Ten Days of Power Design | Day 5: Testing Power Stage Switching Characteristics

   The demands on power supply designers are becoming higher and higher, and they are under tremendous pressure to improve efficiency, reduce costs, and shorten product development cycles. Power supply design is a complex task, and there are many checkpoints in the process. In the power supply design series, we will introduce you to the test requirements for each of the 10 design stages and give you tips to make your testing more efficient and your life easier.

    In this phase we will describe the steps to test the switch characteristics under no-load, nominal load and full-load conditions.

    Before you begin, make sure the startup, shutdown, duty cycle, and dead time of all switches, such as MOSFETs and IGBTs, are as expected. Tektronix oscilloscopes include a high-resolution mode that radically increases vertical resolution so startup and shutdown times can be calculated with the highest accuracy.

    Although energy is lost in almost all components of a power supply, the majority of losses occur when the switching transistor transitions from the off state to the on state (or vice versa). A more complete understanding of switching losses can be obtained using trajectory plots (available in DPOPWR software) of the start-up and turn-off losses for all switching cycles, as shown in the figure below.

Switch track map on/off.

At this point, check all VGS signals for noise and collisions. This is an important step because any unexpected glitches on this terminal can cause unwanted startup and shoot-through. To ensure that shoot-through is not possible, check the dead time of the synchronous rectifier or H-bridge.

The timing relationship between the gate drivers and related instrumentation is then verified to ensure it matches the calculations for the design.

To safely measure signals that are not ground referenced, we recommend using a differential probe with the appropriate voltage rating. Be sure not to float the oscilloscope, as this can lead to poor results. You can consider the TDP1000, TDP0500, or P6251 high-voltage differential probe, depending on the application. Each probe implements high-speed, broadband acquisition and measurement capabilities, provides outstanding electrical performance, universal DUT connections, and is easy to use.

It is undoubtedly difficult to measure floating gate signals. We recommend probing the signal at the gate driver input so that you can verify the dead time between the top FET and the bottom FET.

Measuring current at the lowest voltage slew rate can also help you minimize crosstalk and improve accuracy.