Whether a connector is a high frequency or a low frequency type, insulation resistance, dielectric withstand voltage (also called electrical strength) and contact resistance are the most basic electrical parameters that determine whether it works reliably. They are specified with clear technical requirements and test methods in groups A and B of the routine quality conformance inspection for connector products, and they are also the basis on which users judge connector quality and reliability. In practice, however, manufacturers and users differ considerably in how they apply the technical requirements, and factors such as instruments, test fixtures, operating methods, sample preparation and environmental conditions directly affect the accuracy and consistency of the results. A focused discussion of these three routine electrical tests and the practical problems in performing them is therefore valuable for improving the reliability of connector inspection.

In addition, with the rapid development of electronic information technology, a new generation of multifunctional automatic testers is gradually replacing single-parameter instruments. These new testers will greatly improve the speed, efficiency and accuracy of electrical performance testing.

1. Insulation resistance testing

1.1 Principle. Insulation resistance is the resistance that appears when a voltage is applied to the insulating part of a connector, causing leakage current on its surface or through its interior. That is, insulation resistance in megohms equals the voltage applied to the insulator in volts divided by the leakage current in microamps. The test determines whether the insulation performance of a connector meets circuit design requirements, and whether its insulation resistance still meets the technical requirements after exposure to high temperature, humidity and other environmental stresses. Insulation resistance is a limiting factor in the design of high impedance circuits: low insulation resistance means high leakage current, which disrupts the circuit and normal operation. If a feedback loop forms, the heat and DC electrolysis produced by excessive leakage current can damage the insulation or degrade the electrical performance of the connector.

1.2 Influencing factors. Insulation resistance is affected mainly by the insulating material, temperature, humidity, contamination, test voltage and the duration for which the test voltage is applied.

1.2.1 Insulating material. The choice of insulating material is very important and often determines whether insulation resistance stays consistently acceptable. One manufacturer originally used phenolic glass fibre plastic and reinforced nylon for insulators; these materials contain polar groups and absorb moisture readily, so insulation performance met requirements at room temperature but failed under hot and humid conditions. After switching to the special engineering plastic PES (polyether sulfone), the product underwent 1,000 h at 200℃ and 240 h of damp heat testing and insulation resistance changed little, staying above 105 MΩ with no abnormal variation.

1.2.2 Temperature. High temperature damages insulating materials and reduces both insulation resistance and withstand voltage. In metal shells, high temperature can cause contacts to lose spring force, accelerate oxidation and degrade plating. For example, GJB 598 series II environment resistant quick disconnect connectors specify insulation resistance of at least 5,000 MΩ at 25℃, falling to at least 500 MΩ at 200℃.

1.2.3 Humidity. A humid environment causes water vapour to be adsorbed and to diffuse over the insulator surface, readily reducing insulation resistance below the megohm level. Prolonged exposure to high humidity causes physical deformation, decomposition, outgassing, breathing effects, electrolytic corrosion and cracking in the insulator. For example, GJB 2281 ribbon cable connectors specify insulation resistance of at least 5,000 MΩ under standard atmospheric conditions, falling to at least 1,000 MΩ after 96 h of damp heat testing at 90% to 95% relative humidity and 40 ± 2℃.

1.2.4 Contamination. Cleanliness inside and on the surface of the insulator has a large effect on insulation resistance. Impurities in the moulding compound or in the adhesive bonding the upper and lower insulator mounting plates, metal debris left by repeated mating, and flux residues from soldered terminations penetrating the insulator surface all noticeably reduce insulation resistance. One manufacturer found a circular connector with insulation resistance of only 20 MΩ between contacts at final acceptance testing; sectioning showed the cause was impurities in the moulding compound, and the whole batch had to be scrapped.

1.2.5 Test voltage. The test voltage applied during insulation resistance measurement has a large effect on the result. As test voltage rises, leakage current does not increase linearly but faster than the voltage, so the measured insulation resistance falls. Connector technical requirements specify the test voltage in the referenced method, usually 500 V. This is why ordinary ohmmeters and DC bridges cannot be used to measure insulation resistance.

1.2.6 Duration and reading time. Because there is capacitance between the measuring electrodes, the supply first charges that capacitance, so the indicated resistance tends to rise gradually at the start of the test, which is normal. Many connector test methods therefore require the reading to be taken one minute after the voltage is applied.

