Busbar testing: millivolt drop, pull test and thermal cycling

Why test a busbar at all

A busbar is a simple part until it is installed. Then it becomes a current path with joints, supports, insulation and a fault duty. A dimensional check confirms the shape, and a material certificate confirms the grade, but neither tells you whether a joint will run hot at the rated current or whether a welded tab will hold under vibration.

Testing is how the design assumptions become evidence. The useful tests are not complicated: a millivolt drop across a joint, a pull or shear test on a welded connection, a thermal cycle that repeats the service temperature, and an insulation test on a coated or isolated bar. Each one answers a question that the others cannot.

Millivolt drop

The millivolt drop test passes a known current through the joint and measures the voltage across it. The probes are placed at a fixed distance either side of the joint, and the result is compared with the voltage across an equal length of solid bar at the same current, or with a limit in the specification. A four-wire meter removes the resistance of the leads from the reading.

The test is sensitive to the things that matter: contact area, surface films, plating, pressure and contamination. It is also sensitive to the way it is performed. Probe spacing, probe pressure, surface condition and current all change the result. A reading taken across a painted face, or with the probes too close to the bolt, is not a joint measurement.

Use the test at first article, on a sample from each lot and after any change to the joint preparation or plating. Keep the current and the probe spacing with the result. A millivolt drop figure without those two numbers cannot be compared with anything.

Pull and shear tests

A pull test applies a tensile load to a welded joint, a welded stud or a terminal, and checks that the connection holds. A shear test loads the joint in the plane of the connection. Both are destructive, so they are run on samples or on sacrificial parts, not on the shipment.

For a welded busbar joint, the test is usually a peel, a shear or a bend, depending on the joint geometry. The result should be recorded as a load and a failure mode. A joint that holds a high load but pulls out of the bar, rather than failing in the weld, is telling you something different from a weld that fails in the nugget.

A bolted joint is not qualified by a pull test in the same way. Its strength comes from the clamp force and the friction between the faces, and the useful checks are torque, fastener proof load and contact resistance. If a pull test is specified for a bolted terminal, it is usually a check on the terminal or the stud, not on the electrical joint.

Thermal cycling

Thermal cycling takes the assembly through the temperature range it will see in service, usually from ambient to the rated operating temperature, and back. The purpose is to expose the change that a single room-temperature test cannot: the joint loosens, the plating creeps, the coating cracks or the contact resistance drifts.

A useful cycle test measures the joint resistance and the fastener torque before and after the cycles, and checks the insulation if the bar is coated. The number of cycles, the temperature limits and the dwell time should come from the specification or the applicable standard. A test with an undefined cycle is a demonstration, not a qualification.

Thermal cycling is also the best check for a coating boundary and a masked joint face. The coating and the metal expand at different rates, and the boundary is where a crack starts. A visual check after the cycles, plus an insulation resistance measurement, will show whether the coating survived.

Insulation tests

A coated or isolated busbar needs an insulation test as well as a current path test. The usual checks are insulation resistance and a dielectric withstand test at a stated voltage and duration. The test voltage depends on the system voltage, the standard and the pollution degree, and it should be stated on the drawing or the test schedule.

The test is only meaningful if the bar is clean and dry and the test points are defined. A test from the bar to a metal block is not the same as a test from the bar to the enclosure with the supports fitted. For a production routine, a simple go or no-go test at a defined point is often enough.

A coating pinhole may pass a low-voltage continuity check and fail a dielectric test, or the reverse, depending on the voltage and the moisture. That is why the test method and the acceptance limit belong together, and why a visual check at bends and edges is still part of the routine.

Sampling and sequence

The sequence matters. A destructive test should not be run on the same parts that are used for the electrical test unless the test plan says so. A thermal cycle should come before the final resistance measurement, because the cycle is what changes the joint. An insulation test should come after the mechanical and thermal tests, because those are the tests that can damage the coating.

Sampling should reflect the risk and the production volume. A new joint design, a new plating line or a new welding procedure needs a qualification sample before production. A stable process can move to a periodic sample, provided the periodic test is the same test with the same acceptance limit. Reducing the sample size is a decision, not a drift.

Write the plan as a table with the test, the sample, the sequence, the current or voltage, the duration and the acceptance limit. That table is the difference between a test report and a set of readings.

1. 01

   Define the duty

   State the rated current, the fault current and duration, the temperature range and the insulation requirement.

2. 02

   Qualify the design

   Run the full set on representative samples: drop, strength, cycling and insulation.

3. 03

   Set the routine

   Choose the production checks and the sample frequency, with the same limits as the qualification.

4. 04

   Record the conditions

   Keep the current, voltage, temperature, probe spacing and sample identity with every result.

5. 05

   Act on the result

   Contain the lot, find the process cause and re-qualify after a change, rather than retesting until it passes.

Acceptance criteria and records

An acceptance criterion should be a number with a condition. A joint drop of a stated value at a stated current and probe spacing is a criterion. A note that says the joint should be tight is not. The same applies to a pull test: a load and a failure mode, not a pass stamp.

The records should identify the sample, the drawing revision, the material and plating lots, the test equipment and the operator. If the test is destructive, record the failure location and keep a photograph if the programme allows it. If the test is non-destructive, record the reading and the point where it was taken.