Improve Cable Reliability with Milliohm Resistance Measurement

By Contributed Article | September 08, 2015

Milliohm-sensitive four-wire measurement helps to locate bad solder joints, faulty crimps, recessed pins, pin contact contamination, improper wire gauge, and stress-extruded wire.

 

Four-wire Kelvin measurement makes it possible to accurately measure resistance values less than 0.1Ω while eliminating the inherent resistance of the lead wires connecting the measurement instrument to the component being measured. A high-quality digital multi-meter (DMM) test lead (24″ in length with banana plugs at both ends) will typically present a natural resistance of about 0.1Ω. When using two such leads connected to the unit under test (UUT), we can expect at least 0.2Ω of the measured value to be contributed by the meter leads alone. In addition, the slightest bit of contamination on the banana plug leads, whether from finger oils, dust, or corrosion, will add additional resistance or cause the measured value to appear variable when the lead wires or plugs are flexed.

Precision low-value resistance measurements become necessary when testing cables intended to carry significant current, or when extremely high reliability must be ensured in medical or military applications. For cables intended to transmit power, as you might find in AC power distribution or radio transmitters, low-resistance connections that have faulted to high-resistance create the real danger of fire or explosion. Although cables used in medical applications, for example, do not carry significant current, questionable connections to hair-thin wires connecting a monitor to high-precision sensors could create life-critical measurement errors or result in circuit misoperation. Milliohm-sensitive four-wire measurement helps to locate bad solder joints, faulty crimps, recessed pins, pin contact contamination, improper wire gauge, and stress-extruded wire.

What is Four-Wire Measurement?

Ohm’s law defines “resistance(R), as the ratio of voltage (V) across a component, to the current (I) passing through it:

R = V/I

To measure resistance, we apply a test current to a wire and detect the voltage drop developed. From this, we easily calculate the resistance as shown in the following figure.

Cable resistance

We measure the resistance of interest, RW, between the conductor’s two mating pins. The entire circuit, however, includes the resistance of the lead wires, RL1 and RL2, so the voltage drop used in the calculation includes all three of these resistances. In many situations, the lead wire resistance is much lower than the resistance of the conductor or component we aim to measure and therefore can be disregarded.

In some situations, however, the resistance of interest, RW, approaches the resistance value of the lead wires used to measure it, resulting in an inaccurate reading. We correct this problem by moving the voltage measurement points out to the endpoints of the mating pins, thus bypassing any voltage drop that may occur in the lead wires. Refer to the figure below:

Voltage measurements

OhmmeterThe Ohmmeter then appears to have four wires coming from it. The image at the right shows these terminals on a typical DMM. Because we now use four lead wires instead of two, we refer to this approach as “four-wire measurement,” or alternatively “four-wire Kelvin measurement” in honor of the 19th-century British physicist, Lord Kelvin, who originally developed it.

Note that the current flowing through the voltage-measuring cir­cuit of a four-wire system is extremely small, typically on the order of fractions of a microamp (six or more orders of magnitude less than the source current), so virtually no voltage drop occurs across these lead wires, and its effect on the resistance measurement is negligible. In summary, if there is no current flowing through a wire, there is no voltage drop across it regardless of its length. This important point means that lead wires may now be quite long, sometimes exceeding 10 feet (three meters), without having any effect on the measurement. Long lead wires become necessary when testing large, multi-branch wire harness assemblies, so this situation is not as uncommon as it might seem.

Cable Test fixture

The principal advantage of four-wire measurement is that it eliminates any effect of fixture resistance (the lead wires) to obtain a precise resistance value of the UUT. Because four-wire measurements typically employ test currents well above those needed for two-wire testing, a secondary advantage comes through the use of a high-current stress test for wiring by driving a current of 1A or more through each conductor and the ability to set a dwell time from 100ms to many minutes; observing a slowly increasing resistance during a long dwell period that results from thermal heating may reveal problems not detected with a shorter measurement interval.

Software driving a four-wire measurement system should permit individual conductors within a UUT to be independently disabled from a four-wire test by user selection to avoid potential damage to fuses or other components not intended to carry high test current. Users should also be allowed to independently set different test currents and dwell times for different conductors.

Building Test Fixtures for Four-Wire Measurement

Unlike a benchtop DMM, which has four test connections (two for source and two for sense), modern cable test equipment offers a multitude of programmable test connections, also referred to as test points, to which the UUT may be connected. Typical cable testers start with 128 test points and can be expanded upward into the thousands of points.

The advantages of four-wire measurement come at a cost. First, the test system requires twice the number of test points that would normally be required for two-wire resistance measurement, significantly increasing the equipment cost. Second, test fixtures must utilize two wires for every pin on the mating connector, one wire for the current source, and the other for voltage sense. This increases the cost and complexity of the test fixture.

An example of a four-wire test fixture intended for a 12-conductor cable appears in the image below. If we assume that the UUT is a cable with two connectors, an identical fixture would be needed for both connectors.

The UUT attached to a CableEye test system appears in the accompanying image. You can see here that 48 test points are required to test this 12-conductor cable.

