Tracing a Low-Voltage Run in a Finished House

Why this matters

A connected device that drops out, reboots or reads "no power" is often sitting on the far end of a cable nobody can see: a thermostat wire that detours through the attic, a doorbell run spliced behind a covered-over box, a sensor cable that was cut by a cabinet installer's screw. You can spend an afternoon swapping devices against a cable fault, or you can spend well under an hour finding out what the cable actually does between its two ends. This card is that hour, and the point inside it at which you stop tracing and pull new wire instead.

What you need and what it is

• A tone generator and probe ("toner"). The generator puts an audible signal on a conductor; the inductive probe hears it through drywall and in a bundle. Most are made for telephone and data pairs and are not rated for line voltage. A probe beeping on a wall tells you where the cable is, roughly, not which conductor is good.
• A digital multimeter with a low ohms range. This is the instrument that tells you whether each conductor is continuous and, with a little arithmetic, how long the run really is.
• A short jumper with alligator clips, to tie two conductors together at the far end.
• Labels or tape and a marker.

The procedure

1. Photograph and label both ends before you remove anything. At the device and at the source (the furnace board, the chime, the controller), take a photo of the terminals, then tape-label each conductor by color and terminal. Reason: half of all low-voltage callbacks after a trace are a conductor landed back on the wrong screw. Skip it and you finish the trace knowing exactly where the cable goes and not knowing what it was connected to.

2. Isolate the run at both ends. At the source end, turn off the equipment's service switch or unplug the transformer first, then lift the conductors off the terminals. At the device end, lift them off the device. The cabinet you are reaching into at the source end may also hold line voltage, so the switch goes off before the panel comes off, and the blower and burner compartments stay closed. Reason: a toner signal on a conductor still tied to a transformer or a control board leaks into everything connected to it, and an ohms reading taken through a transformer winding or a board reads the device, not the wire. Never clip a toner onto a conductor that is, or might be, connected to line voltage.

3. Prove continuity, conductor by conductor. At the far end, clip the jumper between two conductors (say R and W). At the near end, touch your leads together first and note the reading; that is your lead resistance, often 0.1 to 0.3 ohm. Then read across R and W and subtract the lead reading. Repeat with every pair until every conductor has been in at least one closed loop. Reason: a toner hears a conductor that is broken ten feet from the end just as loudly as a good one. Only a closed-loop ohms reading proves the conductor goes all the way through.

4. Turn the loop reading into a length. Solid copper has a known resistance per length. At room temperature, 18 AWG is about 6.4 ohms per 1,000 ft and 20 AWG about 10.2 ohms per 1,000 ft (published copper conductor tables). Your loop is two conductors, out and back, so:

• conductor length in feet = (loop ohms / ohms per 1,000 ft) x 1,000
• run length one way = half of that.

Reason: this is the only test that sees a detour. If you expected a 20 ft run and the arithmetic says 50, the cable goes somewhere you have not looked, and that somewhere is usually a splice. Watch the resolution: a meter that reads to 0.1 ohm can only resolve 18 AWG conductor length to about 16 ft (0.1 / 6.4 x 1,000), so use the length as a flag for "much longer than expected", not as a tape measure.

5. Tone the run to find where it physically goes. Put the generator on one conductor of the isolated cable and run the probe along the wall from the device end, then along ceilings, closets and the attic hatch. Mark the wall lightly where the tone is strongest. Reason: once the ohms say the run is longer than the wall, the toner tells you which direction it went. Where the tone suddenly goes louder and spreads, you are usually over a coil of slack or a splice; where it stops abruptly, the cable has gone through a top plate into the attic or down into a floor.

6. Find the splice before you blame the device. Splices hide in predictable places: behind an old device location that was drywalled over after a remodel, at the top plate where a cable was extended in the attic, at a former chime or thermostat position, and in a basement ceiling near the equipment. Reason: a buried splice is where most intermittent low-voltage faults live, because it is the one point where the copper is joined by hand, often with tape, often in an attic that swings 100 F between seasons.

7. Decide: repair, or pull new. Repair a splice when it is reachable without cutting into a finished surface, the cable has the conductors you need, and the cable itself is undamaged on both sides of it. Pull new when any one of these fails. As a shop default, if the trace passes 45 minutes without the fault in your hand, stop and price a new pull; the customer is paying for an answer, and a new run is often the cheaper answer. Tune that number to your own pricing.

8. Re-land, label and record. Land each conductor back on the terminal in your photo. Restore power at the source with the cabinet panels on and stand at the front of the unit for the first call. Then label both ends with what the cable is ("T-STAT, hall, 18/5, R W Y G C") and write the route and any splice location in the job notes. Reason: the next tech who opens this wall will otherwise start from step 1.

Hazards when you open a finished surface

Tracing sometimes means a small inspection hole or a trip into the attic. Each act carries its own control.

