Adaptive Band Synthesis (ABS) in Sonnet
Adaptive Band Synthesis for 3D Planar High Frequency EM Analysis
With Sonnet’s unique Adaptive Band Synthesis (ABS) technique, you can achieve detailed simulation results in a small fraction of the time required by point-by-point EM simulation:
-
You define the band of interest by entering start and stop frequencies
-
Sonnet’s ABS generates a fine resolution response over the entire band with the shortest simulation time possible
ABS uses the smallest number of discrete EM simulation samples possible, and provides a broadband S-, Y- or Z-parameter data sweep, cutting overall simulation time dramatically and filling in the fine spectral behavior with no reduction in accuracy! And it’s reliable and stable for bandwidths exceeding 100x. Compare the results below between measured and calculated using an ABS sweep based on 4 discrete EM analysis frequencies on the right, and a full discrete frequency by frequency simulation on the left.
In an ABS analysis, the Sonnet solver (em) provides several hundred simulation data points for you in only the time needed to analyze a small number of discrete frequencies. The ABS technique uses EM analysis to sample a minimum set of frequencies in order to synthesize a high confidence internal equivalent model. Using this internal model, a very large number of frequency data points are sampled, yielding very accurate and detailed circuit response over the entire band.
The ABS sweep is somewhat similar to other “fast sweep” technologies presently available in some commercial solvers, but special technology developed at Sonnet provides the ability to develop extremely accurate models with less EM frequency samples (usually 50% or better) than current competitive approaches. And the Sonnet ABS sweep technique is stable for bandwidths exceeding 100x.
Consider the following example.
A diplexer implemented in multi-layer Low Temperature Cofired Ceramic (LTCC) is shown in the figure. Sonnet analysis data using ABS is shown in the data plot with 300 frequencies. The full analysis actually requires only 9 discrete EM analysis frequencies, adaptively selected at the frequencies shown with circle markers. The S11, S21 and S31 data are shown in the plot.
Another ABS example is shown here. This design illustrates a Sonnet ABS sweep analysis of a high-Q microstrip bandpass filter. The example below is a Cascaded Quadruplet (CQ) microstrip filter, which employs 8 cross-coupled resonators with a "C" shape. (Design based on "Microstrip Filters for RF/Microwave Applications,Jia-Sheng Hong and M.J. Lancaster, pp. 326-329)
In this example, a Sonnet ABS Sweep requires only 5 discrete EM simulation frequencies in order to synthesize a very detailed response with 1 MHz resolution (501 points). Discrete simulation points are shown by the circle markers on the sweep.
ABS will change the way you view and use EM simulation, making detailed band sweeps possible with minimal simulation time and no loss of accuracy.
