sparsweep_pcell QA

Note: For legacy data format (2022 R2.4 and older) please visit following link:

sparsweep_pcell QA (Legacy)

Photonic Model: sparsweep_pcell

Information on QA tests and QA types: Introduction to Quality Assurance (QA) tests

Instructions on running QA tests: Running QA tests in CML Compiler

QA tests

QA Variables

These variables are defined as below in a struct named "QA" in the source data file: 

[[snippet||12184428062739]]

Statistical QA Variables

(statistical compact models only) 

These variables are defined as below in a struct named "QA" in the "statistical" struct in the source data file: 

[[snippet||12192199680531]]

QA scripts

sparsweep_pcell_passivity_reciprocity_time_domain_qa

sparsweep_pcell_regression_qa

This script generates random values for the parameter set and assign it to the element. It then sets up the the test-bench and extracts the S matrix. For the number of ports the sparameter element has, a test-bench will be set up where the the ONA feeds one port and probes the other ones. Here you can see the MMI set-up as an example.

This is done in frequency domain and time domain and results are compared to makes sure it within the tolerance, S_time tolerance.  Also reciprocity and passivity are tested to be within the defined tolerance, reciprocity tolerance and passivity tolerance . 

The "regression" script also includes a regression test for the absolute and complex values of the S matrix. If reference data is not available or needs to be renewed, "reference_data_available" should be set to false so the S matrix extracted from QA saves as reference data. For consecutive runs, "reference_data_available" should be set to true, and results will be compared to these data and make sure the difference is within the tolerance value. 

sparsweep_pcell_statistical_qa.lsf.lsf

In this test, a testbench is setup in which the ONA feeds the input port of the compact model which is defined in "input_port" and probes the output port defined in "output_port" and Monte Carlo analysis is performed. See test-bench below as an example.  Data is then extracted and compared to data saved in stat_refdata. 

This test is a regression test. If this is the first time running QA, the statistical "reference_data_available" in source data should be set to "false", so the Monte Carlo results from the first run will be saved in the element folder. For consecutive runs, statistical "reference_data_available" should be set to true, and results will be compared to these data and make sure the difference is within the tolerance value