Electrical safety tests can be roughly divided into two areas: those tests carried out during the approvals process by test houses (known as type tests) and those carried out at the end of each production line by the manufacturer (known as routine production tests).

For type tests, a product is subjected to tests and evaluations in accordance with a specific product safety standard. For production tests, a manufacturer can select a few tests, ensuring that each product is subjected to those tests in accordance with its own procedures. Most manufacturers, in particular information technology (IT) equipment manufacturers, choose four primary product safety tests to be routine at the end of the production line. These include dielectric, insulation resistance, ground continuity, and leakage current.

These tests are designed to ensure that the user does not get electrocuted or otherwise hurt by operating a piece of equipment that has hazardous voltages or high fault current as a result of electrical fault. This article looks at the fundamentals behind each test and analyzes the reasoning behind each test, as well as discussing appropriate limits and equipment.

The dielectric strength test, also known as dielectric withstand test or hipot test, is probably the best known and most often performed production-line safety test. It is part of virtually every standard, which indicates its importance. The hipot test is a nondestructive test that determines the adequacy of electrical insulation for the normally occurring overvoltage transient. This is a high-voltage test that is applied to all devices for a specific time in order to ensure that the insulation is not marginal. Another reason for conducting the hipot test is that it also detects possible defects such as inadequate creepage and clearance distances introduced during the manufacturing process.

During type testing, the hipot test is applied after tests such as fault condition, humidity, and vibration to determine whether any degradation has taken place. The production-line hipot test, however, is a test of the manufacturing process to determine whether the construction of a production unit is about the same as the construction of the unit that was subjected to type testing. Some of the process failures that can be detected by a production-line hipot test include, for example, a transformer wound in such a way that creepage and clearance have been reduced. Such a failure could result from a new operator in the winding department. Other examples include identifying a pinhole defect in insulation or finding an enlarged solder footprint.

Most safety standards use the 2U + 1000 V formula as the basis for testing basic insulation, where U is the operating rms voltage. Although this formula is a guideline, each standard—in particular IEC 60950—refers the user to a specific table in the standard showing the exact test voltage based on the working voltage measurements.¹ The reason for using 1000 V as part of the basic formula is that the insulation in any product can be subjected to normal day-to-day transient overvoltages. Experiments and research have shown that these overvoltages can be as high as 1000 V.

Test method. Normally the high voltage is applied between the two parts across the insulation being tested, such as the primary circuit and metal enclosure of the equipment under test (EUT). If the insulation between the two is adequate, then the application of a large voltage difference between the two conductors separated by the insulator would result in the flow of a very small current. Although this small current is acceptable, no breakdown of either the air insulation or the solid insulation should take place. Therefore, the current of interest is the current that is the result of a partial discharge or breakdown, rather than the current due to capacitive coupling.

Another example would be to test the insulation between the primary and secondary circuits of a power supply. Here, all the outputs are shorted together. The ground probe from the hipot tester is placed in contact with this cable bunch, and the high-voltage probe is placed in contact with L and N connectors, which are shorted together (see Figure 1). The EUT does not run during the hipot test. It must also be noted that when applying the high voltage during the type test, the ideal situation would require that not more than half of the prescribed voltage be applied, and then raised gradually over a period of 10 seconds to the full value and maintained for 1 minute. Most test equipment, however, either turns on directly to the full voltage or has an electronically controlled ramp.

Figure 1. Typical hipot test setup.

Test Duration. If the test is part of an agency certification process, then the test duration must be in accordance with the safety standard being used. For instance, the test time for most standards, including products covered under IEC 60950, is 1 minute. However, when testing products in the production line, it is normally not practical to hipot test each product for 1 minute, and manufacturers normally conduct the test to a much shorter time, such as a few seconds, but with higher voltages. A typical rule of thumb is 110–120% of 2U + 1000 V for 1–2 seconds. The duration and procedure should be in agreement with any test houses concerned. It should be noted that although the reduced time and increased voltage are approximate, experiments and the manufacturers' data sheets indicate that each insulating material has its own specific voltage-time characteristics.

