The Garage Zone That Faulted Only in Winter

Why this matters

A detached garage zone faulted on freezing mornings and read perfectly by mid-afternoon, which is exactly when the service tech arrived the first two times. Both visits found the zone normal, swapped a door contact that was not bad, and left. The third visit came at 07:30 on purpose and found the fault in a splice somebody had taped and buried years earlier. A fault that depends on weather has to be measured in the weather, and a loop resistance reading taken at the wrong hour proves nothing.

The complaint

The house had a hardwired panel in the basement. Zone 6 covered the detached garage, about 110 ft away: one wide-gap contact on the overhead door and one recessed contact on the side door, wired in series with the end-of-line resistor at the side door contact, the last device on the loop. See related: End-of-Line Resistor Supervision: Open, Short and Normal.

From December on, the homeowner reported the same thing roughly every morning the temperature had gone below freezing overnight. The keypad showed zone 6 faulted and would not arm away without bypassing it. By the time anyone came home in the afternoon the fault had cleared. Both garage doors were closed throughout.

The two earlier visits were both booked in the afternoon. Each found zone 6 normal on the keypad. The first tech adjusted the overhead door contact's alignment. The second replaced it. Neither visit measured loop resistance, and neither changed anything.

The candidates

Before the third visit the service manager laid out what could make a series loop of two contacts and a resistor go open or out of window on cold mornings and recover by afternoon. There were three.

The overhead door moving. Steel overhead doors contract in the cold and their weatherstrip stiffens, so a door can sit slightly higher or lower in its track on a cold morning. A floor-mounted overhead door contact with a marginal gap could open. This was the hypothesis both earlier visits had acted on.

A failing device or the resistor. A reed capsule with a hairline crack, or a resistor with a cracked body or a corroded lead, can change with temperature.

The wiring between house and garage. A run through cold ground, with whatever connections it has along the way.

All three predict the same symptom on the keypad. They predict different readings at different points on the loop, and that is why the third visit was booked for the hour the fault actually existed.

Account on test, then the readings

The tech arrived at 07:30 with the outdoor temperature at 18 F and the keypad showing zone 6 faulted. Before lifting any zone wire he called the central station, put the account on test for the visit, and had the operator read the test window back. Lifting zone wires on an armed or partially armed system, or restoring them, sends signals. See related: How to Put an Alarm Account on Test and Take It Off.

He disarmed the panel, lifted the two zone 6 conductors from the panel terminals, and measured with a digital multimeter on its resistance range. Then he walked to the garage and split the loop at its far end. The July annual inspection had logged a loop resistance for the zone, which gave him a baseline from a warm day.

 ZONE 6, GARAGE LOOP (EOL value per this panel's
 manual: 2,000 ohm; yours may differ)

 JULY, annual inspection, 84 F
   full loop at panel, doors shut ...... 2,006 ohm

 JANUARY, 07:40, 18 F
   full loop at panel, doors shut
     ........ 3,400 to 9,800 ohm, wandering,
              briefly OL twice in a minute
   devices plus EOL only, measured at
     the garage entry ................. 2,002 ohm
                                       steady
   run only, jumper across the pair
     at the garage entry .... 1,400 to 7,900 ohm
                                      wandering
   pair to each other, both ends open .... OL

 JANUARY, 14:30, 34 F
   full loop at panel, doors shut .... 2,150 ohm

 AFTER REPAIR, next morning, 15 F
   full loop at panel, doors shut .... 2,005 ohm

Reading the readings

The baseline makes sense. The two contacts add essentially nothing when closed. Two conductors of 22 AWG running 110 ft out and 110 ft back is 220 ft of wire, and 22 AWG solid copper is roughly 16 ohm per 1,000 ft at room temperature, so the wire adds about 3.5 ohm. The resistor plus 3.5 ohm, with the resistor a few ohms above its nominal value, gives the 2,006 ohm the July record shows.

The devices were cleared in one reading. At the garage end, with the run disconnected, the two contacts and the resistor read 2,002 ohm, steady, with both doors shut. That is the resistor and two closed reeds. The overhead door contact both earlier visits had worked on was closed and fine. That eliminated the first and second candidates on the morning they were supposed to be failing.

The run carried the whole fault. With a jumper across the pair at the garage end, the run alone should read about 3.5 ohm. It read 1,400 to 7,900 ohm and wandered while the tech watched. Something in series in one conductor was adding thousands of ohms and changing by the second. That is the signature of a corroded or loose connection, not a damaged insulation path. Leakage between two wet conductors would have shown up in the last line as a reading lower than OL with both ends open, and it read OL.

