CHECK FIRSTThe photo eyes
Blocked, bumped or sun-blinded — the cause behind most reversal calls. Both LEDs steady is the pass condition; anything else, clean and re-aim before blaming the opener.
GUIDE · DIAGNOSIS FIRST
The Door Won't Close. Here Is the Ten-Minute Diagnosis
A door that starts to close and reopens is almost always the photo eyes — blocked, nudged, or sun-blinded. A door that will not move at all points at the lock switch, power, or the radio. A door that is heavy or crooked is spring or cable territory: stop and call. The ranked checklist below sorts yours in about ten minutes, free.
// THE NUMBERS UP FRONT
What do you check, in what order?
Ranked by hit-rate. Stop at the first one that fixes it.
Look at both photo-eye lights
Knee height, both sides of the opening. Each sensor carries an indicator LED: both steady means the circuit is happy; one dark or blinking means blocked, nudged or unplugged. A broom handle, a bin, a bike pedal or a cobweb across the lens is the most common culprit in the entire diagnosis.
Clean and re-aim the eyes
Wipe both lenses. Loosen the wing nut on the blinking side and pivot the sensor gently until its LED holds steady, then snug it down. They aim at each other across the opening — small angles matter. This single step resolves the biggest share of reversal complaints.
Check the lock switch
If NOTHING radio works but the wired wall button does, the console’s lock/vacation switch is engaged — bumped switches are silent and common. Toggle it and retest before touching anything else.
Confirm power and the trolley
Opener light on? Outlet live (test with a lamp — garage GFCIs trip quietly)? Red release cord hanging in its engaged position rather than pulled to bypass? A disengaged trolley lets the motor run while the door ignores it.
Watch one full cycle from inside
Send it down and watch where it quits: same mid-travel spot = look for a bent track or seized roller there; at the floor = limit calibration; instantly on the button press = eyes again. The failure location is the diagnosis.
Run the manual balance test
Door DOWN, pull the release, lift to half height by hand. Floats = counterbalance fine, the problem is electronic or calibration. Heavy, crooked, or fighting you = springs or cables, and the opener has been straining against it. That one is a call, not a fix.
The three usual suspects, on camera
CHECK FIRST
MID-TRAVELA binding track
A door that reverses at the same point every trip is meeting resistance there — a bent rail, a seized roller, a rub. The force limit is doing its job; the track needs the attention.
STOP & CALLThe counterbalance
Heavy on the manual lift, crooked in the opening, or a bang before it all started — spring and cable territory. Loaded hardware, licensed hands.
And if it won't OPEN instead?
The mirror problem shares most of the list, with three additions.
The bang you may have missed
A snapped torsion spring is the classic will-not-open cause: the opener strains, lifts inches, gives up. If the door is suddenly enormously heavy on the manual test, stop — that is the spring page, not this one.
Vacation lock, again
Same switch, same signature: every remote dead at once, wall button fine. It blocks opening commands exactly as it blocks closing ones.
The disengaged trolley
Someone pulled the red cord (often during a power cut) and the carriage never re-latched. Run the opener once and gently lift the door toward the trolley until it clicks back in — with the door DOWN.
Frozen to the slab
Hard-freeze mornings can bond the bottom seal to the concrete. Break the bond by hand from inside (door released, lift gently) — never let the opener rip it free, which tears seals and strips gears.
Force limit on the up-stroke
Openers carry an opening force ceiling too. A stiffened door trips it the same way — and the fix is the same balance test, then calibration.
Keypad-only failure
If handhelds work but the keypad does not, it is the keypad battery or its pairing — the programming guide covers the ninety-second fix.
The safety line on this whole page
SAFE — DOEverything observational: watch cycles, check lights, wipe lenses, test outlets, run the balance test with the door DOWN. None of it can hurt you.
SAFE — DORe-aim photo eyes and toggle switches freely — worst case is re-doing it. The eyes are the one safety component designed for owner hands.
PRO-ONLY — DO NOTNever disable, unplug or tape over photo eyes to force a close. That entrapment protection is federal baseline since 1993 for a reason that outweighs every inconvenience.
PRO-ONLY — DO NOTNever crank force or travel screws to bully a binding door closed — mis-set force is the mechanism behind car dents, cracked panels and worse. Find the bind instead.
PRO-ONLY — DO NOTNever touch springs, cables or bottom brackets during any of this. Loaded at rest, unforgiving in failure — the three stop-and-call symptoms exist for them.
Checklist exhausted — how do you make the call count?
Four sentences that turn your ten minutes of diagnosis into a cheaper, faster visit.
Report the checks, not just the symptom
“Won’t close; both eye lights steady; reverses at the same spot mid-travel; balance test floats” — that sentence has already done the pro’s first twenty minutes and priced the visit accordingly.
Name the hardware
Door type, opener brand off the motor head, roughly how old. Truck stock gets loaded from these three answers.