• Cutting drywall or plaster. Cut shallow with a hand drywall saw or an oscillating tool set to depth, never a full-depth blade, because the cavity can hold NM cable, a water line or a gas line. In a house built before 1978, assume painted surfaces may carry lead paint: a palm-sized hole is far under the EPA renovation rule's 6 square foot per room interior threshold for minor repair (40 CFR 745, Subpart E), but it still makes lead dust, so mist the cut, catch it on plastic, and clean with a HEPA vacuum, not a broom. Drywall joint compound from before roughly 1980 may contain asbestos; if the house is that age and nobody has tested, do not sand or power-cut the joint.
• Attic entry. Step only on joists or a decked walkway, never between them. Carry water and limit the stay when the attic is hot. Look at the insulation before you move it: fluffy white, yellow or pink loose fill is usually fiberglass, which needs a respirator under your shop's 29 CFR 1910.134 program, gloves and sleeves. Gray-brown, pebbly, glittery fill may be vermiculite, which EPA advises treating as asbestos-containing. If it is vermiculite, you do not dig through it to reach a splice; you stop, and the run gets rerouted where it does not cross the insulation.
• Drilling a top plate or a floor plate. Before you drill down into a wall cavity, look and probe for other cables entering the same cavity, and keep the bit clear of any line voltage you can see. Some jurisdictions require the hole through a top plate to be fire-stopped afterward; check your local code.

Worked example: the thermostat that reboots in winter

The complaint. A single-story ranch, gas furnace in a hall closet, connected thermostat on the hallway wall about 12 ft from the closet. The thermostat reboots two or three times a day, mostly in cold weather, and the customer has already had it replaced once. The furnace board has never logged a fault.

Steps 1 and 2. Photos and labels at both ends. Furnace service switch off, panel off, conductors lifted from the board's terminal strip. Thermostat off its base, conductors lifted. The cable is 18/5.

Step 3. Meter leads touched together read 0.2 ohm. Jumper on R and W at the thermostat end. Raw reading across R and W at the furnace: 0.8 ohm; 0.8 minus 0.2 leaves 0.6 ohm. Jumper moved to R and C: the reading wandered between 0.8 ohm and open as the cable was flexed at the closet. Y and G, jumped to R in turn, both read 0.8 raw, 0.6 corrected.

Step 4. Using the R and W loop: 0.6 / 6.4 x 1,000 = 94 ft of conductor, so about 47 ft one way. At 0.1 ohm resolution that figure is good to about 16 ft of conductor, 8 ft one way, so the run is somewhere between roughly 39 and 55 ft. The straight-line route from closet to thermostat is about 20 ft, including the drop down each wall. The run is at least twice the expected length, which means a detour.

Step 5. Tone on the C conductor. The probe picked it up in the closet wall, followed it up to the ceiling, lost it, and found it again in the attic over a spot in the hallway about 10 ft past the current thermostat. The homeowner confirmed the thermostat had been moved during a remodel.

Step 6. The attic insulation was loose fill, white and fluffy, consistent with fiberglass, with a bag label still stapled near the hatch. With a respirator under the shop's program, gloves and sleeves, standing on joists, the tech found the old cable coming up through the top plate over the original thermostat location, joined to a new length with five wire nuts wrapped in tape. The C splice was green with corrosion inside the nut. Winter attic condensation at that spot explained the cold-weather pattern.

Step 7. The splice was reachable without cutting anything, but the cable on either side had been chewed in two places by rodents, so repairing one splice would leave two damaged sections in service. That fails the "cable undamaged on both sides" test. The tech pulled a new continuous 18/5 from the closet top plate across the attic to the current thermostat's top plate, about 22 ft, abandoned the old run in place with its ends taped and labeled, and fire-stopped both top-plate holes as the local code required.

Step 8 and the check. New loop, R and W jumped: 0.5 raw minus 0.2 for the leads leaves 0.3 ohm. The expected figure for 22 ft one way is 44 ft of conductor x 6.4 / 1,000 = 0.28 ohm, which the meter displays as 0.3. That agrees within the meter's resolution. Every pair read the same. Conductors landed per the photos, furnace panel on, switch closed, first heat call watched from the front of the unit with the doors on. The thermostat's own diagnostics then showed a steady supply with heat and fan calling.

What would have changed the call. Had the attic fill been vermiculite, digging to the splice was off the table. The route would instead have been a surface raceway along the hallway ceiling line from the closet, which costs looks but disturbs nothing. Had the cable been sound on both sides of the splice, re-making that one joint inside a small accessible box would have been the repair, and a 45 minute job.

When the tone lies

• Signal everywhere. An analog toner on an unshielded cable bleeds into nearby cables in a bundle, so the probe sings on several. Switch the probe to its lowest sensitivity, or short the far end of your target pair to ground so its tone drops while the neighbors' do not.
• No signal at all on a cable you can see. The generator may be on a conductor that is broken close to the near end. Go back to step 3; the ohms test already told you.
• The tone follows a metal duct. Long runs strapped to sheet-metal ductwork can couple the tone into the duct. Follow the strap points, not the sound.

References

• Published copper conductor resistance tables (ohms per 1,000 ft by AWG at 20 C)
• EPA Renovation, Repair and Painting Rule, 40 CFR 745 Subpart E; EPA guidance on vermiculite attic insulation
• 29 CFR 1910.134 for respirator use under a written program
• Tone generator and multimeter manufacturer documentation for voltage ratings and range resolution
• See related: Low-Voltage Wiring Basics for Non-Electricians; Leaving a Device Record the Next Tech Can Use