Home  ›  Blog  ›  Bus Duct Commissioning Tests

How to Test a Bus Duct Before Commissioning

By DHYN Engineering Team  ·  August 17, 2026

Quick answer: Before you energize a new bus duct, do four things. Test insulation resistance with a megohmmeter, 1000V DC on a low voltage duct, where 1 megohm is the pass floor and a healthy new duct reads in the hundreds. Torque every joint bolt to the manufacturer's value and mark it done. Verify phase sequence and protective earth continuity end to end. Then energize at no load, bring the load up, and finish with an infrared scan after 2 to 4 hours. (Answer time: August 17, 2026)

A bus duct arrives on site as a pile of factory-tested sections, but it becomes a system in your building, bolted together by an installation crew. Every one of those field-made joints is a place the factory could not test. That is why commissioning tests exist, and why the failures we get called about months later almost always trace back to a test that was skipped or done in a hurry. Our bus duct installation guide covers the mechanical side.

Test 1: insulation resistance, section by section

This is the main event. Use a megohmmeter at 1000V DC for a low voltage busway (500V is acceptable if the manual says so). Measure between each phase and earth, and phase to phase. The pass criterion most specifications use is 1 megohm minimum, which follows the IEC 61439 idea of 1000 ohms per volt as a functional floor. But treat 1 megohm as the floor, not the target. A new, dry sandwich bus duct typically reads in the hundreds of megohms, often off scale. If a brand new duct reads 2 megohms, something is wet or contaminated, and the right move is to find it, not to file the report.

Two practical points make the difference. First, test each section as it comes off the truck, then again after assembly. A single damp section hides inside an average when you only megger the whole 80-meter run. Second, disconnect anything electronic before you apply 1000V: intelligent tap-off meters, monitoring modules, surge devices. A megohmmeter will kill them quietly.

Sequence matters too. Do this test after the duct is fully installed but before wall penetrations get sealed. Breaking a finished firestop to fix a wet joint, then resealing, is a cost nobody budgets for. We went into the sealing side in how to seal a busway through a fire-rated wall.

Test 2: joint torque, the one that decides the next twenty years

Inside every joint, copper faces are pressed together by bolts, and contact resistance depends on that pressure. Under-torqued joints run hot from day one. Over-torqued joints can crush washers or extrude soft tin plating out of the contact area. So use a calibrated torque wrench and the manufacturer's table, which for typical joint hardware lands somewhere between 20 and 70 N·m depending on bolt size. Guessing is not a strategy.

Many modern joints use shear-head bolts that snap off at the correct torque. They make life easier, but check two things: that every head has actually snapped, and that nobody snapped a head and then left the joint half-seated, because a sheared head proves torque, not assembly. Where a joint pack has two or more bolts, tighten in a cross pattern. Mark each finished joint with a paint pen. On a riser with 40 floors, the paint marks are the only thing standing between you and a missed joint.

Ask the manufacturer whether joints need a re-torque after the first heat cycle. Some designs with Belleville washers do not. Some plainly do. Put the answer in the commissioning report, and if a re-torque is required, schedule it with the first annual maintenance visit.

Test 3: phase sequence, continuity and the earth path

Run a phase rotation meter across both ends of the run. This sounds trivial until a duct feeding motors or paralleling with an existing supply goes in with reversed rotation, and the fix means breaking joints apart. While you are at it, confirm the neutral is not crossed with a phase, something that happens more often than anyone admits on 4-wire runs with look-alike bars.

Then prove the protective earth. On most busways the aluminum or steel housing is the earth path, with bonding plates or spring contacts at each joint. Measure continuity end to end with a low-resistance ohmmeter. You want a reading well under 0.1 ohm for a typical run, and every joint's earth contact visibly seated. A duct that carries 1600 amps of load current and has a broken earth path is an accident with a schedule.

Energizing: no load first, then heat

Energize with every tap-off box switched off. Listen. A healthy duct is nearly silent; a buzz means something loose, and it will not get better on its own. After half an hour at no load, walk the run and feel the housing for warm spots with the back of your hand. There should be none worth mentioning.

Then bring the load up progressively and, after 2 to 4 hours at 30 percent or more of rating, scan every joint and tap-off with an infrared camera. IEC 61439-6 caps temperature rise at connections as part of design verification, so the duct has limits by design; your scan is checking the assembly, not the design. The field criterion that works is comparative: any joint running 10 to 15 degrees C hotter than its neighbors under the same load gets investigated. We covered what hot joints look like and what to do about them in busway joint overheating: 5 warning signs. One warning: an IR scan at 5 percent load shows you nothing, because joint heating scales with the square of the current. Wait for real load.

The checklist at a glance

Test	Instrument	Pass criterion	When
Insulation resistance	1000V DC megohmmeter	1 megohm minimum; expect 100+ megohms on new duct	Per section and full run, before sealing penetrations
Joint torque	Calibrated torque wrench	Manufacturer value (typically 20 to 70 N·m), paint-marked	During and after assembly
Earth continuity	Low-resistance ohmmeter	Well under 0.1 ohm end to end	After assembly
Phase rotation	Rotation meter	Same rotation at both ends	Before energizing
Thermal scan	Infrared camera	No joint 10 to 15 degrees C above its neighbors	2 to 4 hours at 30%+ load

Mistakes that waste the whole exercise

The classic is meggering the run with smart tap-off units still connected and writing off the dead electronics as bad luck. Close behind: testing only the assembled run instead of section by section, so one wet length slips through. Third, scanning with the building half-occupied at light load and declaring the joints healthy. And fourth, throwing away the numbers. Record every value in a commissioning report, because those readings are the baseline your annual inspections will be compared against. A joint resistance that drifts year over year tells you a story; a single number with no history tells you nothing.

None of this needs exotic equipment or more than a day on a typical project. What it needs is the discipline to test before the ceiling closes and the walls get sealed, because that is the last moment a problem is cheap.