Conduit Run Rejected Three Days Before Paving
On a new signal installation the contractor has installed a 4-in PVC conduit run from the controller cabinet to the far-side pole: 260 ft, laid at 18-in cover beneath a gravel shoulder that is scheduled to be paved Monday. The run has three factory 90-degree sweeps and no pull boxes. Inside it, the crew pulled three unjacketed twisted-pair signal cables and one insulated equipment-grounding conductor, terminated in the pole handhole, which has 8 in of slack coiled in the bottom. The cabinet has a new grounding electrode conductor to a 10-ft rod; the pole has its own rod, unbonded, and the two are on opposite sides of the roadway. The agency specification requires pull points at a maximum of 150 ft and no more than two 90-degree bends between pull points, plus NEC and agency grounding compliance. The inspector must accept or reject today.
1. List the deviations, rank them by risk, and state what each one leads to in service.
Read the worked answer
Rank by consequence, not by convenience of fixing.
- Three 90-degree bends with no pull points over a 260-ft run. The agency spec (and standard practice) allows at most two 90s between pull points and pull boxes roughly every 150 ft or less. Three 90s in 260 ft makes the pull tension unpredictable, abrades the jacket, and makes any future cable replacement effectively impossible without trenching. Realistic outcome: damaged insulation now, or a dig-up later.
- Unjacketed cable in a wet, direct-buried conduit. Signal cable in a buried raceway that drains is a wet location; it needs a jacketed, wet-location-rated assembly per the adopted code (Article 800 and the agency detail). Bare twisted pairs wick water into the handhole and into the splice.
- Grounding electrodes not bonded. The cabinet rod and pole rod on opposite sides of the roadway create a potential difference between two metal structures that a pedestrian can touch. That is the classic shock and surge path.
- 18-in cover under a gravel shoulder. Technically defensible today, but the shoulder is being paved Monday; confirm the design cover for a paved area with the agency before backfill, because the trench is cheapest to deepen now.
- Slack coiled in the handhole is acceptable if it is neatly dressed and does not block the handhole or sit in standing water; confirm drainage.
2. What is the minimum you require before the trench is closed and the shoulder is paved?
Read the worked answer
Do not allow backfill or paving until the following are complete and witnessed:
- Install pull boxes to break the run into segments of 150 ft or less with no more than two 90s between pull points. This may require opening the trench further - accept that cost now.
- Replace the unjacketed cable with jacketed, wet-location-rated signal cable, or obtain a written, sealed acceptance from the engineer of record if the agency has an approved alternate detail. Get it in writing; verbal approvals disappear.
- Bond the two ground rods into a single grounding electrode system with a properly sized copper bonding jumper (sized per the grounding-electrode-conductor rules, typically 6 AWG minimum for this application - verify with the agency detail), using listed clamps, and record the resistance-to-ground reading. Many agencies require a specific maximum; measure and record whatever the spec says.
- Verify conduit grade/drainage and cap open ends; confirm cover depth for the paved condition.
- Photograph the open trench, the pull boxes, the sweeps, and the bonding connections, and get the readings into the project file before the concrete and asphalt go down.
Records to close out: trench inspection form, insulation resistance test results after the re-pull, grounding test results, and a photo log tied to stationing.
3. The contractor argues that because the cabinet and pole rods are separate electrodes, they do not need to be bonded. Respond, and give the practical field test that settles it.
Read the worked answer
The argument is backwards.
- Separate grounding electrodes at a single premises or structure must be bonded together to form one grounding electrode system. Two isolated rods on opposite sides of a roadway sit at different potentials during a nearby strike, a utility fault, or a neutral shift. Any metallic path between them - the signal conductors, the conduit, the pole itself - becomes the bonding conductor, and it is not sized for that current.
- The failure mode is real and common: a person touching the pole and the cabinet, or standing on wet pavement between them, becomes part of the gradient.
- The signal conductors are also the surge path into the controller. An unbonded pole rod gives surge current a route through the controller's terminal facility instead of to earth.
Field test to settle it: with the system de-energized, measure resistance between the pole and the cabinet grounding terminals using a low-resistance ohmmeter (not a standard multimeter). A bonded system reads near zero; an unbonded one reads the resistance of the soil path, which is unstable and often surprisingly low - which is why the measurement must be made, not assumed. Then measure each electrode's resistance to remote earth with a fall-of-potential or clamp-on tester and record both values. If the bonding jumper is missing, install it and re-measure.