The consequences and costs of a failure at such a late stage can, of course, be significant not only in terms of the time and money that will be spent investigating the cause of failure but also measured as potential disruption of the production process, loss of output, and impact on profitability.

Immediate production problems and the sacrifice of profit margin to redesign expenses are important matters, and bad enough. But sometimes device failure emerges over a longer term. Identifying the reasons for failure too late—after product release—can often prove to be more serious and costly.

For example, a live conductor with insufficient insulation might nevertheless pass a final test if its isolation due to clearance is sufficient to allow it to pass the high-voltage stress test at final quality control. However, this design or assembly defect could, during use by the customer, present a potentially lethal problem that may become apparent only after years of service.

The confusion between a protective ground (earth) and functional grounding is a source of errors in design and assembly that might only come to light if there is testing during the construction process. Such problems can go undetected at the final test stage and subsequently present a long-term safety hazard in the field.

Problems like these can be identified and eliminated at an early project stage through the performance of testing at strategic steps in the assembly process. A well-planned and well-managed safety testing policy implemented throughout the manufacturing process need not involve undue complication and expense. And there can be no better plan than to ensure that safety and quality are inherent in the product during the design and assembly stages.

This article discusses the four fundamental electrical safety tests that designers of electrical products should employ during the product-building process in order to design safety into any device. It relates specifically to products covered by the LVD, those designed for use at a rated voltage of 50–1000 V ac and 75–1500 V dc and intended to be marketed in the EU. The focus should not obscure the fact that all the tests named in the relevant harmonized standards are important unless they are found to be inapplicable. Nor should it be interpreted to slight requirements to determine the levels of other risks presented by an LVD product, including mechanical risks and the danger of fire, explosion, burns, and other potential causes of death or injury.

Inadequate Clearance Distances for Connectors

Clearance distances within device assemblies may have been reduced as a consequence of postdesign changes adopted for reasons unrelated to product safety, for example, to streamline a production process. Unsatisfactory clearance may be attributable to a different placement of components or to the unavailability at build of the style of component envisioned during product design and substitution with a physically different part. A placement-related problem is illustrated in Figure 1a. In some cases, a problem like this might be identified only by safety testing performed during the production stage. Another possibility is that an inappropriate fitting of connectors has produced a clearance distance that is unacceptably small. An example of this sort of problem appears in Figure 1b. Diagram 1 shows what could happen if individual connectors were used to fit connection pins in close proximity. An adequate clearance distance (X) at the point of connection is reduced to a dimension (Y) below the required minimum at the point where the incorrectly applied, unrestrained individual connectors approach each other. In a situation like this, a successful test result would depend heavily on stringent quality control. Diagram 2 shows the effect of fixing the connectors in alignment by attaching them to an insulator. The adequate dimension at X now remains the clearance distance for the assembly at Y. This illustration makes clear the advantage of using preformed plugs and sockets.

Figure 1. Inadequate clearance distances.

Meeting the Standard

From the outset of the design process for any new electrical product, the principal objective of the designer, with respect to the product's electrical safety, must be to ensure that the correct level of protection against electric shock is incorporated into the design. This involves considering protection for both the person who has to build the product and the eventual end-user. The environment in which the product is to be used, the users' technical ability, the materials from which the product is fabricated, and the technical specifications required by relevant standards, directives, and legislation are all factors that need to be taken into account. A similar process must be followed for mechancial and other concerns.

At least one relevant standard applies to nearly every electrical product, whether it be issued at a national, European, or international level. Standards offer not only guidance regarding the design and construction of specific types of products but also advise about the electrical safety tests to be performed, as well as other tests such as those for temperature rise and short circuit and overload. Care should be taken that tests required during the design stages not be confused with those used in the manufacturing process, as they do differ.

The central precompliance electrical safety tests that electrical product designers should consider are ground testing (earth bonding), dielectric strength (flash or hipot) testing, insulation resistance testing, and ground (earth) leakage testing. The test procedures to be applied are usually specified in the particular standard relevant to the product under design and to the type of protection it provides under its classification as Class I or Class II equipment.

Class I equipment or products are defined as those having protection against electric shock not solely reliant on basic insulation, but rather having all external conductive parts referenced to the ground of the fixed wiring installation. In the event of breakdown of the insulation, the user is protected by the grounding of the equipment or product. Class II equipment or products incorporate two layers of insulation that provide the user with sufficient protection from electric shock, and they do not rely on a ground reference.

