Measurement and analysis with MSO/DPO oscilloscopes - Part 1

Power supplies can be found in many different electronic devices, from children’s toys to computers and office equipment to industrial equipment. They are used to convert electrical power from one form to another for proper device operation. Common examples are AC-to-DC converters which change AC voltages into regulated DC voltages or DC-to-DC converters which convert battery power into required voltage levels.

Supplies range from traditional linear supplies to high-efficiency switchmode supplies designed for complex, dynamic operating environments.

The load on a device can change dramatically from one instant to the next, and even a commodity switchmode power supply must be able to withstand sudden peak loads that far exceed average operating levels.

Engineers designing power supplies or the systems that use them need to understand their power supply’s behaviour under conditions ranging from quiescent to worst-case.

Historically, characterising the behaviour of a power supply meant taking static current and voltage measurements with a digital multimeter and performing painstaking calculations on a calculator or computer.

Today, most engineers turn to the oscilloscope as their preferred power measurement tool.

This article looks at common switchmode power supply measurements, shown in Figure 1, using a Tektronix MSO/DPO4000 or MSO/DPO3000 series oscilloscope.

Figure 1: SMPS components that are characterised with DPOxPWR power analysis software.

With the optional power measurement and analysis software (DPOxPWR), these scopes provide automated power measurements for fast analysis and simplified set-up and deskew of probes for maximum accuracy.

Ideally, a power supply would operate exactly as designed and modelled. In reality, components are imperfect; loads vary; line power may be distorted; environmental changes alter performance.

Design is further complicated by demands to increase performance, improve efficiency, reduce size and cut cost.

Given these challenges, the measurement system must be set up correctly to accurately capture waveforms for analysis and troubleshooting. Important topics to consider are:

• Oscilloscope acquisition modes;
• Eliminating skew between voltage and current probes;
• Eliminating probe offset;
• Current probe degauss;
• Bandwidth limiting filters.

The oscilloscopes’ acquisition modes control how electrical signals are sampled, processed and displayed. The resulting waveform points are digital values that are stored in memory and displayed to construct the waveform.

Most scopes support different acquisition modes and the acquisition mode chosen may affect the accuracy of power measurements.

It’s important to understand how acquisition modes function and the effect they will have on the waveform and ensuing power measurements.

Every scope offers sample mode, which is the simplest acquisition mode. As depicted in Figure 2, the oscilloscope creates a waveform point by saving one sample point during each waveform interval (waveform intervals are shown as 1, 2, 3 and 4 in the figure).

Figure 2: Sample mode.

Sample mode is suggested for measurements such as ripple and noise analysis that require multiple acquisitions on non-repetitive signals.

Another acquisition mode offered by most scope manufacturers is average mode. Here, the oscilloscope saves one sample point during each waveform interval as in sample mode. However, in average mode, corresponding waveform points from consecutive acquisitions are then averaged together to produce the final displayed waveform as depicted in Figure 3.

Figure 3: Average mode.

Average mode reduces noise without loss of bandwidth but requires a repetitive signal. Average mode is especially useful when performing harmonics analysis or power quality analysis measurements such as true power, reactive power and apparent power.

The company also offers hi-res mode, in which multiple consecutive samples are taken within one waveform interval and are averaged together to produce one waveform point from a single acquisition as shown in Figure 4.

Figure 4: Hi-res mode.

The result is a decrease in bandwidth and therefore noise and an improvement in vertical resolution for low-speed signals. Hi-res is especially useful for conducting modulation analysis when powering up a supply and acquiring data in a single acquisition.

Hi-res may improve the accuracy of measurements such as switching loss, which are based on mathematically calculated values like instantaneous power.

To make power measurements with a digital oscilloscope, it is necessary to measure voltage across and current through the device under test. This requires two separate probes: a voltage probe (often a high-voltage differential probe) and a current probe. Each probe has its own characteristic propagation delay and the edges produced in these waveforms more than likely will not be automatically aligned.

The difference in the delays between the current probe and the voltage probe, known as skew, causes inaccurate amplitude and timing measurements. It is important to understand the impact of the probes’ propagation delay on maximum peak power and area measurements as power is the product of voltage and current.

If the voltage and current signals are not perfectly aligned, results will be incorrect.

The MSO and DPO series of scopes offer a deskew feature to remove the skew between the probes. When the deskew menu is selected, an information box is displayed that describes the probe model, nominal propagation delay, recommended deskew and actual deskew for each channel.

The voltage and current waveforms have about 8 ns of skew and the propagation delay for each probe is shown in the information box. The TDP1000 (differential voltage probe) has a nominal propagation delay of 6.5 ns whereas the TCP0030 (current probe) has a nominal propagation delay of 14.5 ns. The difference in propagation delays is 8 ns.

