The Capacitive Discharge Test

Hidden Tolerance Issues Make This Test Difficult to Conduct

The capacitive discharge test has been around for a while. As most of us know, its purpose is to set a limit for the voltage present on a plug face when disconnected from mains power. The method of conducting this test has been defined and labs are conducting this test as a routine matter. However, because of limitations in the measuring equipment used, and the particular requirements of conducting this test, results may be in strict tolerance, but still lack accuracy.

This article will define the test requirements, and discuss the methods used to obtain results. Limitations of the equipment used will also be discussed, so that lab personnel can be made aware and do their best to minimize tolerance errors.

A Brief Look at the Requirements of the Capacitive Discharge Tests

The capacitive discharge test appears in different iterations in the following standards:

• IEC 61010-1, Laboratory Equipment
• IEC 60065, A/V Equipment, (dow 2019)
• IEC 60335-1, Household Equipment
• IEC 60601-1, Medical Equipment
• IEC 60950, Information Technology Equipment, (dow 2019)
• IEC 62368-1, A/V, Information and Communication Technology Equipment

Briefly stated below, requirements vary between these standards. There are more differences in conducting the tests, and in some cases there are supplemental circuit evaluations to conduct to determine results. But the following information is presented to show the different pass/fail voltage levels defined per standard:

• IEC 61010-1: (For pins) <60V after 5 seconds.
• IEC 61010-1: (Terminals for external conductors) <60V after 10 seconds.
• IEC 60065: <60V after 2 seconds.
• IEC 60335: ≤34V after 1 second.
• IEC 60601-1: <60V after 1 second.
• IEC 60950: <37% of mains voltage after 1 second (Pluggable Type A).
• IEC 60950: <37% of mains voltage after 10 seconds (Pluggable Type B).
• IEC 62368-1: <60V after 2 seconds (Ordinary person)
• IEC 62368-1: <120V after 2 seconds (Instructed person or SFC)

Although the pass/fail points of these tests vary, test setups do not contradict each other. EN 60950 requires a 100MΩ input impedance with <25pF capacitance to minimize the influence of the test setup on the results. Other standards do not stipulate input networks, but CTL DSH 0716 requires compliance to the values in EN 60950 for plug discharge tests performed for any standard.

Voltage tolerance is not specified for any of these tests, so the tolerances of OD 5014 apply.

Conducting the Capacitive Discharge Test

In general, laboratories conduct the test using a three-pole switch, which mimics the removal of the plug cap from the mains socket. All three poles are disconnected at approximately the same time in this scenario. (Switches that disconnect only the line and neutral, but leave the ground lead connected will not give correct results.)

The output side of the three-pole switch is connected to the device under test (DUT), and to the 100MΩ, 25pF probe which is connected to a scope channel. The scope is set to read ~±100V/ div to view the entire mains voltage waveform at the maximum voltage generally used with pluggable equipment, 264V rms, which is ±374Vpeak, or 748V from negative peak to positive peak. While the DUT is connected to mains with the three switch contacts closed, the mains waveform is displayed on the scope, and the DUT is placed in the condition which results in the least power draw, usually a standby mode.

Then the three pole switch is opened. The scope waveform is evaluated to see if the disconnect happened within ±5% of the peak of the waveform. If not, the setup and switch disconnection are repeated until peak disconnection is seen. Then the waveform is analyzed to see if the voltage is below the standard’s pass/fail threshold at the specified time. The test is repeated the number of times required, and results are entered into the report.

OD 5014 Voltage Tolerance

The voltage tolerances controlling the capacitive discharge test are found in OD 5014, and for voltages < 1000V, up to 1kHz, the instrument accuracy of measuring range is ±1.5%. In our case, for the full scale range of 800V (assuming 100V/div is used to see the 264V mains waveform), the accuracy is ±12V. But, since the OD 5014 accuracy is specified as the accuracy of the range, every voltage measurement, no matter amplitude, can have a swing of ±12V and still be in accordance with OD 5014. In most testing, because of this fact, tests are conducted at the minimum range possible, so the error is acceptable, and approaches ±1.5% of the voltage reading. This best case of ±1.5% of the voltage reading will only be attained if the voltage being measured fills the range completely.

In the case of the capacitive discharge test, however, the voltage at the pass/fail threshold is small with respect to the range that is required to be used. Because a large range must be used to view the mains voltage waveform, the voltage of interest is a small fraction of the range, and therefore the error is large. Figure 2 graphs the allowable error under the requirement of OD 5014, and as the voltage read diminishes, the error rises.

Oscilloscope Error and Tolerance

Vertical accuracy of oscilloscopes is sometimes difficult to determine. The oscilloscope is a very versatile tool for waveform evaluation, but voltage accuracy is hampered by the 8-bit D/A converter used on nearly every scope sold. The 8-bit converter only allows 256 voltage levels over the range of the instrument, and in our case of an 800V range, the D/A converter’s maximum theoretical resolution is ±3.125V, which corresponds to an error of 1 bit. This can be seen in Figures 3 and 4, where Figure 3 is a representation of a 10kV waveform, and Figure 4 shows an excerpt of that waveform. The dithering of the D/A converter can be clearly seen and, in this case, the 1-bit dithers represent the smallest voltage that this particular scope can measure on this scale: 39 volts. This full-scale error is constant regardless of the voltage being measured, so at the point shown in Figure 4, the readings near zero will have large errors.