1.3 Discussion.

1.3.1 Inspection environment temperature and humidity. Connector technical requirements usually specify the service ambient temperature and humidity, such as -55℃ to 125℃ and 40 ± 2℃ at 95% ± 3%. The inspection environment is not the same as the service environment. The fact that a product may work in those conditions does not mean the manufacturer must meet the standard atmospheric pressure acceptance criteria when measuring insulation resistance in them. If insulation resistance is measured at the upper service temperature of 125℃ or under damp heat at 40 ± 2℃ and 93% ± 3%, the criteria for high temperature and damp heat testing in the technical requirements apply, not the standard atmospheric criteria.

In practice it is common to find the same batch measuring above 1,000 MΩ and passing at the factory in the dry northern climate with humidity below 50%, then measuring only 100 to 200 MΩ and failing on re-inspection at a user in the humid south with humidity above 80%. Cleaning with alcohol and drying may give a pass immediately, yet the next day the result fails again. Manufacturers should therefore set acceptance levels with a suitable margin above the specified value and should not pass products that only just reach the limit in a dry environment, so that differing inspection climates do not cause disputes. To clarify requirements, some test methods now specify both the temperature and humidity for testing, in a relatively wide range, and the tighter values to be used for arbitration. GJB 1217-91, Test Methods for Electrical Connectors, for example, specifies standard atmospheric conditions of 15 to 35℃, 20% to 80% relative humidity and 73 to 103 kPa, and arbitration conditions of 25 ± 1℃, 50% ± 2% relative humidity and 86 to 106 kPa.

1.3.2 Test fixtures. Connector technical requirements state that insulation resistance between all contacts and between all contacts and the shell must meet the specified value, and that the voltage must be applied for more than one minute. Many manufacturers therefore prepare two or three inspection fixtures with different contact arrangements for each product type, applying the test voltage in parallel between contact pairs, between rows and between all contacts and the shell. Applying voltage in parallel through a fixture is more severe than applying it between individual contacts, so if a fixture test shows a failure, it is permissible to retest without the fixture using probes on single points. Some manufacturers and most users, however, do not use fixtures at all and simply touch two probes connected to the insulation resistance tester between each contact pair or between contact and shell. This has drawbacks: it is highly random and very likely to miss defects, and each point cannot be held for a full minute before reading, so misjudgement is possible and reliability is poorer. Even with a fixture, the fixture itself must be verified, clean and dry, with adequate insulation resistance margin.

2. Dielectric withstand voltage testing

2.1 Principle. Dielectric withstand voltage testing, also called electrical strength testing, applies a specified voltage for a specified time between contacts and between contacts and the shell to determine whether the connector can work safely at its rated voltage and withstand overpotentials caused by switching surges and similar phenomena, thereby assessing whether the insulating material or insulation gap is acceptable. If the insulation contains a defect, the applied voltage will cause breakdown discharge or damage, appearing as flashover (surface discharge), spark discharge (air discharge) or puncture (breakdown discharge). Excessive leakage current may change electrical parameters or physical properties. Because overpotential can damage insulation or reduce its safety factor even below the breakdown voltage, dielectric withstand testing must be performed carefully. In routine testing, if the voltage must be applied repeatedly, it is best to reduce the potential in subsequent tests.

2.2 Influencing factors. The main factors are insulating material, cleanliness, humidity, atmospheric pressure, contact spacing, creepage distance and the duration of the applied voltage.

2.2.1 Insulating material. Suitable engineering plastics must be selected to meet the intended withstand voltage. PES special engineering plastic with a breakdown voltage of 16 kV/mm, for example, meets the 1,500 V standard atmospheric requirement of GJB 598 YB series II products. Fluoroplastics (F4) offer higher dielectric withstand and insulation resistance than other materials and are widely used for RF coaxial connector insulators.

2.2.2 Cleanliness. Cleanliness inside and on the insulator strongly affects dielectric withstand. During supplementary screening of a circular connector rated at 1,500 V, one unit broke down between two contacts at only 400 V. Joint sectioning with the manufacturer showed the breakdown occurred at the adhesive interface between the upper and lower insulator mounting plates because of impurities in the adhesive.