Correcting for the skew between the probes is as simple as selecting the ‘set all deskews to recommended values’ side bezel button. Choosing this option adjusts the probes’ actual deskew values to the recommended deskew values. The recommended deskew value is based on the probe’s nominal propagation delay which is stored in the internal memory.

Selecting ‘set all deskews to recommended values’ accounts for the probes’ nominal propagation delay differences which will come very close to deskewing them correctly, but still may not precisely align the waveforms. To precisely align the waveforms for greatest measurement accuracy, the TEK-DPG (deskew pulse generator) and deskew fixture are required.

The TEK-DPG provides a source signal to the power measurement deskew fixture. With the probes connected to the deskew fixture, the ‘actual deskew’ may be manually dialled-in to change the deskew value to precisely align the waveforms.

Differential probes tend to have a slight voltage offset. This can affect accuracy and must be removed before making measurements. Most differential voltage probes have built-in DC offset adjustment controls, which makes offset removal a relatively simple procedure.

Current probes may also need to be adjusted before making measurements. Current probe offset adjustments are made by nulling the DC balance to a mean value of 0 A or as close as possible.

TekVPI-enabled probes, such as the TCP0030 AC/DC current probe, have an automatic degauss/autozero procedure built in that’s as simple as pressing a button on the probe compensation box.

A current probe should also include an easy-to-use degaussing feature. Degauss removes any residual DC flux in the core of the transformer which may be caused by a large amount of input current. This residual flux results in an output offset error that should be removed to increase the accuracy of the measurements being made.

TekVPI current probes offer a degauss warning indicator that alerts the user to perform a degauss operation. Since current probes may have significant drift over time which affects measurement accuracy, a degauss warning indicator is a useful feature.

Limiting the oscilloscope’s bandwidth removes noise or unwanted high frequency content from the displayed waveform, resulting in a cleaner signal.

The MSO/DPO series offer built-in bandwidth limiting filters. In some cases, the probe may also be equipped with bandwidth limiting filters.

The user should be careful when using these filters, as high frequency content contained in nth order harmonics may be removed from the measurement.

For example, if measuring a 1 MHz signal, and evaluating it out to the 40th harmonic, at least 40 MHz of system bandwidth is required. Setting the bandwidth limiting filter to 20 MHz, which is an available option, would eliminate the frequency content required for this measurement.

Once the measurement system is properly set up, the task of performing power measurements can begin. The common power measurements can be divided into three categories: input analysis, switching device analysis and output analysis.

Real-world electrical power lines never supply ideal sine waves and there is always some distortion and impurity on the line. A switching power supply presents a nonlinear load to the source. Because of this, the voltage and current waveforms are not identical. Current is drawn for some portion of the input cycle, causing the generation of harmonics on the input current waveform.

Key measurements for analysing the input of the power supply are:

• Harmonics;
• Power quality.

Switching power supplies tend to generate predominantly odd-order harmonics, which can find their way back into the power grid. The effect is cumulative and, as more and more switching supplies are connected to the grid (for example, as an office adds more desktop computers), the total percentage of harmonic distortion returned to the grid can rise.

Since this distortion causes heat build-up in the cabling and transformers of the power grid, it’s necessary to minimise harmonics.

Determining the effects of these distortions is an important part of power engineering and the benefits of using an oscilloscope rather than a multimeter are significant.

The measurement system must be able to capture harmonic components up to the 50th harmonic of the fundamental. Power line frequency is usually 50 or 60 Hz; though for some military and avionics applications, the line frequency may be 400 Hz.

It should also be noted that signal aberrations may contain spectral components with even higher frequency components.

With the high sampling rate of modern oscilloscopes, fast-changing events may be captured with great detail (resolution). In contrast, conventional power meters can overlook signal details due to their relatively slow response time.

Performing harmonics analysis is as easy as taking an ordinary waveform measurement. Since the signal in this case is a repeating periodic waveform, it’s a simple matter to trigger and display it.

At least five cycles should be displayed to ensure good frequency resolution and the vertical scale should be set such that the signal occupies as many vertical divisions on the display as possible to optimise the oscilloscope’s dynamic range.

In the display menu, measurements on a specific harmonic may be selected. Users may choose to view the results as a table or a graph and can select whether to view all, odd, or even harmonics. Harmonics data may be saved as a CSV file to a USB storage device or CompactFlash card.

Total harmonic distortion values relative to the fundamental and RMS value are also displayed. These measurements are useful in analysing compliance to standards such as IEC61000-3-2 and MIL-STD-1399 that are included in the DPOxPWR power application software.

Power quality does not depend on the electricity producer alone. It also depends on the power supply and the end-user’s load. The quality characteristics at the power supply define the health of the supply and determine the effects of distortions caused by nonlinear loads.

The DPOxPWR power application software provides a results table with the following automatic measurements: V_RMS and I_RMS, voltage and current crest factors, true power, reactive power, apparent power and power factor.

Part 2 will publish next month.