Above is a small comparison with the AICC I guess at some point we have to compare it with the IEEE Open Standards MATLAB code to compare the accuracy of skrf. There is another question : Is it better to fit with IEEE implementation as much as possible and have class and supporting fonctions duplicated into two files of de-embedding implementation extending deembedding class or shall we merge these at one place in deembedding.py (which would probably be more clean) ?Ī first attempt for Z-corrected IEEE de-embedding is implemented. Skrf de-embedding was used to remove Z lines while AICC relies on removing ABCD parameters, maybe I should have a try with that. Results show very good agreement with original dut, AICC is the best. De-embedded with Z correction : skrf (green square) show kind a 1-sample lag with AICC (green curve).The Overal Z of both results are offset from original dut. De-embedded with No Z correction : skrf implementation (red diamond) show excellent agreement with AICC (red curve). ![]() The fix-dut-fix (blue line) embedd additionnal thru with same microstrip. DUT is a skrf simulated 3xWidth portion of microstrip fed with 1xWidth 45 ohm microstrip (black dots).Above is a small comparison with the AICC tool. However, some of you fellas here definitely have the Credentials to do so.Ī first attempt for Z-corrected IEEE de-embedding is implemented. I have had the opportunity to bump into Jason Ellison at Y2019 in a conference and spoke to him a little bit of scikit-rf, as a conference attendee after his presentation then. We should also probably invite the Gitlab Code Owners over here to have a look at scikit-rf. I have not tested it extensively enough yet (so far so good on just 2 samples until 70 GHz, no real world measurement yet) but looking at how some of the big Connectors, measurement & EDA companies are already member of this group, is hard to ignore its relevance. Looking that now we have a De-Embedding Class inside scikit-rf (thanks to Vikram) perhaps we can now consider adding the IEEE P370 de-embedding method into scikit-rf? ![]() exe program over here (although it may not be the latest version):įortunately the codes are open sourced and constantly updated. There is an IEEE Committee of SI Engineers working in parallel like you guys over here to push their own version of the characterization & de-embedding Class for SI Community to implement using MATLAB.įor those who doesn't own MATLAB, you can excess the. I have been observing this standard for quite some time from its early days.Īnd only recently that it has been released as a Standard ( Published Date: )
0 Comments
Leave a Reply. |