2.2.3 Humidity. Increased humidity reduces dielectric withstand. The J36A rectangular connector, for example, specifies 1,000 V under normal conditions, falling to 500 V after 48 h of damp heat testing at 40 ± 2℃ and 93% ± 2%.

2.2.4 Low air pressure. At high altitude where air is thin, the insulator outgasses and contaminates the contacts, corona is more likely and withstand performance falls, which can short the circuit. Non-hermetic connectors used at altitude must therefore be derated: the Y27A circular connector specifies 1,300 V under normal conditions but only 200 V at 1.33 Pa.

2.2.5 Contact spacing. Miniaturisation and high density, particularly in rectangular and printed circuit connectors, push pitch down to 0.635 mm and even 0.3 mm, with the critical height dimension reduced to 1 to 1.5 mm. Surface mount technology is closely linked to miniaturisation, and insulating materials with higher withstand performance are needed to meet the smaller dimensions.

2.2.6 Creepage distance. This is the shortest distance measured along the insulator surface between contacts or between contact and shell. Short creepage readily causes surface discharge (flashover). Some connectors therefore mould the pin or socket holes in the insulator mounting plate with a stepped profile to increase creepage distance and improve resistance to surface discharge.

2.2.7 Duration. Connector technical requirements generally specify one minute at the rated voltage with no breakdown, flashover or discharge. To speed up final acceptance testing, however, many manufacturers raise the test voltage by 20% and shorten the duration to 5 or 10 seconds. There is no functional relationship between the two. From the mechanism of AC breakdown, breakdown is mainly caused by leakage, that is a leakage current above the specified value, or by thermal breakdown, where raising the voltage forces leakage; thermal breakdown is unrelated to duration. GJB 1217-91 specifies that the test voltage be held for one minute, and that in-house quality conformance testing may reduce this to a minimum of 5 seconds where permitted. In practice, products that passed at the factory under this rule have still been found to break down during 100% supplementary screening at the user because of defects inside the insulator. The likely cause is that shortening the duration to 5 seconds does not allow enough time to charge the insulator capacitance and raise leakage current above the specified value.

2.3 Discussion.

2.3.1 Measurement methods. To apply a high voltage between contacts or between contact and shell for one minute, a suitable test fixture is needed, the same type used for insulation resistance and interchangeable with it. For connectors with larger contact pitch, a two-step method can be used: first connect all contacts in even rows in parallel and all contacts in odd rows in parallel and measure between the two groups, then connect all contacts in parallel and measure between the group and ground. For a rectangular connector with contacts on an equilateral triangular grid, row pitch 2.8 mm, row spacing 2.5 mm and adjacent row pitch 2.87 mm, the two-step method measures 2.87 mm rather than the minimum 2.8 mm, but because the withstand voltage is around 1,000 V with a large margin, the 0.07 mm difference in wall thickness is negligible. The two-step method is economical but still unreliable, since it cannot screen out breakdown hazards caused by internal defects between contacts in the same row.

For high density and subminiature connectors, where the specified withstand voltage and the margin are both small, the insulator wall between adjacent contacts is very thin, and even a tiny bubble, porosity or impurity seriously affects dielectric withstand. A three-step method is therefore required: adding to the two steps above a third step in which all odd-numbered contacts in all rows are connected in parallel and all even-numbered contacts in all rows are connected in parallel, and the withstand voltage between the two groups is measured. For connectors with high reliability requirements, especially those with contact pitch of 1.5 mm or less and insulator wall thickness of 0.4 mm or less, the three-step method should be used so that the withstand voltage between every contact and all of its neighbours is measured.

2.3.2 Setting the leakage current. When performing dielectric withstand testing, the leakage current threshold must be set strictly according to the referenced test method. One rectangular connector specified leakage current not exceeding 1 mA, but the instrument was set too low at 0.5 mA, producing false breakdown alarms. Because large leakage currents harm the electrical parameters and physical properties of connectors and coaxial contacts, the maximum leakage current in testing should be limited to 5 mA. Most products specify no more than 1 mA, and some, such as GJB 101-86 small circular environment resistant quick disconnect connectors, allow up to 2 mA.