Say where it reverses
Instantly / mid-travel at one spot / at the floor — the three locations map to three different repairs. Your observation is the triage.
State the urgency honestly
Door parked down and car out = scheduled rates. Trapped car or unsecurable opening = say so and pay the premium knowingly, not accidentally.
The four parts this page keeps naming — a fast anatomy
Ninety seconds of vocabulary that makes every diagnosis above concrete.
Photo eyes
The two sensors facing each other at ankle height — an invisible tripwire the door must not cross while broken. Both carry indicator LEDs; both steady is the pass state.
The trolley & release
The shuttle connecting door to opener along the rail, with the red cord that disconnects it. Home of the “motor runs, door ignores it” family of faults.
Travel & force limits
Two stored calibrations: where to stop, and how hard to try before assuming an obstruction. Drift here produces floor-bounce, header gaps, and phantom reversals.
The counterbalance
Springs, cables, drums — the machinery that actually lifts the weight. When THIS side fails, every opener symptom is secondary, and the page you need is the spring one.
What stops this from happening again?
The five-habit prevention layer — every one free or nearly.
Monthly: the balance test
Door down, release pulled, one-hand lift to half height. Floats = counterbalance healthy. This single habit catches the spring fade that eventually trips force limits — about half a spring-life before it strands the door.
Quarterly: the reversal test
Board flat under the door, close it. Contact must produce immediate retreat — that is the entrapment system proving itself. No reversal means the opener stays unplugged until a pro calibrates it.
Quarterly: ten seconds of eye care
Both lenses wiped, both brackets checked square, both LEDs steady. The most common cause on this page, prevented by the cheapest habit on it.
Fall: lubricate before the freeze
Door-rated lithium or silicone on springs, hinges, rollers and the opener rail — never penetrating oil, never grease in the tracks. The pass that keeps winter from raising the door’s effort past the force ceiling.
Yearly: the measured visit
Forces verified against spec, limits re-learned, cables read at the brackets. The calibrations behind half this page drift silently; once a year, someone with a gauge should read them.
What will the pro actually do that you could not?
When the free layer is exhausted, this is what the visit adds.
Bench the electronics
Receiver memory, board faults, intermittent eye wiring — diagnosed with a meter rather than by part-swapping. On supported models this ends in an affordable component, not a new unit.
Calibrate with a test block
Travel and force set to spec and PROVEN with the reversal test — the difference between adjusted and safe. Mis-set force is the failure mode that damages what a door meets.
Find the bind
A reversal at one spot has a physical address: a rail out of plumb, a flat-spotted roller, a hinge cracking at the knuckle. Pros find it by feel on the released door in minutes.
Read the counterbalance
If the root cause is spring fade, the visit prices it honestly — and the wont-close symptom retires along with the tired steel. The spring cost guide on this site is the homework for that conversation.
The full symptom map — where does yours land?
Every close/open failure mode the trade actually sees, with its class of fix.
| Symptom | Most likely cause | Class |
|---|---|---|
| Closes, reverses instantly at the button | photo-eye circuit fault (blocked/misaligned/wired) | free check first |
| Closes only while holding the wall button | the designed eye-fault fallback — fix the eyes | free check first |
| Reverses at the same mid-travel spot | binding there: bent track, seized roller, rub | repair |
| Reverses right at the floor | close-limit overshoot or threshold change | calibration |
| Will not close in afternoon sun | sun-blinded sensor | free — shade/re-aim |
| Nothing responds, wall button included | power: outlet, GFCI, unplugged unit | free check first |
| Remotes dead, wall button fine | lock switch or radio side | free — programming guide |
| Motor runs, door ignores it | stripped gear or released trolley | repair / free re-engage |
| Lifts inches, quits, sags back | counterbalance failing — springs | STOP & CALL |
| Crooked, one cable slack | single-side cable/drum failure | STOP & CALL |
Why does winter create this exact problem?
The force limit is a fixed ceiling; winter raises the load underneath it. Cold spring steel delivers less counterbalance assist, thickened grease drags in every bearing, and hardened weather seals grip the slab — so a door that cleared its force limit by a comfortable margin in September trips it in the first hard week of January. The opener is not broken; it is honestly reporting that the door got harder to move.
The pattern is measurable in service data everywhere winters are real: close-failure calls cluster into the first sustained freeze, right alongside spring failures — the same climate load driving both. The prevention is unglamorous and works: a fall lubrication pass with door-rated product, the balance test while the weather is still kind, and seals replaced before they harden into grip. The city pages on this site carry each county’s NOAA-derived seasonal calendar for exactly this reason.
What are travel and force limits, really?
The two calibrations behind half of this page — worth understanding even though setting them is pro work.
Travel limits: WHERE to stop
Stored positions for full-open and full-closed. Set shallow, the door hovers above the slab; set deep, it presses into concrete and self-reverses as if obstructed. Floor changes — new epoxy, thicker seal — move the target under a fixed setting.