Current Setting. Most modern hipot testers allow the user to set the current limit. However, if the actual leakage current of the product is known, then the hipot test current can be predicted. Choosing the trip level really depends on the product being tested. The best way to identify the trip level is to test some product samples and establish an average hipot current. Once this has been achieved, then the leakage current trip level should be set to a slightly higher value than the average figure. Another method of establishing the current trip level would be to use the following mathematical formula:

Test Voltage. The majority of safety standards allow the use of either ac or dc voltage for a hipot test. When using ac test voltage, the insulation in question is being stressed most when the voltage is at its peak, i.e., either at the positive or negative peak of the sine wave. Therefore, if one decides to use dc test voltage, one must ensure that the dc test voltage is Ö2 (or 1.414) times the ac test voltage, so the value of the dc voltage is equal to the ac voltage peaks. For example, for a 1500-V-ac voltage, the equivalent dc voltage to produce the same amount of stress on the insulation would be 1500 x 1.414 or 2121 V dc.

One of the advantages of using a dc test voltage is that the leakage current trip can be set to a much lower value than that of an ac test voltage. This would allow a manufacturer to filter those products that have marginal insulation, which would have been passed by an ac tester. It must be noted that when using a dc hipot tester, the capacitors in the circuit could be highly charged and, therefore, a safe-discharge device or setup is needed. However, it is a good practice to always ensure that a product is discharged, regardless of the test voltage or its nature, before it is handled.

Another advantage of a dc hipot tester is that it applies the voltage gradually. By monitoring the current flow as voltages increase, an operator can detect a potential insulation breakdown before it occurs. A minor disadvantage of the dc hipot tester is that because dc test voltages are more difficult to generate, the cost of a dc tester may be slightly higher than that of an ac tester.

One of the advantages of an ac hipot test is that it can check both voltage polarities, whereas a dc test charges the insulation in only one polarity. This may become a concern for products that actually use ac voltage for their normal operation. The test setup and procedures are identical for both ac and dc hipot tests.

A minor disadvantage of the ac hipot tester is that if the circuit under test has large values of Y capacitors, then, depending on the current trip setting of the hipot tester, the ac tester could indicate a failure. Most safety standards allow the user to disconnect the Y capacitors prior to testing or, alternatively, to use a dc hipot tester. The dc hipot tester would not indicate the failure of a unit even with high Y capacitors because the Y capacitors see the voltage but don't pass any current.

The insulation resistance test is also known as a Megger test. Its objective is to measure the total resistance between any two points separated by insulation. The test, therefore, determines how effective the insulation is in resisting the flow of electrical current. The voltage is typically around 500–1000 V dc; hence, the current is very low. Because the current is so low, this test is useful for checking the quality of the insulation not only when a product is first manufactured, but also over time as the product is used.

Test Procedure. The EUT is connected to the measuring instrument, and the voltage is ramped up from zero to the final value, which in most cases is 500 V dc. Once the voltage reaches the selected value, it is kept at that value for a brief period (typically up to 5 seconds) before the resistance test is measured. The measured value should be very high (typically in the megohm region). The insulation resistance test is mandatory in some product safety standards, including IEC 60065 and UL 6500.^2,3

Also known as the ground bond continuity test, the ground bond test must be conducted on all Class I products. The purpose of the test is to ensure that all accessible conductive parts of the product that could become live in the event of a single insulation fault are connected securely to the final earth point of the supply input. In other words, a ground bond test verifies integrity of the ground path by applying a high-current, low-voltage source to the ground-path circuit.

Compliance is checked by measuring the resistance of the connection between the protective earthing terminal or earthing contact and the parts to be earthed to ensure that resistance does not exceed certain values when subjected to a high current as specified in various product safety standards. It is important to bear in mind that from the constructional and design points of view, the protective earthing conductors should not contain switches or fuses.

Test Requirements. Most safety standards require the following parameters for conducting the ground bond test:

Understanding Resistance Values. With the exception of the Canadian standard, most standards require 25 A for 1 minute. The value of 25 A for 1 minute represents the worst current and the longest operation time of a mains overcurrent device. The maximum 25 A is approximately 1.5 times the mains circuit breaker value installed for most pluggable type A cord-connected equipment rated up to 16 A. The Canadian National Wiring Code requirements are very similar to these in the sense that they assume that fuses are expected to operate no more than 1 minute at twice their rated current. Because most mains circuits are protected with a 15–16 A fuse, the fault current would be 30 A for no more than 2 minutes.

Some standards, including IEC 60950, 3rd ed., have named the leakage current test "touch current." This refers to the electric current through a human body or through an animal body when it touches one or more accessible parts of installation or equipment. There is also another concept known as "protective conductor current," and this refers to the current that flows in a protective conductor. A protective conductor current, therefore, can never be the source of an electric shock because, by definition, the protective conductor is connected to earth.