2.3.3 Test fixtures. Dielectric withstand and insulation resistance testing share the same fixtures, so that the specified voltage can be applied between all contacts and between contacts and shell for one minute while checking for discharge, flashover and breakdown. Many manufacturers, however, do not use such fixtures and instead touch the two probes of the instrument at random between points and between points and the shell. This method is unreliable and very prone to false and missed results.

2.3.4 Insulation resistance testing cannot replace dielectric withstand testing. Some believe that a connector with sufficiently high insulation resistance does not need a withstand test, that the high voltage involved is dangerous for operators and does no good to the connector, and therefore avoid the test. In fact insulation resistance is measured with DC voltage, whereas dielectric withstand uses AC, and the power of the insulation resistance source is far lower than that of the AC withstand source. A connector with high insulation resistance therefore does not necessarily withstand a higher AC voltage. Megohmmeters, even with test voltages of several kilovolts, have low output power, around 10 mA even with the terminals shorted, so improper use cannot cause a fatal shock, whereas AC withstand testing involves much higher power and demands attention to personal and equipment safety. Internal defects in a connector insulator can only be revealed with high power and high voltage. Insulation and withstand are not equivalent: a clean, dry insulator may have high insulation resistance yet fail the withstand test, while a dirty or damaged insulator may have low insulation resistance yet not break down at high voltage.

3. Contact resistance testing

3.1 Principle. Contact resistance testing determines the resistance produced when current flows through the contact surfaces of the contacts. When a large current passes through a high resistance contact, excessive energy dissipation can occur and the contact can overheat dangerously. Many applications require low and stable contact resistance so that the voltage drop across the contact does not affect circuit accuracy. In practice a contact resistance tester (milliohm meter) based on the four-terminal Kelvin bridge principle is used, with a dedicated fixture clamping the termination ends of the contact under test. The measured total contact resistance R consists of three parts:

R = Rc + Rf + Rp

where Rc is the constriction resistance, appearing because the conducting cross section contracts where current passes through the contact interface; Rf is the film resistance formed by surface films and other contaminants; and Rp is the conductor resistance, the ohmic resistance of the mating contacts and lead wires. Besides the milliohm meter, contact resistance can also be measured by the volt-ampere method or the ammeter-potentiometer method.

3.2 Influencing factors. Contact resistance is affected mainly by contact material, normal force, surface condition, applied voltage and current.

3.2.1 Contact material. Technical requirements specify different contact resistance limits for mating contacts of the same size made from different materials. GJB 101-86, for example, specifies that for 1 mm diameter mating contacts the contact resistance shall be no more than 5 mΩ for copper alloy and no more than 15 mΩ for iron alloy.

3.2.2 Normal force. The normal force of a contact is the force produced by the contacting surfaces and acting perpendicular to them. As normal force increases, the number and area of contact microspots increase and the spots pass from elastic to plastic deformation, so constriction resistance falls and contact resistance decreases. Normal force depends mainly on the geometry and material properties of the contact.

3.2.3 Surface condition. Surface films are of two kinds. The first is a loose film formed by mechanical deposition of dust, rosin and oil, which traps particulate matter in microscopic pits on the contact surface, reducing contact area and raising contact resistance unstably. The second is a contamination film formed by physical and chemical adsorption, mainly chemical adsorption on metal surfaces, which follows physical adsorption accompanied by electron transfer. Products with high reliability requirements, such as aerospace connectors, therefore need clean assembly environments, thorough cleaning processes and necessary sealing, while users need good storage and operating conditions.

3.2.4 Applied voltage. Above a threshold voltage the contact film breaks down and contact resistance falls sharply. Because heating also accelerates chemical reactions near the film and partly repairs it, the resistance becomes non-linear. Near the threshold voltage a small change in voltage drop can change the current twentyfold or more, causing large changes in contact resistance. Misunderstanding this non-linearity leads to errors in testing and in use.

3.2.5 Current. When current exceeds a certain value, Joule heating (I2R) at the microscopic contact points softens or melts the metal, affecting constriction resistance and reducing contact resistance.

