Switch Mode Power Design and Component Analysis Using Keysight’s Oscilloscopes

If you are designing your own power supply, or characterising an off-the-shelf design, or even individual components, you will need to make a broad range of measurements - most of which can be performed using a Keysight oscilloscope.

However, many engineers still perform these measurements manually, which can be very time-consuming. Keysight’s InfiniiVision X-3000T oscilloscope with Power Measurement option (DSOX3PWR) automates and simplifies measurements for Input, Switching and Output analysis.

Frequency Response Analysis is another performance metric of great interest to engineers - this typically requires a Vector Network Analyser (VNA) but the 3000T includes this as well!

Let’s look at just some of the most-frequently used measurement; the unique frequency response analysis capability of the 3000T, plus some tips on probing.

Power quality analysis measures the input AC line signal quality parameters: Real power; Apparent power; Reactive power; Power factor; Voltage crest factor; Current crest factor and Phase angle.

The screenshot below shows the input voltage (yellow), current (green) and power (pink) waveforms plus all of the measurements on one easy-to-read screen.

Screenshot of input power analysis

These are all automated measurements - the user just needs to connect the current probe and voltage probe to the right part of the circuit. The built-in set-up guide even shows where to put the probes:

Screenshot of the probe connection diagram for input analysis

Current harmonics analysis measures the amplitude of frequency components that can be injected back into the AC lines. End-products must often meet specific standards of compliance in order not to disturb other equipment connected to the AC supply grid. This measurement will perform an FFT measurement on the current waveform, compare the results of the amplitudes of odd and even harmonics against a user-selected IEC standard, and also provides colour-coded pass/fail indicators for each tested frequency up to the 40th harmonic. The results are shown below.

Screenshots of current harmonic content analysis

Other input characteristics that can be measured include Inrush current analysis (the peak input current when power is initially turned on).

Switching loss analysis measures power and energy losses of your switching device (typically a FET). In a switch mode power supply (SMPS), most power and energy losses occur during the switching phases of the transistor when it turns on and turns off. During these phases, the switching transistor goes in and out of saturation and momentarily operates in its linear region. Energy losses also occur during the conduction phase of the switching transistor. This is when voltage is at the transistor’s saturated minimum and current flows. Losses during the non-conduction phase are typically insignificant, and should theoretically be zero. The Switching Loss measurement automatically measures losses over one switching cycle. But you can then optimise the scope’s horizontal and vertical settings to perform more accurate power and energy loss measurements during particular switching phases.

The 3000T oscilloscope can measure a range of switching losses, including: Rds_on and Vce_sat, total switching loss, Slew rate and Modulation analysis. These measurements can be used to select different devices to optimise the component selection process.

Modulation analysis is typically used to characterise the pulse-width modulation (PWM) feedback from the DC output to the switching transistor’s gate terminal for dynamic voltage regulation.

Screenshot of the trend of Duty Cycle in Modulation Analysis

Output analysis features include output ripple, turn-on/off time, transient response and overall efficiency.

Output ripple analysis measures the peak-to-peak and the AC_RMS of the power supply’s DC output. The AC_RMS measurement is the same as standard deviation (σ), which is commonly used to characterise random noise. Output ripple is typically dominated by switching noise, but can also include other random noise and signal coupling from various sources in your system.

Screenshot of output ripple analysis showing ripple content: Ripple, AC_RMS, Frequency range

Unique Frequency Response Analysis (FRA) features include Power Supply Rejection Ratio (PSRR) Analysis and Control loop response measurement. Control loop response measures the stability of the power supply and is represented as a Bode plot of gain and phase over frequency.

Screenshot of Bode Plot of gain and phase response.

The Bode plot feature can also be used to characterise other components or circuits, such as filters and amplifiers.

Probing Tips

We’ve discussed some of the features and automatic measurements available in the 3000T oscilloscope with the Power option. Now let’s look at the other important element in power measurement: probing. Our three top tips are:

(1) Match the probe attenuation and Maximum Voltage rating for your application.
Most scopes come with 10:1 attenuation passive probes. For low voltages, 1:1 probes provide better noise performance - great for signal integrity applications such as checking VCO, ADC or DAC performance.

For higher voltages, higher attenuation passive probes are used. But remember, the probe needs to be safety-rated with appropriate sized clearances for the voltages being measured; not just offer high signal attenuation. Using a high-attenuation probe and increasing the gain on the scope always increases measured noise and often impacts bandwidth and rise-times.

(2) Differential probes offer isolation as well as attenuation.
You can measure differential signals using two standard single-ended passive probes and the (A-B) math function of the scope. But this technique does not offer the isolation commonly needed for power measurements that true differential probes offer, such as Keysight’s N2891A (100:1 / 1,000:1, 7kV).

(3) Current probes – the right size for the job.

Common current probes are typically good from about 100mA, and to measure lower currents. A common trick is to make several winds around the current probe jaws, but this also increases noise. Keysight have dedicated low-current probes (N2820A/21A) which work down to 30µA and up to 5A.

After extended use, a magnetic field can build up in a current probe so you should occasionally "degauss" (de-magnify) the probe. You should also de-skew the probes by checking the current and voltage waveforms and correct for the delay between signals. Keysight scopes can auto-deskew with a dedicated fixture.

For more tips, download your free Power Supply Testing Toolkit today!