Designers of electrical products can conform to LVD requirements by referring to a number of harmonized European standards for guidance and for references to design principles and electrical safety testing values. These product safety standards include:

The ground bond test performs a measurement of resistance to determine if the equipment would remain safe in the event that internal live conductors touch the case. The test ensures that sufficient current will flow through ground to blow the protective fuse in the plugtop or circuit breaker. It uses a high current from a low-voltage test to make sure that ground connections are correctly made. A failure indicates loose or poor ground paths. The ground bond test applies to Class I equipment only; it cannot be used on Class II equipment in which protection is supplied by double insulation.

This test is designed to ensure the safety of the product. Thus, it is seen as the first test to carry out, having priority over such tests as those for dielectric strength or insulation resistance. Both recent legislation such as the LVD and issues involving product liability have made it more necessary for manufacturers to illustrate due diligence. In addition, the trend toward using ground bond techniques to deal with issues of electromagnetic compatibility, particularly conducted emissions, means that the quality of the ground path becomes a matter of pertinence to product performance as well.

To perform the ground bond test, a low voltage of less than 12 V (for operator safety) is introduced to the input ground connection, which often includes the supply cable. A measurement is then taken using a clip or probe connected to each separate external metal surface. The test is performed without the equipment connected to line voltage (a mains supply). Rather, the connection is made with the power input lead to be used with the equipment. Test current varies between standards, although generally 25 A ac is used. Exceptional examples include lighting fixtures, which require a test at 10 A ac, and machinery under the the Machinery Directive standard EN 60204, which requires a constant current of 10 A ac, after which the voltage drop is measured. The permissible drop is determined by the cross-sectional area of the grounding conductor.

The pass-fail limit for the ground bond test measurement of resistance also varies between standards. In general, a test at 25 A requires a resistance of 0.1 W. Lighting equipment tested to EN 60598-1 using 10 A has a pass-fail threshold of 0.5 W.

Ground bond tests should not be confused with ground leakage. Ground bond ensures a sound connection of the metal case to the power-line (mains) ground reference, whereas ground leakage monitors the leakage from the ground to live and neutral conductors at normal supply voltage (230 V). Thus, good ground bonding is a prerequisite to dealing with the presence of ground leakage.

Sensitive electronic equipment calls for special test considerations. A common design feature of personal computers, for example, is that they have secondary, or peripheral, grounding points after the primary grounding position, which can use PCB tracks. Care has to be taken not to apply the high test current to these points, as damage may be caused. The only way to avoid this problem is through visual identification of the primary ground. This requires a combination of training and the foresight to label or indicate the correct test point clearly.

Some equipment has grounding points that are difficult to access. For example, the element in an electric kettle is the only grounded component. Other devices have grounding points that are hard to access because of plastic shrouds. These difficulties sometimes must be addressed during the product design stage so that proper consideration is given to the positioning of ground test points and so that the test process requirements of production testing are taken into account.

To accommodate all possible design-related causes of failure, the ground bond test should include the power input lead. Since the ground bond measurement of resistance is dependent on the length of the conductor as well as its cross-sectional area, an unnecessarily long power supply lead may give a failure reading. Leads of 2–3 m will not cause any problems for achieving the normal pass-fail criteria, but some products, such as floor cleaners, that require much longer supply leads owing to the nature of their use will require careful assessment of their test results. Thus, attention should be paid to the length of lead supplied.

It is also important to consider how manufacturing tolerances in production may invalidate assumptions made for the product as designed. For example, joints that rely on compression may not seal as completely as anticipated during the design phase, and hinges may not provide the degree of continuity that was originally expected. Faults like these can result in product assemblies that have no effective reference and that therefore present a potential electric shock hazard and noncompliance with requirements of the LVD.

The Safety Ground Terminal

The safety ground terminal on Class I electrical equipment must be a dedicated terminal not used for any other purpose. Were it to be used for fixing components such as transformers or circuit boards to the case, and if such components were disturbed during maintenance, modification, or other activity, the ground integrity of the case could be inadvertently and unknowingly compromised. Testing and inspection of this important terminal during the production process can expose any problems at an early stage. A suggested method of construction is presented in Figure 2.

Figure 2. Safety ground terminal.

Dielectric Strength Testing

Dielectric strength testing does not in fact involve a measurement but rather is a procedure that aims to illustrate whether a product remains safe when subjected to a high voltage. The test is designed to determine if gaps or clearances between conductive parts and a ground are sufficient and if degradation, such as pinholes and cracks in insulation or other protective devices, has occurred as a result of production processes or other sources of wear and tear. High voltage is applied to the product to test the insulation between live conductors and all exposed metal surfaces. For Class I equipment, the high voltage is applied between the conductors and a ground. For Class II equipment, the application point is between the conductors and the outer surface of the product.

Voltage levels applied vary by product type in accordance with the relevant product-specific standards, but Class I equipment items normally require between 1000 and 1500 V applied for 1 minute with a trip level of 0–100 mA, depending on the relevant product standard. Voltages for Class II equipment are usually higher, between 2500 and 4200 V, but the time and trip settings are similar.

