised by hand to full open and clamped, cables slack, locking pliers on the shaft bearing against the header bracket, shaft immobile under hand pressure.

Step 6: three wraps re-laid one per groove, drum rotated to take up slack, set screws torqued per the drum maker's sheet.

Step 7: the cable had been running on the drum flange and two outer strands were broken about a foot above the bottom bracket loop. Both cables replaced as a pair.

Step 8: as found, the opener was driving the door about an inch and a half past floor contact, which is what unloaded both cables at the end of every close. Limit reset so the seal just compresses with no chain jump.

Step 9 needed one correction. First lower measured the bottom rail 3/8 in low on the left, outside the 1/4 in acceptance, so the door was clamped and locked again and the left drum given a small additional rotation. Second measurement, called out by the second tech, was 1/8 in. Three full cycles run with both cables taut throughout and no groove jump. Ticket closed with the cause field reading "down limit overtravel, seal crushed, chain jumping" rather than "cable re-seated."

References

• 29 CFR 1910.147, control of hazardous energy, including the cord-and-plug exemption at (a)(2)(iii)(A) and a wound torsion spring as stored mechanical energy an unplugged operator does not control.
• The drum manufacturer's published set-screw torque and shaft-flat requirement for the part number stamped on the drum. Where shop habit and that sheet disagree, the sheet governs.
• 16 CFR Part 1211, the CPSC entrapment-protection requirements for residential garage door operators, verified by the reversal tests in step 9.
• See related: garagedoor-broken-spring-emergency-call-response, garagedoor-door-off-track-recovery, garagedoor-roller-and-hinge-replacement-service.