Why Gauss 2D?

The Old Paradigm: Signal Integrity Test Vehicles

The two most important attributes, especially for high frequency, high data rate applications, from a Signal Integrity standpoint in PCB transmission lines are Impedance and Loss. Most advanced Signal Integrity simulation software is reasonably accurate for impedance, but insertion loss, especially at higher frequencies, shows a substantial mismatch between simulation and measurement, with simulation significantly underpredicting measured loss results. Historically, this has been attributed to either more lossy dielectric (higher DF than reported) or more lossy copper (higher roughness than reported). In order to characterize this “missing loss”,  the industry has turned to loss parameter extraction from specially designed Signal Integrity Test Vehicles.

Why Doesn’t the Old Paradigm Work?

There is an inherent flaw with the Old Paradigm: dielectric property measurement and copper roughness measurement are actually sufficiently accurate. With the extraction methods built on Signal Integrity Test Vehicles, product developers would see exorbitant mismatches between reported DF values and those that were extracted, when dealing with ultralow loss materials, the delta would often be >100%. The measurement accuracy of the typically used characterization methods is more than an order of magnitude better than this, so this doesn’t track. Additionally, the fact that different Signal Integrity Test Vehicles on the same material can result in considerably different extracted DF values indicates that there is something wrong with this approach. 

So What’s Missing?

The reason that simulations don’t agree with measurement is, by and large, not an issue of inaccurate inputs, but of inaccurate simulation software. While the underlying methods behind the simulation, especially with the higher end software, may be sound, none of the legacy and incumbent simulation software accounts for ground plane losses. Ground plane losses grow with frequency and become particularly important as frequencies go above 2-5 GHz, which is why this gap between simulation and measurement increases with frequency. A high-fidelity simulation environment that accurately accounts for ground plane losses is the solution.

The New Paradigm

With Gauss 2D, you don’t need to ever build another Signal Integrity Test Vehicle! Gauss 2D was built from the ground up with a high-order Finite Difference and high-order Integration computational engine, to provide the highest accuracy at very high speed. Gauss 2D’s superior underlying technology results in higher accuracy than not only the lower end Boundary Element based software, but even the highest end Finite Element based software, with zero setup time, and hundreds of predefined geometries. Most importantly, because Gauss 2D accounts for ground plane losses – unlike any other solution in the market today – it shows exceptional agreement between simulation and measurement for insertion loss, truly eliminating the need for insertion loss measurement of Signal Integrity Test Vehicles.

Insertion Loss Benchmarks

In the below two examples, we show a comparison between measured insertion loss and simulated insertion loss as reported by two well-regarded competitors, as well as by Gauss 2D. These examples were run with the reported dielectric properties and copper roughness data and the dimensions of the measured traces – no fudge factors or tuning. As evidenced by the two examples, Competitor B shows very poor accuracy for insertion loss over the full domain. Competitor A, while not as poor as Competitor B, also shows poor accuracy that gets perceptibly worse with frequency. However, Gauss 2D (the blue line) tracks extremely closely with the measured insertion loss, something not previously attainable with transmission line simulation, finally bridging the gap between simulation and measurement.

Break free of the old paradigm for PCB transmission line simulation with Gauss 2D Today!