CableEye Test System

Typical high-performance cable test equipment like the CableEye system not only provide four-wire measurement, but also high-voltage testing to check for dielectric breakdown and insulation resistance. With the cost of this equipment ranging from $25 to $50 per test point, depending on the total number of points ordered, the equipment cost can be significant, especially for larger assemblies.

Construction of the test fixture itself contributes to the overall cost of testing because of the number of wires employed, and the requirement that source and sense leads be joined at each pin in the mating connector. When high-density mating connectors must be used, there may not be sufficient space within the backshell of the mating piece to accommodate two wires if the connector was designed for one, and this then requires splicing a short single-extension wire from each mating pin to the wire pair, soldering the three pieces together, and insulating with heat-shrink tubing. Because test fixtures are not mass-produced but instead custom-designed and hand-assembled, a typical test fixture for four-wire harness testing may cost from many hundreds to thousands of dollars.

When building four-wire test fixtures, we typically assign odd-numbered test points for the source and even number pins for the sense. This may also be reversed, if that is your standard. However, once you agree on a standard, all fixtures should be wired in this manner. Keep in mind these additional considerations:

  • Because the source pin can drive a current of 1A or more into a pin, we recommend using 22-gauge or larger wire for this purpose. The sense pin, however, will carry almost no current at all, so the wire used for sense can be much thinner, which might be an advantage when trying to crimp two wires into a single pin of the mating connector.
  • The length of the wire in the test fixture is unimportant in four-wire measurements since the lead wire is not part of the resistance measurement. Four-wire methodology may be especially advantageous when the UUT is “remotely” located. Consider, for example, a UUT placed in an environmental chamber some distance away from the test equipment. Test leads of considerable length can be fed through a sealed access port into the chamber to obtain precision measurement remotely. The total effective fixture length is long, yet its resistance is not part of the measurement. Note that for very long source wires with resistance exceeding about 5Ω, we must ensure that the voltage of the four-wire system can rise sufficiently to drive the specified current through this wire.
  • Before building the fixture, be sure that your test equipment has sufficient test points for the fixture you need – two test points for every pin in your mating connectors. To determine the minimum test point requirement, add up all the pins on all the connectors of your cable or harness, including any ground or shell conductors, and double this number to determine the total required test points.
  • ampmodu socketSeveral testers presently on the market use 64-pin dual-row latch headers as seen below. For this type of test point interface, we recommend using Ampmodu connectors or their equivalent, which consist of a 64-pin socket body and gold-plated crimp-and-poke pins. These connectors offer superior low-resistance connections to the header, high voltage isolation exceeding 1500Vdc, and rugged construction.

When building fixtures for multi-headed cables or wire harnesses with this type of connector, you may wire multiple connectors to a single Ampmodu connector to avoid wasting test-valuable points. When you do this, always ensure that an odd-numbered test point “n” and the even-numbered test point immediately next in the sequence “n+1” attach to the same pin on the mating connector.

64-pin Latch Header

What Difference Does One Strand Make?

This demonstration of four-wire measurement sensitivity begins with a 3.5″ (8.9cm) length of 22-gauge seven-strand wire, UL07-730 connected between two screw terminals. The UUT runs between source test points 49 and 55, with corresponding sense test points 50 and 56 appropriately linked, as shown in the photo at the right.

ampmodu testers

The strands will be cut, one-by-one, with resistor measurements made at each step, to determine how the resistance varies with the number of intact strands. This photo shows three cut strands with four strands remaining.

Test point wiring

Here you see the screen report from the CableEye tester with all but two strands cut.

Test with all but to cables cut

CableEye Tester Report

A table summarizes how the resistance changes as strands are cut, and as a function of dwell time. Interestingly, with only one strand remaining to carry the 1A test current, no heating was detectable by human touch, although clearly the resistance increased slightly with the current applied for one second or longer compared to the initial short 50ms dwell.

Resistance changes table

Application of four-wire Kelvin measurement techniques will improve the quality and reliability of your cable and harness products. Precision resistance measurements of less than 0.1Ω reveal wiring defects not visible to less-sensitive measurements, including bad solder joints, faulty crimps, recessed pins, pin contact contamination, improper wire gauge, and stress-extruded wire. Resistive losses resulting from these defects in applications carrying current above 1A may cause excessive heat generation or fire in wiring, or in the case of measurement circuits which obtain input from precision sensors, may cause false reporting or circuit misoperation. The four-wire Kelvin resistance method not only makes it possible to obtain milliohm- or microohm-sensitive measurements, but eliminates any effect of incidental resistance that would be introduced by test leads or the test fixture. Achieving these benefits, however, requires test equipment with twice the number of test points than would otherwise be necessary and a test fixture with two wires leading from the tester to every pin on the mating connector.

Author Christopher E. Strangio is the president and founder of CAMI Research and holds degrees in electrical engineering from Villanova University and MIT. He has been awarded two patents, developed CAMI’s CableEye PC-Based Cable and Harness Test System, and is a senior member of the IEEE. 

Contributed Article
Get the Latest News
Download Bishop's new Top 100 Connector Manufacturers research report
eBook 2024 The Robots are Coming
x