The direction matters. A series fault pushes loop resistance up, toward what the panel treats as an open. The panel saw the loop well outside its normal window on the high side and reported zone 6 faulted, which is the same thing it would report for an open door. By 14:30 the run had recovered enough that the full loop read 2,150 ohm, inside the window this panel accepts as normal, so the keypad showed ready. That is why two afternoon visits found nothing.

Finding the joint

The homeowner did not know how the cable got to the garage. The basement told part of it. The cable left the panel, ran along a joist and went out through the rim joist into a PVC conduit that turned down into the ground. At the garage, a conduit came up beside the slab and turned in through the wall at the side door.

The original installer had been short of cable. Partway along, the run had been spliced: two lengths of ordinary indoor alarm cable, twisted together conductor by conductor, wrapped in electrical tape, and pushed into the conduit. Nobody had recorded it. Indoor alarm cable is not made for a wet location, and underground conduit is wet: water condenses in it and runs into it, and a buried run holds that water against whatever is inside.

The tech did not dig. He pulled the old cable out from the house end to use it as a pull line for the replacement, and the splice came out with it about 60 ft from the house. The tape was soaked and split. One conductor's twist was green with corrosion and had broken almost through. The other was dull but intact.

The best explanation that fits every reading is that the corroded joint was holding contact through a thin, damp, partly broken twist. When the ground and the water in the conduit froze, the joint was flexed and the copper contracted, and the contact became intermittent. In the afternoon it relaxed enough to conduct. The tech could not reproduce that mechanism directly, but he did not need to: the joint was the only connection in the run, the run carried all the excess resistance, and the joint was visibly failing.

If the pull had failed and the conduit had needed to be dug up, the job would have stopped for a utility locate first. State one-call laws require excavators to request a locate before digging, with details and response times that vary by state, and the one-call center does not mark privately owned lines. The garage's own electrical feeder was exactly such a line, so a private locate would have been needed as well.

The repair

A new continuous run. He pulled a single length of cable rated by its maker for direct burial and wet locations through the conduit, house to garage, with no splice anywhere underground. Cable rated for burial carries a jacket and fill that resist water; ordinary indoor cable relies on staying dry. See related: Alarm Cable Types and Which Run Gets Which.

Splices only above grade, and sealed. At the garage, the burial cable ended inside a small weatherproof box on the inside wall above the slab, where it joined the interior cable to the side door contact. The splice used gel-filled connectors made for damp-location low-voltage splices, which fill around the conductor and exclude moisture. The house end was terminated directly at the panel with no splice.

The resistor stayed at the last device. He did not move it, and he did not add a second one at the panel. With the EOL at the side door contact, the panel supervises the entire run, which is the run that had just failed. A resistor fitted at the panel to make the zone read normal would have stopped supervising the 110 ft where the fault lived. See related: End-of-Line Resistor Supervision: Open, Short and Normal.

The finish test

Before closing the weatherproof box he measured at the panel with the zone still lifted: 2,004 ohm, doors shut. He opened the overhead door and the reading went to OL; shut it, 2,004. Same for the side door. He landed the zone, confirmed zone 6 ready on the keypad, armed the system, opened the side door, and called the station to confirm the zone 6 alarm had arrived under the test window. Then he closed the box, took the account off test, and confirmed with the operator that it was back in normal service.

The repair was only proven the next freezing morning, so the tech booked himself a short stop at 07:30. At 15 F the keypad showed ready. He put the account on test again, lifted the zone at the panel, read 2,005 ohm steady, relanded it, confirmed a test signal and took the account off test. That reading went on the ticket next to the July baseline and the failing ones, and the homeowner reported a ready keypad on every freezing morning for the rest of the winter.

What the first two visits needed

Neither earlier tech was careless. They tested what they could see at the hour they were sent, and the zone was normal. The missed step was a loop resistance reading compared against the baseline, which the annual inspection had already recorded. A 2,150 ohm afternoon reading against a 2,006 ohm July baseline was 144 ohm high on a loop that should vary by a few ohms, and that alone would have pointed at the wiring on the first visit, without waiting for a cold morning. Dispatch also helped by sending the third visit at the hour the fault existed. A fault that follows the weather or the time of day gets booked for that weather or that hour.

References

• See related: End-of-Line Resistor Supervision: Open, Short and Normal
• See related: How to Put an Alarm Account on Test and Take It Off
• See related: Alarm Cable Types and Which Run Gets Which
• Panel manufacturer installation manual for EOL value and normal loop window
• Cable and connector manufacturer ratings for direct burial and damp-location splices