If touch current is excessive, an operator could receive an electric shock, which could result in a serious injury, depending on a person's body weight. Typically, currents of more than 1.0 mA can cause an electric shock to an operator. The shock may or may not be serious, depending on the amount of the current and the body weight.

Like the other tests, the leakage current test is also a very important safety test, and most safety standards require the test to be conducted under various conditions such as normal operating condition, switches open as well as closed, reversed line polarity, and so on. The measured earth leakage current must not exceed specified limits during any of the tested conditions. Table I shows some typical limits.

One of the biggest contributors to leakage is the capacitance between ac lines and earth, i.e., the Y capacitors. These Y capacitors are normally placed in the circuit to control electromagnetic interference. It should be noted that some standards, such as IEC 60950, allow higher earth leakage current than 3.5 mA only for Class I stationary equipment that is either permanently connected equipment or that is pluggable equipment type B, provided certain conditions are met. These conditions are listed in Clause 5.1.7 of IEC 60950.

Equipment. When measuring the earth leakage current of IT equipment, it must be noted that the measuring instruments should satisfy the requirements of Annex D of IEC 60950, which simulates the worse-case impedance of the human body. The use of an isolating transformer during the test is also highly recommended (see Figure 2).

Figure 2. Typical earth leakage current setup.

Any capacitive leakage in the transformer must then be taken into account. If for any reason the use of an isolating transformer is not possible, then the EUT must be mounted on an insulating stand, and appropriate safety precautions must be taken. Such measures compensate for the possibility that the body of the EUT may carry a hazardous voltage.

Test Method. Most standards—in particular IEC 60950—require that the EUT be energized for this test. The input voltage applied to the EUT is typically adjusted to 110% of the highest rated mains voltage and the highest rated line frequency. As mentioned previously, for safety reasons, it is highly recommended that the EUT be powered via an isolating transformer.

These tests are performed on both Class I and Class II products. For Class II equipment, the test is made to accessible conductive parts and to a metal foil with dimensions of approximately 10 x 20 cm, which is attached to the enclosure of the product. The metal foil simulates a human hand contact.

This test should also be conducted in all possible combinations such as normal operating condition, switches open as well as closed, reversed line polarity, etc. Equipment designed for multiple power sources, only one of which is required at a time (e.g., for backup), must be tested with only one source connected. Although most standards do not require the earth leakage current test to be carried out for 100% of the units in a production line, some standards, such as those for medical products, do require a 100% production-line test.

Because any electrical safety test involves some risk of electrical shock, it is crucial that certain precautions be taken to avoid shock and injury to operators. Listed below is a sampling of precautions that can minimize the danger of electrical shocks and ensure all-around safety:

CENELEC has released a standard, EN 50116, which essentially defines the routine electrical safety tests and their procedures to be applied during or after the manufacturing process of IT equipment certified or declared as complying with EN 60950.^6,7

It is extremely important for manufacturers of electrical products to ensure that all test results, including routine production tests, are clearly and adequately documented and kept on file for possible inspection. Although this may not be required by all agencies, keeping accurate test records is not only good engineering practice, but could also be vital in defending a legal action should such a case arise. Most modern test equipment can produce test results in an electronic format that can be stored or printed when needed.

It is also necessary for manufacturers to ensure that all test equipment is calibrated regularly and that such records are kept on file. In fact, most test houses expect to see a log for daily calibration of the hipot tester on the production line. This log confirms that the hipot tester has indeed been applying high voltage to the EUT. It is also vital to ensure that all electrical safety tests are carried out on units that have been returned for repair or service.

On the production line, there are generally three product states: not tested, tested-passed, and tested-failed. The status of any product must be clearly apparent to ensure that untested products are not shipped and that only tested-passed products are shipped. For instance, a red tag attached to the unit can indicate tested-failed. Untested products are those that have neither a tested-passed nor a tested-failed indicator.

Electrical safety testing at the design and development stages as well as at the production stage is vital to ensure that all products are safe before reaching the end-user. The four tests described in this article are among the most fundamental tests that manufacturers of electrical and electronic products should conduct routinely. It is important to use correct test equipment and adequate and accurate test procedures so that sufficient testing is conducted and operator safety is considered.

The author wishes to thank Rich Nute of Hewlett-Packard for his valuable help.