Beyond Design: Plane Crazy, Part 1


Reading time ( words)

A high-speed digital power distribution network (PDN) must provide a low inductance, low impedance path between all ICs on the PCB that need to communicate. In order to reduce the inductance, we must also minimize the loop area enclosed by the current flow. Obviously, the most practical way to achieve this is to use power and ground planes in a multilayer stackup. In this two-part column, I will look at the alternatives to planes, why planes are used for high-speed design, and the best combination for your application.

Back in the mid-eighties, when I worked at the University of Western Australia, one of my duties was to fix the departmental mainframe: the dreaded DEC PDP-11/40. When it broke down, it was a two-week sentence to solitary confinement in the frigid computer room. This monster machine had card after card with rows of TTL logic chips. Figure 1 illustrates a typical Unibus board. It had 8K, 16-bit word core memory, which I believe could be expanded to 80K if the need ever arose. The core had a 400ns access time, which means the system clock would have been a blazing 2.5MHz.

I always used the “divide and conquer” methodology. First, eliminate the power supplies then start dividing the system in half, then half again until the fault was localized within a small circuit. But, as it took about half an hour to reboot, with a specific sequence of octal latches, it was a very time consuming process. Plus, there were always numerous engineering students banging on the window, to the terminal room, enquiring when the “mother” might be fixed so they could complete their assignments.

The boards were double-sided and used a power finger, type A or B layout configuration on the top side of the board, as shown in Figure 2. The bottom side could then be used entirely for routing. This provided some mutual inductive coupling between the wide power and ground traces and saved on board area. However, it meant that the return current had to flow all the way around the board perimeter, creating a large loop area. Fortunately, the PDP-11/40 was manufactured way before the era of FCC-mandated radiation guidelines. Needless to say, this is not a good supply configuration for high-speed design. Don’t try this at home!

To read this entire article, which appeared in the December 2015 issue of The PCB Design Magazine, click here.

Share


Suggested Items

PCB Cooling Strategies, Part 1

01/19/2018 | Bin Zhou, EDADOC
With the development of communication and IT industries and the ever-increasing demand for information analysis, many chip makers have racked their brains trying to provide customers with better technology, such as increased computing power and storage capacity of chips as well as diversifying their product offerings.

Streamlining Thermal Design of PCBs

01/10/2018 | Dr. John Parry, CEng, Mentor
When designing a PCB, thermal issues are often locked in at the point of selecting and laying out the chip package for the board. After that, only remedial actions are possible if the components are running too hot. Assumptions made about the uniformity of the airflow in these early design stages can mean a disaster for the commercial viability of a PCB if those assumptions are incorrect. A different approach is needed to improve reliability and to optimize board performance. Dr. John Parry of Mentor explains.

Mike Jouppi Discusses his Drive for Better Thermal Data

01/12/2018 | Andy Shaughnessy, PCB Design007
If you mention thermal management in a group of PCB designers and design engineers, Mike Jouppi’s name usually pops up. Mike is an engineer and founder of the Thermal Management LLC consulting firm. He spent years updating IPC’s charts on current-carrying capacity, which had been unchanged since the 1950s. I recently caught up with Mike and asked him to give us his views on the state of thermal management, as well as the tools and standards related to thermal design.



Copyright © 2018 I-Connect007. All rights reserved.