Understanding Power Integrity: VRM Noise, PDN Impedance & the Role of MXO5 & Bode 100
- Benjamin Dannan
- Apr 29
- 5 min read
Time and time again, we discuss that flat PDN impedance is important to achieving low noise power delivery solutions to a load. However, most PDN designs, especially reference designs from power supply vendors, do not follow this design guidance. Any unwanted noise or transients in the power rails can lead to jitter, reliability issues, voltage droop, power failures, etc.
If you want to learn about power integrity, check out this blog.
Let's provide an example that compares two PDN designs based on an LTM4637 voltage regulator module (VRM) EVM from Analog Devices. The LTM4637 is a complete 20A output high-efficiency switch-mode step-down DC/DC µModule (micromodule) regulator. The VRM package includes the switching controller, power FETs, an inductor, and some compensation components.


Based on the figure above, when we apply the same 3.7A step load to these two PDNs, we see a 63 mV-pk-pk (38%) reduction in the large signal transient response (with a step load) on the PDN with flatter impedance (Z).
So why is the transient response so much better?
LTM4637 EVM – Impedance - High Q PDN vs. Flatter PDN
To answer this question, let's first look at the PDN impedance of these two EVMs.
We used a Bode 100, a P2102A 2-port probe, and a J2113A semi-floating differential amplifier to do this measurement.
The High-Q PDN is the LTM4637 EVM designed with 12Vin and 1Vout out of the box. The Flatter PDN is a second LTM4637 EVM in which only BOM changes have been made to the design to flatten the impedance. These BOM changes consisted of changing PDN capacitors from the primary design. These impedance results are shown in Figure 3.

Looking closer at the non-flat PDN, we can use Non-Invasive Stability Measurement (NISM), invented by Steve Sandler of Picotest, to assess its stability and determine that it has both an active instability from the VRM's control loop at 104 kHz and a passive instability from the PDN at 2.7 MHz.
The instability from the VRM's control loop shows a stability margin of 6.7 degrees, which is quite unstable. In contrast, the passive PDN Q point at 2.7 MHz shows a stability margin of 32 degrees. These NISM results are shown in Figure 4.

LTM4637 EVM – DC Voltage Ripple - High Q PDN vs. Flatter PDN
We have looked at the large signal transient response and the small signal impedance of these two PDN designs on this LTM4637 EVM. Now, let's dive deeper with some voltage ripple analysis of these two PDN designs using the Rohde & Schwarz MXO5 oscilloscope and the P2104A 1-port transmission line PDN probe. Figure 5 depicts the voltage ripple measurement setup for these two EVMs with a 10A resistive DC load applied.

Figure 6 below depicts the voltage ripple result of these two PDN designs with the 10A DC load applied.

The flatter PDN design is 8.4 mV-pk-pk (23%) lower under DC load conditions. Again, we have shown that the voltage noise is lower on the PDN design, which has a lower Q and is flatter.
LTM4637 EVM – Step Response and Noise Spectrum - High Q PDN vs. Flatter PDN
As a last piece of our analysis, let's look at one more thing: the Noise spectrum from the step response of these two EVMs. Both EVM designs are applied a 12Vin, 1Vout, 5A DC Load, 100kHz, 3.7A step Load.
The flatter PDN designs show over 14dB reduction in noise between 1 MHz to 2 MHz. These results are shown in Figures 7 and 8.


Wrapping Up
If there is a single takeaway from this blog, it is that "noise follows impedance!" We even made a t-shirt to drive home this takeaway. A flatter, lower Q PDN will always provide a lower noise, lower voltage ripple, lower jitter, lower EMI, and a better transient response.
Any unwanted noise or transients in the power rails can lead to jitter, reliability issues, voltage droop, power failures, etc. This makes using our State-Space Average VRM models or measuring a vendor design to ensure a stable, flat PDN impedance essential to achieving low noise in your design applications. Another example of the importance of measuring the impedance on a vendor design can be found in this blog.
If you're interested in learning how to make these measurements in your lab, Signal Edge Solutions is happy to help you with our training services..
If your team needs help with PDN measurements, design, or simulation, don’t hesitate to contact us about our modeling services.
References:
R&S® MXO58-ULTM Configured Oscilloscope | Signal Edge Solutions
Picotest P2104A 1-Port Transmission Line PDN Probe | Signal Edge Solutions
Picotest P2102A 2-Port Probe | VRM, Power Plane, & Decoupling Measurements | Signal Edge Solutions
Omicron Bode 100 Vector Network Analyzer | Signal Edge Solutions
Picotest J2113A Semi-Floating Differential Amplifier | Ground Loop Breaker | Signal Edge Solutions
Picotest Port Saver | Instrument Protection DC Block | Transient Protection | Signal Edge Solutions
LTM4637 EVM - DC1872A Evaluation Board | Analog Devices
LTM4637 VRM Datasheet - 4637fc.pdf
Non-Invasive Stability Measurement (NISM)
What is Power Integrity? Understanding Power Distribution Networks (PDN) | Signal Edge Solutions
Large Signal Phenomena using the MXO5 - Sequential Trigger Measurement Setup | Signal Edge Solutions
Protect Your Instruments with the Picotest Port Saver DC Block | Signal Edge Solutions
Video - How to Quantify Flat Impedance and Control Loop Stability with PSpice - picotest.com/wp-content/uploads/2024/09/How-to-Quantify-Flat-Impedance-and-Control-Loop-Stability-with-PSpice.mp4
Power Integrity Station with Bode 500 VNA - PWR500 | Signal Edge Solutions
Comentarios