Other errors are present as well, and Table 1 shows amplitude accuracies of some representative higher-end state-of-the-art oscilloscopes. As can be seen from the examples, these scopes have tolerance that skirts the maximum tolerance allowed by OD 5014 and in some cases does not hold that tolerance. Two of these scopes are in compliance with the ±1.5% full scale tolerance required for voltage measurement by OD 5014. Scope 3 has a 12 bit D-A converter, so is very accurate when measuring voltages. Scope 2 has stated an acceptable ±1.5% tolerance, but it does not mention any offset errors.

Table 1: DC vertical gain accuracy of some oscilloscopes

The 12 bit A-D converter used in Scope 3 allows greater accuracy. Instead of 256 steps available with the other scopes’ 8-bit D-A converter, the 12-bit converter has 4096 steps to display the voltage. This scope would be well suited to conduct the capacitive discharge test.

Full Scale Tolerances and How They Apply to the Capacitive Discharge Test

In the preceding sections, we have examined the voltage tolerance requirements of OD 5014, at ±1.5% of range, or ±1.5% full scale, and looked at the accuracy of some oscilloscopes and determined that 8-bit scopes have accuracies approximately equivalent to the OD 5014 requirement. Let’s assume for the sake of argument that the oscilloscope accuracy is within the tolerances of OD 5014. Let’s also assume that the 100MΩ <25pF probe that will be needed for every scope can be inserted into the measurement with zero error. We will also assume that the stated DC vertical gain accuracy holds at mains frequencies.

Using these assumptions, it seems fair to assume that the actual measurement error of a typical oscilloscope used to judge this test will mimic the OD 5014 tolerance requirement. In Table 2, we have calculated the allowable error at the pass/fail points noted in the various standards for the capacitive discharge test, assuming the test is conducted at 264Vrms mains voltage; using the stated tolerance of ±1.5% of range. Acceptable measurement errors range from ±9.5% at 126V in EN 60950 to ±35% at 34V in EN 60335. Most capacitive discharge test setups using 8-bit oscilloscopes would have similar accuracy.

^Standards’ Specified Tolerance; Assumes 800V range

Table 2: OD 5014 errors with respect to the pass/fail points of the capacitive discharge tests

Conclusion

The capacitive discharge test, conducted to determine the voltage present on a plug cap immediately after disconnection from mains power, is relatively straightforward to conduct, but allowable voltage tolerances specified in OD 5014 are not adequate to provide an accurate result in this particular test.

In addition, after looking at tolerances of a sample of available oscilloscopes, it is clear that these instruments, while widely used to conduct this test, lack the DC vertical gain accuracy required to improve on the tolerances of OD 5014. In some instances, the oscilloscopes are not possessing adequate accuracy to be used to conduct the test in accordance with the CTL Guidelines. Additional inaccuracies, imposed by 100MΩ probes or other additions required to conduct this test will add tolerance errors and might push even a complying scope out of tolerance required by OD 5014.

The accuracies mentioned in Table 1 are representative, but do not cover all scopes. You should evaluate the DC vertical gain accuracy of the scope you are planning to use, add in tolerances for the required 100MΩ probe and scope offset, determine if the resulting accuracy applies to line frequencies, and thereby know if your scope is in accordance with OD 5014 before you use it to obtain results from the capacitive discharge test. Based on Table 1, especially adding in errors for offset and 100MΩ probes, it appears that a D-A converter of greater than 8 bits will be needed to meet the voltage accuracy requirements of OD 5014.

Also, we have demonstrated that the tolerance requirements of OD 5014 may not be sufficiently strict to provide reasonable results when conducting the capacitive discharge test, because the OD 5014 accuracy is based on full scale of the range used. Since the result of the capacitive discharge test is small with respect to the range needed to view the entire mains waveform, this small result can be inaccurate even when taken with an instrument in compliance with OD 5014 tolerances.

Users should bear these ideas in mind when specifying equipment to conduct this test. It would be best to use an instrument which uses a tolerance based on the instrument reading, instead of a full scale or range accuracy, to ensure that meaningful results are obtained from capacitive discharge tests. If this is not possible, testers should calculate the specified instrument error resulting at pass/fail levels to ensure that accurate results can be obtained.

Jeffrey D. Lind has over 38 years of electrical engineering expertise. He launched his career working at Underwriters Laboratories (UL) and then for Atari™ and Sega Gremlin™. In 1997, Lind started Compliance West. He received his Bachelors of Science in electronic engineering from Cal Poly, San Luis Obispo. Lind can be reached at jlind@ compwest.com.