A Guide to Effective Probe Selection for Low Noise Power Integrity Measurements – Industry Articles

The Picotest P2105A and P2104A probes, when used with the Rohde & Schwarz ZPR20 power rail probe, provide the best low-noise measurement solution for the MXO 5. These single-ended 50Ω browser probes provide similar performance to having a direct measurement coax connection from your oscilloscope to your device under test (DUT).

When acquiring a signal and making a low-noise measurement, we traditionally only consider the oscilloscope’s features, including the bandwidth and dynamic range. It is easy to forget the importance of looking at the signal chain between the scope and DUT. A high-bandwidth oscilloscope is worthless if your probing solution does not have the dynamic range or low enough noise performance necessary for today’s power integrity measurements.

With voltage compliance requirements getting lower and tighter, understanding the noise impact of your probing solution is critical to making good voltage ripple measurements. This is important at your point-of-load or anywhere in your system when assessing performance or validating compliance to specifications.

At Signal Edge Solutions, we take power integrity measurements very seriously. So, we wanted to explore the impact of noise during measurement using various browser probe solutions connected to a Rohde & Schwarz MXO 5 series oscilloscope. This scope has many new features and capabilities and a 2 GHz bandwidth option.

Our Test Configuration

Figure 1 shows a TPS7H4003 evaluation board from Texas Instruments that will serve as the DUT for our experiments. The TPS7H4003 is a radiation-tolerant buck converter configured for a 1 V output. All measurements will be performed using a 5 V input and a 2 A load.

Figure 1. Texas Instruments TPS7H4003 EVM evaulation module. Image used courtesy of TI

Figure 2 depicts our measurement setup using the MX058 oscilloscope model. Only the probing solution is changed during our experiments. The MXO58 oscilloscope is set to 350 MHz bandwidth, and the probe termination is also set as appropriate for each probe under evaluation.

Figure 2. Voltage ripple measurement setup in which only the probe is changed.

We compared two R&S and two Picotest probe solutions:

1. R&S RT-ZP11 10X probe
2. R&S ZPR20 1X power rail probe
3. Picotest P2104A-1X probe
4. Picotest P2105A-1X probe

As shown in Figure 3, we also performed an additional experiment using soldered coax connected to the board in conjunction with the R&S ZPR20 power rail probe. All probes are placed on the same measurement point, an unpopulated capacitor landing area (C23).

Figure 3. Five different probe configurations for our experiment.

Testing the R&S RT-ZP11 10X Probe

Figure 4 depicts the voltage response at the TPS7H4003’s output for the RT-ZP11 probe. The green box highlights the cursor measurements, which capture the peak-to-peak voltage response of the waveform.

Figure 4. Voltage ripple measurement using the R&S RT-ZP11 probe.

A Note on 10X Attenuating Probes

It is important to note that the high probe impedance (10 MOhm) of the RT-ZP11 significantly limits the dynamic range of the measurement setup. As indicated in Figure 4, the vertical range is set to the absolute maximum range available in MXO58. This is because the RT-ZP11 is a 10X probe that attenuates our signal of interest by 10x. This means our SNR is also divided by 10.

Why is this important? When making a measurement on an oscilloscope, it is crucial to utilize the scope’s entire dynamic range to achieve an accurate measurement. This is done by ensuring the signal on the scope screen uses at least 3/4 of the scope’s vertical range. In other words, we are maximizing the signal captured on the scope screen, which ensures we are maximizing the scope’s dynamic range.

This means that with a 10X probe, we have just limited our measurement setup dynamic range in our measurement setup. In short, a 10X probe should never be used to measure small signals. These probes are designed to protect the scope’s front end for much higher voltages. However, engineers still use 10X probes to make small signal measurements, and in doing so, they limit their measurement fidelity.

This means that due to the 10 MOhm RT-ZP11 probe impedance, the signal-to-noise ratio is limited, which limits the max signal level to the MXO 5 and hinders our measurement from taking full advantage of the scope’s dynamic range.

Testing the R&S ZPR20 1X Power Rail Probe

Figure 5 shows the results of the same measurement using the ZPR20 1X probe. The peak-to-peak noise ripple has been reduced from 19.50 to 12.43 mV.

Figure 5. Voltage ripple measurement using the R&S ZPR20 power rail probe.

Testing the Picotest P2104A-1X Probe

Repeating our test using the Picotext P2014A-1X probe provided the output waveform captured in Figure 6. This provides an even lower peak-to-peak ripple voltage of only 10.40 mV.

