This article was written by Magna-Power Electronics and originally appeared in the Vol. 1 No. 2 June 2014 issue of the IEEE Power Electronics Magazine.
Switching power supplies utilize power semiconductors to switch between conducting and non-conducting states. Together with passive filters, waveforms of different magnitudes can be produced by defining the on/off periods of the switching states. Efficient power conversion requires switching periods on the order of 25 to 500 ns which produce unwanted voltage transients at the output terminals of the power supply. Further transients can be produced by silicon diodes during the reverse recovery period of current. Reducing the magnitude of switching voltage transients is difficult and depends on careful placement of low impedance capacitors physically located across the output connections of the power supply. Measurement of these transients requires a special setup to obtain repeatable results.
In addition to the challenges of differential measurements, common mode electrical noise can further complicate the measurement process. For power supplies with isolated outputs, internal stray capacitance causes the output terminals to vary in voltage with respect to ground, mainly at the switching frequency. Capacitors applied between the output terminals and ground helps mitigate these voltages, but like switching transients, they are difficult to manage.
Measurement of switching transients with high common mode voltages requires an oscilloscope and probe with a sufficient bandwidth to measure fast transient signals with high common mode rejection. Secondly, and more importantly, is the technique for feeding signals to the oscilloscope. Oscilloscope probes with conventional ground leads cannot be used. The area defined by the ground clip connection to the output terminals of the power supply form a loop susceptible to stray magnetic fields. Magnetic fields, produced by di/dt in the output bus bars and leads to the load, can greatly exaggerate the reading; the readings can be several orders of magnitude in error.
Figure 1 shows the electrical measurement method and sources of EMI. Figure 2 shows the physical test fixture used at Magna-Power Electronics. The test fixture utilizes the ground connection on the probe tip, a short BNC connection to the output bus bars, and an integrated coaxial, common mode filter in the probe cable. Even with the efforts taken to properly measure peak-to-peak output voltage, a helpful task is to first make a common mode measurement by shorting the BNC connection and removing one lead from the power supply. Any resulting erroneous signals can be minimized by adjusting the common mode filter and physical movement of the leads to the oscilloscope.
Magna-Power Electronics has historically taken the position not to publish limits on output peak-to-peak voltage; instead we specify the output rms voltage ripple. The peak-to-peak output voltage measurement requires a setup that is difficult for customers to replicate.
Figure 1 The electrical diagram of the test fixture with an illustration of EMI sources.
Figure 2 The physical test fixture used for measuring peak-to-peak output voltage.