Key takeaways
- A millivolt drop test measures the voltage across a joint at a known current, so the result converts to resistance in micro-ohms.
- Use four wires: current through the joint, sense probes on the joint, and keep the sense leads out of the current path.
- Take the reading at a stated current and temperature, after the value has stabilised.
- Acceptance is a limit for that joint or a comparison with an equal length of the same braid, recorded for every joint.
What the test measures
A millivolt drop test passes a known current through a joint and measures the small voltage that appears across it. Ohm's law turns the reading into resistance: one millivolt at one amp is one milliohm, and one millivolt at ten amps is one hundred micro-ohms. The result is the resistance of the joint plus whatever length of conductor sits between the probes.
The test is used because a joint resistance is too small for an ordinary ohmmeter to resolve. A good braid joint may sit in the tens of micro-ohms. A handheld meter with a two-wire lead set includes the lead resistance and the probe contact resistance, which can be larger than the joint itself.
The value of the test is comparative. A single reading tells you little unless you know the limit for that joint. A reading before and after thermal cycling, or a reading compared with an equal length of the same braid, tells you whether the joint is changing.
Choosing the test current
The test current has to be high enough to produce a readable voltage and low enough not to heat the joint during the measurement. A current that is too low gives a reading buried in noise and offset. A current that is too high raises the joint temperature while you are measuring it, so the resistance drifts.
Many procedures use a fixed reference current, such as 10 A or 100 A, so results from different joints and different dates can be compared. Others use the rated current of the link. Either is workable if the basis is stated. A reading at an unstated current cannot be compared with anything.
Use direct current for the measurement. Alternating current introduces reactance and phase effects, and the braid geometry makes the current distribution frequency dependent. A DC reading is the stable, repeatable number that a specification should quote.
State the current with the result
A resistance value without the test current, the probe spacing and the temperature is not a result. It is a number that cannot be repeated or compared.
Probe placement and lead routing
A millivolt drop test uses four wires. The current source connects to the two ends of the assembly, so the test current flows through the joint. The sense probes touch the conductor on either side of the joint, and the voltmeter measures only the small voltage between them. The sense leads must not carry the test current.
Place the sense probes as close to the joint as the geometry allows, and mark the spacing. If the probes are placed far apart, the reading includes the resistance of the braid and the result looks worse than the joint. If they are placed on the plating or on a painted surface, the reading includes the contact resistance of the probe.
Route the sense leads away from the current leads and away from any source of electrical noise. Twist the sense pair or use a shielded lead. A long, loose sense lead can pick up enough noise to shift a micro-ohm reading, especially when the test current is switched.
Temperature and stabilisation
Copper resistance rises with temperature, by roughly 0.4 per cent per kelvin near room temperature. A joint measured cold and the same joint measured warm will give different readings. Record the conductor temperature at the time of the test, or let the assembly settle to a stated ambient before measuring.
Let the reading stabilise. When the current is first applied, the joint warms and the voltage drifts. Wait until the value is steady over a stated period, then record it. A reading taken immediately after switch-on can be several per cent away from the steady value.
For a critical joint, repeat the measurement after thermal cycling. A joint that reads well cold but rises after cycling has a contact problem that the first reading hid. The change between the two readings is often more informative than either value on its own.
Setting an acceptance limit
There are two practical ways to set a limit. The first is an absolute resistance for the joint, derived from the design and the current it carries. The second is a comparison with an equal length of the same braid, with an allowance for the joint. The second is often easier to defend because it accounts for the conductor.
For a braided link, a common approach is to measure the braid alone over a stated length, then measure the assembly over the same length including the joint. The difference is the joint resistance. A limit on that difference isolates the termination from the conductor.
Set the limit before the first production part, not after seeing the results. A limit chosen to fit the parts already made is not a specification. If the process cannot meet a sensible limit, the joint design needs to change, not the acceptance number.
| Item | Record | Why it matters |
|---|---|---|
| Test current | DC value in amps | Converts millivolts to resistance |
| Voltage | Reading in millivolts | The measured value |
| Probe spacing | Distance and location | Separates joint from conductor |
| Temperature | Conductor or ambient | Copper resistance changes with heat |
| Stabilisation | Time or steady-state rule | Prevents a drifting reading |
| Result | Micro-ohms and limit | Pass or fail against a stated basis |
| Traceability | Part, lot, date, operator | Links the record to the shipment |
A record with the voltage but not the current cannot be converted to resistance. Both belong on the certificate.
The errors that give a false result
The most common error is a two-wire measurement. The lead resistance and the probe contact resistance are added to the joint, and the result is dominated by the test set rather than the part. A four-wire set removes both, provided the sense leads are connected correctly.
The second error is a probe on an oxidised or plated surface. Tin, silver and nickel all have different resistivities, and an oxide layer can add a contact resistance that has nothing to do with the joint. Clean the probe point, or probe the bare copper where the drawing allows it.
The third error is comparing readings taken at different currents, temperatures or probe spacings. A joint that appears to improve may simply have been measured closer to the pad. Keep the method fixed and record the conditions so a later reading can be compared with an earlier one.
Recording and traceability
A millivolt drop result belongs with the part record, not in a notebook on the bench. Record the current, the voltage, the calculated resistance, the probe spacing, the temperature and the limit. Add the part number, the lot, the date and the operator.
For a production run, a sample of joints per lot is usually enough if the process is stable. For a first article or a design change, test every joint on the sample assembly and keep the results. The pattern across the joints shows whether the process is centred or whether one termination is carrying the risk.
If the joint is critical, keep the record with the shipment. A buyer who receives a braid assembly with a millivolt drop record for each joint has evidence that the termination was checked, not just that the conductor was supplied.
Frequently asked questions
What is a good millivolt drop for a braid joint?
There is no universal number. It depends on the joint size, the current and the probe spacing. Set a limit from the design or from the resistance of an equal length of the same braid, then test every joint against that limit at a stated current.
Why use four wires instead of a normal ohmmeter?
A two-wire meter includes the lead and probe contact resistance, which can be larger than the joint. A four-wire measurement passes current through separate leads and senses only the voltage across the joint, so micro-ohm values are meaningful.
Should the test be done hot or cold?
Do it at a stated temperature and record it. Copper resistance rises with heat, so a hot reading is higher for the same joint. For critical joints, measure cold and again after thermal cycling to see whether the contact has changed.
Can a millivolt drop test find a loose bolt?
It can show that the joint resistance has risen, which is a strong indication. It does not tell you whether the cause is torque, contact surface or corrosion. Follow a failed reading with a torque check and a visual inspection of the interface.
Sources and further reading
- IEC 61238-1, Compression and mechanical connectors for power cables
- IEC 61439-1, Low-voltage switchgear and controlgear assemblies, Part 1: General rules
- Qumek millivolt drop test method notes
Qumek Engineering Desk
Engineering team
The engineering team behind Qumek drawings, DFM notes, tolerances and inspection standards. Every guide is reviewed against the drawings and processes used on live programmes.
Need a millivolt drop method for a braid assembly?
Send the joint drawing and the current. The engineering desk will set out the probe positions, test current and acceptance record.
Send a drawing