3.3 Discussion.

3.3.1 Low level contact resistance testing. Because contact films can break down mechanically under high contact pressure or electrically under high voltage and current, some small connectors are designed with very low contact pressure and are used only at mV or mA levels where the film is not readily broken down, which may affect signal transmission. Test Methods for Electrical Connectors therefore specifies two methods: the low level contact resistance test and the contact resistance test. The low level test evaluates contact resistance under voltage and current that do not alter the physical contact surface or any non-conductive oxide film present; the open circuit test voltage does not exceed 20 mV and the test current is limited to 100 mA, a level sufficient to show contact interface performance under low level electrical excitation. The contact resistance test measures the resistance across a mated pair or between contact and gauge at a specified current much larger than the former, usually 1 A.

3.3.2 Single contact separation force. Maintaining a stable normal force is the key to reliable contact. Normal force is a direct indicator of contact pressure and clearly affects contact resistance, but it is very difficult to measure in the mated state, so the separation force of a single contact, measured as the contact moves from rest, is used to estimate it indirectly. The theoretical relationship is F = FN × μ, where FN is the normal force and μ the coefficient of friction. Because separation force is governed by both normal force and friction coefficient, a large separation force does not necessarily mean a large normal force or reliable contact. As contact manufacturing precision and plating quality have improved, keeping separation force at an appropriate level is enough to ensure reliable contact. In practice, a separation force that is too small can cause intermittent signal loss under vibration and shock. Assessing contact reliability by separation force is more effective than by contact resistance, because contact resistance rarely fails in practice while sockets with low separation force often still show acceptable contact resistance.

3.3.3 Passing contact resistance does not mean reliable contact. In many applications - automotive, motorcycle, rail, power machinery, automation instruments, aviation, aerospace and marine - connectors operate in dynamic vibration environments. Testing static contact resistance alone cannot guarantee reliable contact in such conditions, and connectors that pass the contact resistance test often still show momentary interruptions during vibration, shock and centrifugal testing. For high reliability connectors, many designers therefore ask for 100% dynamic vibration testing to assess contact reliability. A Japanese company has introduced a small benchtop electric vibration table for use with a continuity tester, successfully applied to contact reliability inspection of many commercial wiring harnesses.

4. Instruments for electrical testing: current use and development

The connector industry is growing strongly in the electronic and information age, with demand rising year on year. Harness assemblies in particular are now used very widely, from household appliances and communication equipment to computers and peripherals, aircraft, vehicles and military instruments. To check harness assemblies for open circuits (poor contact), short circuits (poor insulation) and mis-wiring, a power supply with indicator lamps was once connected in circuit to observe lamp brightness, or a multimeter was used to see whether circuit resistance tended towards zero or infinity. This method is primitive, slow and prone to false and missed results. It can only judge continuity and cannot assess whether insulation resistance and dielectric withstand meet safety requirements, so for important connectors the technical requirements call for insulation resistance testers and withstand voltage testers.

Because such instruments have only two input terminals, checking insulation resistance and dielectric withstand between all contacts and between contacts and shell requires changing instruments and swapping two or three test fixtures for every parameter, which is slow and inefficient. Some domestic instrument makers have introduced combined insulation and withstand testers, but the basic procedure is unchanged: withstand testing first, then insulation resistance, on one instrument.

More recently, new instruments dedicated to the electrical testing of connectors and harnesses have appeared, including the EE30 continuity tester, the NM-10A intermittent contact detector and the NM-30X multifunctional automatic tester. These use internal automatic logic switching and memory circuits and offer the following features:

(1) Fast, accurate automatic testing of continuity, insulation, withstand voltage and intermittent contact after a single mating, replacing the traditional approach of repeatedly changing single-parameter instruments and two or three fixtures.

(2) Self-checking before testing to confirm the instrument is working correctly.

(3) Comparison of the connector or harness under test with stored reference data to judge pass or fail.

(4) Automatic printout of results for record keeping.

(5) LCD displays, red and green indicators and voice prompts on many models.

Such instruments suit on-line inspection at connector and harness manufacturing plants and at electrical assembly stations for important military products in aviation and aerospace. Although they are still expensive and some do not yet cover the full range of parameters, they indicate the direction in which instrument application is moving. Domestic manufacturers should introduce, absorb and localise this technology so that such instruments are used more widely in connector and harness production and service.