Production-line dielectric testing calls for special considerations because certain practicalities affect the voltage levels and duration of application. These tests need to be faster and yet be equally rigorous. In general, a 10% overvoltage is applied during a test time of a few seconds. Thus, a type test with a voltage rating of 1250 V would be carried out on the production line at 1375 V with a trip level of 5 mA.

Note that it is possible to get an apparently satisfactory test result when in fact the equipment is switched off or not properly connected. Obviously, the equipment under test must be checked to ensure that it is switched on and connected appropriately. This circumstance can be a stumbling block even for experienced test operators.

In Class II equipment, the absence of a ground makes protection via primary and secondary insulation necessary. A common test problem, particularly with new equipment, is that a failure occurring on the primary insulation is undetectable by a dielectric test on the outer surface, which checks the secondary insulation only. All testing regimes should include tests of both these levels.

To test the primary level of protection, a method of gaining access to the primary insulation must be available. This is somewhat paradoxical, since this connection needs to be an inaccessible conductor. However, experience has shown several options to be feasible. First, the primary insulation can be tested prior to the final assembly of the device. The important concomitant here is that checking should be done upon assembly to determine that no degradation of this level of protection has taken place, such as screws penetrating the insulation. A second alternative is to design the product with an access point that can be permanently sealed after testing. Product designers do not often foresee this need. Another possibility is to design test jigs and probes that afford access through the enclosure while ensuring that the integrity of the product in terms of the relevant standard finger tests is maintained.

Testing of the secondary protection also needs to be carried out. Standards generally specify that the device under test be wrapped in aluminum foil so that high voltages can be applied to all the outer surfaces. Such a test may be practical for the laboratory environment, but it is impractical in production scenarios, both because of the complexity of test setup and the time required and because the outer surfaces of the product can easily be marked. The use of conductive foam in a special jig to create a nest or envelope around the product enables test voltages to be applied to the subject without marring its finish. Although not quoted in standards, this methodology devised by Clare Instruments (Worthing, W Sussex, UK) is recommended by standard authorities.

Concerns have also been raised that ac dielectric testing could corrupt sensitive electronic components. A possible solution is to use a dc dielectric test. The voltage applied needs to match the specified peak ac voltage; this is calculated by multiplying the specified ac voltage by 1.414. A discharge facility to ensure that no residual voltage remains following application is required.

Another available dc dielectric test, known as the soft dc test, involves ramping up to the required voltage. In some instances, this test can benefit from ramping down as well. The complete test would consist of a slow ramp from zero up to the required value, which is then held for a timed period before ramping back down to zero and discharging the unit under test.

Correcting Ground Bond Test Failures

Because ground bond test results should be very low measurements of impedance, an unsatisfactory test outcome is not uncommon. However, satisfactory retests are relatively simple to achieve if the fault is detected during the assembly phase of production. Following are three situations that can result in ground bond test failure and the solutions for converting a fail to a pass. Problem: The test probes are not making a good contact with the product's conductive metalwork and the meter terminals because of a decorative or protective coating. Although such coatings as paint or anodizing must not be considered as insulation for purposes of construction, they will present a high impedance to the test current if they are not penetrated by the probes for the test process. Solution: Make the connections again with penetration and retest. Problem: Parts of the equipment case have been assembled after a decorative or protective coating has been applied. In such a situation, ground continuity will be fortuitous; that is, it will only occur where, by chance, the edge of a fixing screw happens to touch the edge of the cover fixing hole as illustrated in Figure 3. Solution: Change the surface-finishing process by masking the required ground continuity arrangements before the coating is applied, use a penetrative star-type washer for intentional grounding, or reorder the sequence of manufacturing processes. Problem: The ground conductor between the equipment ground input pin and the safety ground terminal does not have at least the same cross-sectional area as the circuit-protective conductor within a correctly rated flexible cord. Solution: Replace the conductor with one of adequate size.

Figure 3. Ground bond test failure.

Insulation Resistance Testing

The insulation test acquires a measurement in megohms of the resistance of a product's insulation protection by applying a dc voltage between the phase (live) and neutral conductors to the ground conductor for Class I equipment and between the phase and neutral conductors to the outer case for Class II equipment. The test is designed to ensure that protective insulation forms a barrier sufficient to prevent electricity from coming into contact with a user so as to cause injury, or to prevent other systems in the vicinity from being adversely affected. Legislation such as the LVD requires evidence of due diligence in the production environment. Results of the insulation and other tests can be used to fulfill this requirement.

