Key takeaways
- Use a four-wire measurement so the probe and lead resistance stay out of the reading.
- State the test current, the probe spacing and the joint condition, or the millivolt value cannot be compared.
- Acceptance comes from the drawing, the customer standard or a reference joint, not from a convenient round number.
- A millivolt drop proves electrical continuity, not mechanical strength or thermal performance.
What the test measures
A millivolt drop test measures the voltage across a joint while a known current passes through it. The result is a small voltage, usually reported in millivolts, and it is proportional to the resistance of the joint at that current. A low drop indicates a large effective contact area and a sound connection.
It is used on busbar joints, cable lugs, contact assemblies, braids and welded or bolted connections. It is a comparative test: the number means something against a reference, a limit or a previous result, not on its own.
The four-wire method
Use four wires. The current leads carry the test current through the joint; the voltage probes measure only the drop across it. This keeps the probe contact resistance out of the reading. A two-wire measurement adds the lead and probe resistance and can make a good joint look bad.
Place the voltage probes on the conductors either side of the joint, at a defined spacing. The measured value then includes the joint and a known length of conductor. If you want the joint alone, subtract the conductor resistance for the same length, or compare with a reference sample.
Keep the probe tips clean and sharp. A probe that sits on paint, oxide or a plated surface with high contact resistance adds a variable that has nothing to do with the joint.
Choosing the test current
Choose a current that gives a readable voltage without heating the joint. Too low and the resolution of the meter dominates. Too high and the joint warms, the resistance rises and the reading drifts during the test.
The test current should be stated in the specification. Some programmes use a current related to service, others use a fixed test current for comparison. Either is workable if it is recorded and applied consistently.
For illustration, 100 A through a joint of 0.5 milliohms gives 50 millivolts. That is arithmetic, not an acceptance limit. The limit comes from the drawing, the customer standard or a reference joint.
Preparing the joint
Prepare the joint as it will be installed: clean, dry, correctly torqued and free of grease. A test on a finger-tight joint measures the fixture, not the design.
Record the torque, the surface condition and the time since assembly. A joint tested immediately after assembly may still be settling. A joint that has been in service may have oxide or relaxation. Both are valid states, but they are not the same test.
Test at a stable temperature. Copper resistance rises with temperature, so a joint tested hot will read higher. Note the ambient or the joint temperature with the result.
Acceptance criteria and test variables
The acceptance criterion should be a resistance or a voltage at a stated current, defined for the joint type. Where no limit exists, compare with a reference joint of the same design, or with the resistance of an equal length of conductor. A joint that reads much higher than the conductor deserves review.
A single reading is weak evidence. Take several, rotate the probe position and record the spread. A joint with a stable low reading is different from one that averages low but varies.
| Variable | Typical choice | Effect on the reading |
|---|---|---|
| Test current | A fixed value stated in the procedure | A higher current gives a larger reading but heats the joint |
| Probe placement | Either side of the joint at a defined spacing | Wider spacing includes more conductor resistance |
| Probe contact | Clean, sharp tips on bare metal | Oxide or paint adds resistance and noise |
| Joint torque | The specified value, checked before the test | A loose joint reads high; over-torque can damage it |
| Surface condition | As installed, or as agreed for the test | Grease, oxide and plating change the interface |
| Temperature | Recorded at the time of the test | A higher temperature raises the reading |
| Stabilisation | Read after the value settles | A drifting reading is not a result |
State the current, the probe spacing and the joint condition in the test record. Without them the millivolt value cannot be compared with anything.
Common errors that change the number
Most bad results come from the setup rather than the joint. The same joint can pass and fail on the same day if the method changes between tests.
- Using a two-wire meter and including the lead resistance in the result.
- Placing the probes across the plating only, rather than across the joint and into the conductors.
- Testing at a current too low to resolve the difference between a good and a marginal joint.
- Testing a joint that has not been torqued to the specification.
- Reading the display before it stabilises, or while the joint is still warming.
- Comparing results taken at different currents, probe spacings or temperatures.
Recording and trending
Record the part, the joint, the torque, the current, the probe spacing, the temperature, the reading and the pass or fail. Link the record to the lot and the operator so a later query can be answered.
Trend by joint type and by shift. A rise across a production run can point to a tool, a torque setting or a surface preparation change before a failure appears at the customer.
When a millivolt drop result is not enough
Millivolt drop is an electrical check. It does not prove mechanical strength, thermal performance or corrosion resistance. Use it with a pull test, a torque test, a thermal cycle or a cross-section where the joint is critical.
For a joint that will see thermal cycling, test before and after the cycles. A joint that passes cold can open when the materials expand differently, and the millivolt drop after the cycle is the evidence that matters.
A low reading is not a guarantee
It shows that the current found a path. The pull test, the torque record and the thermal cycle show whether the path will stay there.
Frequently asked questions
Why use four wires instead of two?
The two extra wires separate the current path from the voltage measurement. The meter then sees only the drop across the joint, not the resistance of the leads and probes. On a low resistance joint, that difference can be larger than the result.
What test current should be used?
Use the current stated in the drawing, the customer standard or the test procedure. If none is given, choose a value that gives a stable readable voltage without heating the joint, record it, and use the same value for every comparison.
What is a passing millivolt drop?
There is no universal number. The limit depends on the joint design, the current and the conductor. Use the specification limit, a reference joint or the resistance of an equal length of conductor as the basis, and record the method so the result can be reproduced.
Can the test be done on a live joint?
No. The joint should be isolated and the test current supplied by the test set. Testing a live joint risks damage to the meter and to the person, and the service current is not a controlled test condition.
Sources and further reading
- ASTM B539, Standard Test Methods for Measuring Resistance of Electrical Connections
- IEC 61238-1, Compression and mechanical connectors for power cables
- IEC 60943, Guidance concerning the permissible temperature rise for parts of electrical equipment
Qumek Quality Desk
Quality team
The quality team responsible for first article inspection, PPAP packs, supplier audits and the records shipped with production parts.
Setting up a joint test procedure?
Send the joint design and the current. The quality desk will check the probe placement, the acceptance basis and the record fields.
Send the joint details