Force limits: HOW HARD to try
The maximum effort before the opener assumes an obstruction. It is a safety ceiling, not a power setting — raising it to bully a binding door defeats the protection that stops the door meeting a person.
Why pros set them with a block
Correct calibration ends with the reversal test: a board flat under the door must trigger immediate retreat on contact. Anyone can turn the screws; the test with an obstruction is what makes the setting safe.
The auto-relearn wrinkle
Many modern units re-learn travel on their first cycles after power events — which is why a door sometimes "fixes itself" after an outage, and why odd behaviour right after one deserves a watched full cycle before any panic.
Won't close, won't open — the questions that matter
Why does the door close only while I hold the wall button?
That is a designed behaviour, and it is diagnostic gold: constant-pressure close is the opener’s legal fallback when the photo-eye circuit reports a fault. The eyes are blocked, misaligned, or their wiring has failed — the opener will still close under your continuous command but refuses automatic closing. Fix the eyes (clean, align until both indicator lights are steady, check the wire staples) and automatic closing returns.
Both sensor lights are on and steady — why does it still reverse?
Then the eyes are telling the truth and something else is triggering reversal: most commonly the closing-force limit tripping against a door that has stiffened (cold weather, dry rollers, failing springs), or a bent track section binding mid-travel. Watch WHERE it reverses — at the same spot every time suggests binding there; at the floor suggests the close-limit setting overshooting into the slab. Force and limit calibration is pro territory: mis-set force is how doors damage what they meet.
Why will it not close in bright afternoon sun?
Low direct sun can flood one photo eye and mimic an obstruction — the classic west-facing-garage complaint, worst in winter sun angles. Confirm by shading the sunstruck sensor with your hand or a piece of cardboard while someone tries the door. Durable fixes: a sensor hood (cardboard works, purpose-made shields look better), swapping which side emits (some pairs allow it), or slight realignment that keeps both LEDs steady.
The remote does nothing but the wall button works — closed case?
Radio side, yes. In order: remote battery; the lock/vacation switch on the wall console (it blocks all radio inputs by design — the everything-at-once signature); range collapse from a cheap LED bulb in the opener socket jamming the receiver; then re-pair the remote via the Learn button. The programming guide on this site walks every step. If the receiver pairs and then forgets, that is electronics — a bench repair.
It reverses right at the floor and pops back up — what is that?
Two candidates: the close travel limit set a touch too deep, so the opener presses into the slab and interprets it as an obstruction; or a threshold/weather-seal change (new epoxy floor, thicker seal) that moved the effective floor height. Both are calibration answers, not parts. If the door also slams the last foot when closing manually, add fading springs to the suspect list — that is the counterbalance no longer easing the final arc.
When is a won’t-close door actually dangerous?
Three states mean stop using it entirely: the door is visibly crooked or one cable hangs slack (single-side failure — the door can jump the tracks); a loud bang preceded the problem (snapped spring — the opener is lifting dead weight it was never rated for); or the door free-falls any distance when the release is pulled. Those are stored-energy failures. Everything observational on this page stays safe; those three go straight to a licensed pro.
Why did it stop closing the week it got cold?
Cold stiffens the whole system: spring steel loses assist, lubricants thicken, weather seals harden and grip the floor. A door that was marginal in October fails its force limit in the first hard week of winter — which is why won’t-close calls cluster with the first freeze. A door-rated lubrication pass and a balance check in fall prevents most of it; a door that got dramatically heavier is springs, not weather.
Can I just hold the wall button every time and live with it?
You can — constant-pressure close is designed to work through eye faults — but understand what you have chosen: automatic closing is gone, the entrapment system is effectively bypassed for that operation, and everyone else in the house inherits a door that behaves strangely. The underlying fix is usually a ten-minute alignment. Living with the workaround costs more attention than the repair — and it masks the day a second, genuinely mechanical fault arrives, because the door already “always acts weird”. Fix the eyes once and the door goes back to being a machine whose complaints mean something — the diagnostic value of a quiet baseline is real, and it is the cheapest instrumentation a garage will ever have — a door that only speaks when something changed.
Why does it close fine at night but not during the day?
That schedule is the sun talking: direct low-angle light flooding one photo eye reads as an obstruction during specific hours and resolves at dusk. West-facing openings show it on winter afternoons; east-facing on summer mornings. The shading test in the sun FAQ above confirms it in one door-cycle, and a sensor hood or emitter-side swap retires it permanently.
The door closed on my car — is that a settings problem?
It is a failed safety system, and it outranks every convenience issue on this page. Either the photo eyes were bypassed, misaligned or wired out, or the closing force is set far above spec — in any case the entrapment protections both missed. Stop using automatic close entirely and book the calibration visit: the reversal test against a block, passed and demonstrated, is the only acceptable end state.