The insulation resistance test can be applied to both Class I and Class II equipment. It is carried out with the use of probes or insulated clips, without the equipment being connected to a power supply. Test voltages vary between standards; the most common application called for is 500 V. The voltage is applied for 2 to 3 seconds. In general, Class I equipment must display a resistance greater than 2 MW to pass the test. Resistance greater than 7 MW is required for Class II equipment. Both classes call for a 5 or 10% accuracy, with specific exceptions.

Equipment has been designed to supply up to 1000 V for testing such devices as automobile ignitions that call for greater levels of protection. Equipment is also available with voltages as low as 100 V to test switches and other devices that might be damaged by higher voltages.

Insulation resistance testing is designed to provide a quantitative representation of the quality of insulation. If a wire was positioned 0.5 mm from exposed metal, an insulation test—conducted in dry air—could well provide a pass reading. A dielectric test would be more likely to reveal the danger in this situation. Similarly, if insulation is somehow contaminated, a dielectric strength test would result in a pass but an insulation test would highlight the deficiency. For example, the normal minimum insulation resistance value for Class I appliances is 2 MW; with a 1500-V dielectric test, the current would be 0.75 mA and would not be detected by the 5-mA trip that has to accommodate the capacitive losses that occur. Obviously, a dc dielectric test with a leakage meter can provide insulation resistance monitoring, as the capacitive component is overlooked after the initial inrush.

The ground leakage test is the most misunderstood safety test. It was devised to provide some measure of the quality of the insulation resistance at line voltage potential when it is inappropriate or impossible to perform dielectric strength or insulation resistance tests.

In any domestic or industrial power supply, a substantial current normally flows from the live conductor through the various loads—motors, heating elements, and so on—and returns to the neutral conductor. This current then flows, via the neutral conductor, to the local substation where it is bonded to true earth. This is literally the "ground," or planet Earth. To provide protection and screening from the live and neutral conductors, a third conductor, ground, is included.

In a perfect world where conductors had no resistance and insulators had infinite resistance, the current drawn by an appliance would remain in the live-to-neutral loop, and no current would flow in the ground conductor. However, in the real world, the current in the live conductor can leak, via insulation, to any other conductor. The neutral conductor can also become elevated to 40 V above true earth, so current can also leak from this conductor. The net result is that these leakage currents flow back to true earth via the ground conductor—hence the term ground leakage.

If the ground conductor in a Class I appliance were to become disconnected, then this leakage current, or part of it, would flow via the lowest-impedance path to true earth. Unfortunately, this could be the person holding or touching the appliance if that individual were in direct contact with the ground.

Class II appliances have no grounding wire; their leakage current would normally flow to true earth via the neutral conductor. In this case, if the neutral conductor were to become disconnected, the leakage current would again take the lowest-impedance path to true earth, which again could be a person in contact with both the appliance and the ground.

Measurement of ground leakage in an electrical appliance can be performed by two distinct methods, both of which are conducted with the product powered up. Care must be taken that no hazard to operators can arise from moving parts or heating elements.

The first test method involves the insertion of a sensitive current-measuring device directly between the ground wire of the appliance and the supply ground. A reading of the current flow in the ground path is taken with the appliance turned on. The power connections are then reversed and a second reading taken, which takes into account a possible polarity reversal in manufacture or installation. In some instances, the impedance of the measuring circuit is designed to replicate the impedance of the human body.

For Class II appliances in which there is no ground wire, the measurement is taken between the outer surface of the appliance, wrapped in conductive foil, and true earth. This particular method can be hazardous and must only be undertaken by skilled personnel under strictly controlled—usually laboratory—conditions.

The second method uses current transformers and Hall-effect devices, which do not require the disconnection of the ground conductor. Ground leakage current in Class I appliances can be directly monitored via this method. Because this method is less hazardous, it is often preferred in the production environment.

Class II appliances might have the measurement taken as with the first method or a different way, specific to this second method. The alternative involves taking a differential measurement: the difference between the measured current flows in the line and neutral wires is taken to be the ground leakage (i.e., line current – neutral current = ground current).

Permissible levels of ground leakage current vary among product types and related standards. The selection of values presented in Table I gives an indication of the range.

The precompliance tests described above, performed with in-house instruments by operators knowledgeable about the role and limitations of each procedure, are invaluable tools to apply during the design and assembly phases of electrical product manufacture. Used in a timely manner, they can help ensure that user safety is an integral component of the final product design.

Clearly, the extent and severity of testing will vary from product to product. Test conditions may necessitate modifications of the working environment for the protection of assembly or test personnel. (EN 50191 is a good guide for the erection and operation of test equipment both within and outside the EU.)

There can be no doubt that an investment in test instrumentation and in the time required for effective electrical safety testing to be carried out during the product-building process will be repaid many times over if the potential for a product to be dangerous or susceptible to damage is eliminated and if product recalls can be avoided.