What is partial discharge?

Partial discharges (hereinafter also referred to as PD) are localized electrical discharges which only partially bridge the insulation between conductors and which can occur adjacent to a conductor, but do not have to (according to standard IEC 60270). The partial discharges occur on high-voltage lines, in windings and other electrical equipment. PD can occur due to faults in windings caused by faulty enamelled wires, air pockets during casting, insulation faults or similar.

Fig. 1: Sliding or surface discharges during hipot test (AC)
Fig. 2: Partial discharges during surge test

How are partial discharges generated?

A high-voltage pulse is applied to the DUT and potentially occurring partial discharges are detected. The measurement of the partial discharge of winding against body (housing) can be done with a hipot tester. A surge tester is used to measure the partial discharge of winding against winding (and body).

What is measured?

A partial discharge generates high-frequency voltages and radio beams. Parallel to the measurement of the test voltage, the high-frequency voltage components can either be measured via a line or the radio radiation via an antenna. Both methods are offered by SPS electronic and are interference-free due to narrow-band and active measurement technology. They can therefore be used without restrictions in both laboratory and production environments.

Testing EOL

In the routine test, the test voltage is first determined. For the series testing of many DUT, testing is often only carried out at this determined voltage and with fewer voltage pulses. The reduction of steps and pulses is done to avoid too high loads on the final product. While in the past these tests were not very suitable for series production due to the long cycle time, this is now possible by using SPS electronic test technology. Due to the considerable reduction of the test time, EOL measurements on a large scale are also possible!

Measurement result

The measurement result is output logarithmically in millivolts (mV). A doubling of the partial discharge voltage from 50 mV to 100 mV corresponds to a tenfold increase in discharge energy (from 100 mV to 200 mV corresponds to a tenfold increase again). The background noise with antenna is (depending on the environment) typically 30 mV - 50 mV, with line coupling the background noise is typically 180 mV to 230 mV.  The threshold value for detection as partial discharge should be approx. 20 mV to 50 mV above the background noise (customer-dependent).

Why we don't measure in picocoloumb

The measurement of partial discharge in coulombs is useful for "twisted pairs of conductors", if the equivalent capacitance CS is considerably smaller than the detection capacitor Cd, and is only required in the IEC 60270 standard for PD. However, standards of practical relevance for winding materials (e.g. IEC/TS 60034-27-5 and IEC 61934) do not specify the measured value in Coulomb (or Picocoulomb), since the electrical charge cannot be measured in a meaningful way. Measuring the electric charge in Coulomb leads to inaccuracies and corresponding difficulties in comparison.

More meaningful than the charge energy is, in practice, the level of the inception- (PDIF) and extinction voltage (PDEV). The higher the voltage, the better the insulation.

Why an active measurement is preferable

The accuracy of the measurement is increased by active antennas (MW 40) and active line coupling (HW 40), the sensitivity to interference is reduced, the range of applications is increased and the line length to the test device may be longer.

Active antenna compared to active line coupling

The microwave antenna operates in the range of 1.57 GHz (GPS frequency), as no electronic devices may interfere in this frequency range. The active antenna processes the signal already in the antenna head, so that the accuracy is increased and the line length to the test system can be as long as required. The measurement with an antenna is only useful on open stators. On closed stators (assembled motors) the active line coupling is used. The measurement with line coupling is more sensitive, but has a higher background noise and only frequencies in the range up to 400 MHz can be evaluated.

Measuring method according to IEC 61934

Since partial discharges are not a 100% reproducible event, a measuring method in connection with surge pulses was defined in the IEC 61934 standard to determine the insulation quality of winding materials by partial discharge measurements. For this purpose not only one surge pulse is evaluated, but a whole series of surge pulses with different voltages.

The IEC 61934 standard proposes to work with 10 pulses per voltage. Initially, the test voltage is continuously increased. As soon as a partial discharge occurs for the first time, this voltage level is recorded as the partial discharge inception voltage (PDIV). The test voltage is then further increased until 50% of the surge pulses produce partial discharge. This voltage is called the repeated partial discharge inception voltage (RPDIV). Now the voltage level of the surge pulses is reduced again. If partial discharges are detected in less than 50% of the surge pulses, this value is recorded as the repeated partial discharge interruption voltage (RPDEV). As soon as no more partial discharges are registered at a voltage level, the partial discharge interruption voltage is reached (PDEV) and the test is terminated.

With these four characteristic values the insulation of the winding is clearly qualified and can be provided with tolerances for production. Testing to IEC 61934 is particularly important for winding materials that are exposed to steep switching flanks during use, as potential early failures in production are detected that cannot be detected by other methods.