Technical Articles

RSS

There are many questions about measuring Power Delivery Networks (PDN), but these two are very common ones.  Why do we calibrate the 2-port measurement with a 1Ω shunt resistor and why do I use DC blockers on both ports?  In this article I’ll provide responses to both of these questions.  The measurement setup in Figure 1 is an example where I used both the 1Ω calibration and the inclusion of the DC blockers.

As systems designers work hard to squeeze more and more features into less board space, the power delivery paths are becoming increasingly complex. The current mature VRM designs based on Silicon MOSFETs are hardly meeting present day requirements. One of the promising technologies touted to solve this conundrum of space and performance constraints is GaN HEMT. However, many engineers are hesitant to design very high frequency GaN VRMs from the ground up. This paper evaluates the steps required to modify existing Si-MOSFET designs for use with eGaN HEMT devices. The paper also compares the expected performance of GaN vs. Si in linear and switching regulator topologies and covers some of the measurement challenges as well.

Models of transmission lines and transitions accurate over 5-6 frequency decades are required to simulate interconnects for serial data channels operating at 10-100 Gbps. Extremely broadband modeling of conductor properties for such high-speed channels is a challenging task. This paper explains physics of the conductor-related signal distortion effects in PCB and packaging interconnects. After reading this paper, you should be able to setup simple experiments in your EDA tool to figure out the limitations and will be sufficiently qualified to ask your EDA tool vendor questions about the accuracy of the conductor modeling effects.

This paper will demonstrate the application of DOE and RSM to a CEI 28G VSR design. We will show the process of creating a DOE, fitting the data to models, determining the goodness and reliability of the fit and then using the model to perform “what if” analysis, optimize design factors and quantify the impact of manufacturing variation. 

In the GB/s regime, accurate modeling of conductor losses is a precursor to successful high-speed serial link designs. In this paper, a practical method for modeling conductor surface roughness is presented. Obtaining the roughness parameters solely from manufacturers’ data sheets, conductor loss can now be accurately predicted from first principles. By using a close packing of equal spheres model, the radius of the spheres and area of the multi-sphere tiled base are determined then applied to the Huray “snowball” model. A case study, based on Megtron-6 and N4000-13EP dielectric with HVLP and VLP copper foils respectively, validates the model’s accuracy.