AnsweredAssumed Answered

FEM electromagnetic Simulation: consistency and convergence issues

Question asked by apaglia on Oct 18, 2013
Latest reply on Oct 23, 2013 by volker_muehlhaus

Like • Show 0 Likes0
Comment • 2

Dear all,

I'd like to point to you the following issue in the FEM Simulator of ADS 2013 (not verified with previous versions).
When you simulate any layout, simple or complex, the simulator shows consistency problems and fails to converge to a proper solution during the Mesh Refining process, IF any layout shape (considered in the Substrate definition) is not in the first quadrant (positive x, positive y) of the cartesian system of coordinates of the Layout Editor window.
Don't know if it is a bug of the FEM simulator. If it is not the case, then I recommend to you to move all the shapes in the layout in the first quadrant of the Layout Editor window.
Best Regards,

Angelo Paglia

No one else has this question

Outcomes

janvh Oct 23, 2013 2:50 AM

Thank you for your feedback. We have conducted specific tests to reproduce the problem you indicated, but did not observe the problem.
Could you provide the test case to an Agilent support representative so that we can examine the specific case in detail?

Actions

volker_muehlhaus Oct 23, 2013 8:26 AM

Hi Jan,

I tested the scenario that Angelo described, and found a similar behaviour with my LINUX installation of ADS 2013. The Windows version works just fine.

Attached is a screenshot of the layout where convergence fails. It does work after shifting the layout into the upper right quadrant with x,y >0.

One thing I noticed on my machine is an error message in the solver log windows. I see this in both cases (no matter where the layout objects are located), so it might be totally unrelated to Angelo's issue. It might be something that happens on my machine only.

Setting EMPROHOME to /usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin
FemEngine
Version:FEM FemEngine_64 10.24 (*) 2012.08-HF4 May 3 2013
Target dataset file: IHP_ind2_2.5nH_N2_w20_s5_d374_FEM
empro-build-2012-10-25-livermorium-ship 260.Ship #282 (64-bit)
ERROR:root:code for **** md5 was not found.
Traceback (most recent call last):
File "/usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin/tools/linux2/lib/python2.7/hashlib.py", line 139, in
    globals()[__func_name] = __get_****(__func_name)
File "/usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin/tools/linux2/lib/python2.7/hashlib.py", line 91, in __get_builtin_constructor
    raise ValueError('unsupported **** type %s' % name)
ValueError: unsupported **** type md5
Running
Simulation path:   /home/volker/Desktop/ADS_Projects/ads2012_08/test_SG25H1/simulation/test_SG25H1_lib/IHP_ind2_2.5nH_N2_w20_s5_d374/layout/emSetup_FEM/proj.ep/Simulations/000001
Simulation status: Running

Best regards
Volker

Attachments

FEM_2013_Linux_Issue.PNG65.7 KB

Actions

2 Replies

janvh Oct 23, 2013 2:50 AM
Mark Correct
Correct Answer
Thank you for your feedback. We have conducted specific tests to reproduce the problem you indicated, but did not observe the problem.
Could you provide the test case to an Agilent support representative so that we can examine the specific case in detail?
Like • Show 0 Likes0
Actions
volker_muehlhaus Oct 23, 2013 8:26 AM
Mark Correct
Correct Answer
Hi Jan,

I tested the scenario that Angelo described, and found a similar behaviour with my LINUX installation of ADS 2013. The Windows version works just fine.

Attached is a screenshot of the layout where convergence fails. It does work after shifting the layout into the upper right quadrant with x,y >0.

One thing I noticed on my machine is an error message in the solver log windows. I see this in both cases (no matter where the layout objects are located), so it might be totally unrelated to Angelo's issue. It might be something that happens on my machine only.

Setting EMPROHOME to /usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin
FemEngine
Version:FEM FemEngine_64 10.24 (*) 2012.08-HF4 May 3 2013
Target dataset file: IHP_ind2_2.5nH_N2_w20_s5_d374_FEM
empro-build-2012-10-25-livermorium-ship 260.Ship #282 (64-bit)
ERROR:root:code for **** md5 was not found.
Traceback (most recent call last):
File "/usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin/tools/linux2/lib/python2.7/hashlib.py", line 139, in
    globals()[__func_name] = __get_****(__func_name)
File "/usr/local/ADS2013_06/fem/2012.09/linux_x86_64/bin/tools/linux2/lib/python2.7/hashlib.py", line 91, in __get_builtin_constructor
    raise ValueError('unsupported **** type %s' % name)
ValueError: unsupported **** type md5
Running
Simulation path:   /home/volker/Desktop/ADS_Projects/ads2012_08/test_SG25H1/simulation/test_SG25H1_lib/IHP_ind2_2.5nH_N2_w20_s5_d374/layout/emSetup_FEM/proj.ep/Simulations/000001
Simulation status: Running

Best regards
Volker
Attachments
- FEM_2013_Linux_Issue.PNG65.7 KB
Like • Show 0 Likes0
Actions

FEM electromagnetic Simulation: consistency and convergence issues

Outcomes

Attachments

Related Content

Recommended Content