Heart rate monitors arrive for refurbishment in Polar's office on an almost weekly basis, despite our having no connection with the Finnish company well known for its fitness monitors.

Ralph, our IT Manager, technical author and Webmaster, is quite concerned. "Think of all those people with heart problems who sent their monitor to the wrong place for repair." A colleague replies, "No, Ralph, you misunderstand, only fit and healthy people use heart rate monitors."

Where is this story going? Well, the same could be said for signal integrity. And this new column aims to highlight the smoothest way to transition quality designs from prototype through to production from a signal integrity perspective. In the world of signal integrity, only the best designers and fabricators measure, model, then measure again to ensure the highest quality predictable signal integrity in their designs and products.

I'll be exploring a number of themes over the coming months: For example, a major topic that is set to grow and grow is the emergence of new silicon families designed to push traditional materials into the multi-gigahertz arena. These new chipsets lift transmission speeds up to a point where signal losses rather than reflections become the predominant concern from an SI perspective.

Another of my favourite subjects is digging for the root cause when analysing measurement and modeling conflicts, something that can be applied to any engineering challenge. In addition, I'll be introducing a range of practical application notes and stories that ensure that the designer's original intentions do not get lost as the design flows from design to prototype and onwards, possibly through a broker, possibly out to a volume fabricator in Asia.

Expanding a little on the topic of transmission line losses, it is worth thinking back to a previous generational shift in high-speed considerations for PCBs; way back in the early 1990s I can vividly recall the emerging need to consider controlling the characteristic impedance of PCB tracks, something we now casually call "controlled impedance" or "impedance controlled." What happened at that time to make PCB transmission line impedance a driving factor in delivering the high-speed boards of that era?

The driving force was the arrival of chipsets with sub-nanosecond switching times. Any interconnect that is long enough at a given speed will exhibit transmission line characteristics and the predominant characteristic of interest will depend on a mix of factors: The line length, the switching speed, the conductor characteristics and the dielectric properties of the insulator carrying the conductor. In 1990s technology, speeds were high enough and traces long enough that without correct impedance matching, reflections could become a significant hindrance to error-free high-speed performance.

Were there losses in 1990s era PCB signal paths? Yes, but in the majority of cases the losses were small enough to be ignored. Roll the clock forward to the "noughties" and a whole variety of logic families is succeeding in running at much higher speeds on the same basic interconnect.