Figure 6. Voltage ripple measurement using the Picotest P2104A probe.

Testing the Picotest P2105A-1X Probe

The Picotest P2105A is the only other browser probe with even lower inductance and probe loading than the P2104A. However, the Picotest P2105A is only available in a 1X attenuation. Therefore, when your voltage is greater than 5 Vrms, the P2015A must be combined with a Picotest Port Saver DC block solution like the P2131A to protect the front of your MXO 5 scope. The Picotest P2015A-1X probe with the P2131A provided the lowest noise measurement result of 9.32 mV, as captured in Figure 7. 

Figure 7. Voltage ripple measurement using the Picotest P2105A probe.

After initial analysis, the Picotest P21045-1X probe solution provides the lowest noise measurement solution with the MXO 5.

Testing with the Picotest P2104A-1X and the R&S RPZ20

As an additional area of exploration, we wanted to combine the P2104A-1X probe with the R&S RPZ20 power rail probe to assess the performance of this measurement solution. This measurement setup is shown in Figure 8. 

Figure 8. Test setup using the R&S ZPR20 with Picotest P2104A probe.

This combination of probes provided slightly better results than the P2104A alone: 10.08 mV versus 10.40 mV, as shown in Figure 9. We could have tried a similar setup combining the P2105A with the ZPR20, but they were unavailable at the time we were finishing this article.

Figure 9. Voltage ripple measurement using the R&S ZPR20 with Picotext P2104A probe.

Testing with Soldered Coax and the R&S RPZ20

Even though this guide is focused on probing solutions with the MXO 5 series, we still want to cover all of our bases and do one last evaluation to explore the noise performance of soldered COAX in conjunction with the R&S ZPR20 power rail probe. This setup is shown in Figure 10 below, and a close-up view of the soldered COAX solution is shown above in Figure 3.

Figure 10. Test setup using the R&S ZPR20 with soldered coax.

The peak-to-peak voltage ripple for this configuration was measured as 11.40 mV (Figure 11). Although a soldered solution should typically provide better performance, it is essential to remember that it does not offer the repeatability of a browser probe. The advantage of using a low-inductance browser probe such as the P2104A or the P2105A is that they provide low noise repeatability with your measurement setups.

Figure 11. Voltage ripple measurement using the R&S ZPR20 with soldered coax.

Probe Test Results Comparison

As shown in Table 1, the best probing solution evaluated is the Picotest P2105A. The P2105A provided the absolute lowest noise solution that we have evaluated thus far. This low noise is due to higher shielding on the Picotest PDN cable and the very low probe loading from the P2104A probe solution.

Table 1. Voltage Ripple Measurement Results Summary

As demonstrated, the probing solution is crucial to achieving a low noise measurement solution. Even a great passive transmission line probe can outperform a high-end active probing solution. 

The P2105A can also be combined with the ZPR20 to provide an even better lower-noise solution, but this was beyond the scope of this article.

When Testing, Being Right Matters!

When you are testing, you want to get the best possible results. As this testing has shown, the Picotest P2105A and P2104A-1X offer a very high-fidelity and reliable method for measuring low-noise signals. Our results demonstrated that this browser probe is even a better, lower-noise solution than the soldered coax solution and is obviously much easier to use!

The P2105A and P2104A probes combined with the Rohde & Schwarz ZPR20 Power Rail solution provide the best low noise measurement solution for the MXO 5. The benefit of using the ZPR20 with these Picotest probes is that engineers can now take advantage of the highly accurate DC voltmeter features built into the ZPR20. In addition, the ZPR20 has a +/-60V offset compensation. Combining this with a 50-ohm probe like the P2104A-1X or P2105A offers the highest signal-to-noise ratio (SNR) and dynamic range.

In short, This also means if more than one measurement is done on the same DUT simultaneously, a ground loop exists, and a coaxial transformer (J2102B) should be added to your measurement setup.

If you have questions about probing solutions for any oscilloscope, please contact us on our website or by email at info@signaledgesolutions.com.

Except as noted, all images used courtesy of Signal Edge Solutions.

Industry Articles are a form of content that allows industry partners to share useful news, messages, and technology with All About Circuits readers in a way editorial content is not well suited to. All Industry Articles are subject to strict editorial guidelines with the intention of offering readers useful news, technical expertise, or stories. The viewpoints and opinions expressed in Industry Articles are those of the partner and not necessarily those of All About Circuits or its writers.