[Bug 47592] New: HFSS antenna modelling software running but erroneous output !

[wine-bugs at winehq.org]

https://bugs.winehq.org/show_bug.cgi?id=47592

            Bug ID: 47592
           Summary: HFSS antenna modelling software running but erroneous
                    output !
           Product: Wine
           Version: 3.0.2
          Hardware: x86
                OS: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: -unknown
          Assignee: wine-bugs at winehq.org
          Reporter: friedtj at free.fr
      Distribution: ---

I wanted to share with this community a rather amazing observation
that has been baffling me for the last week: ANSYS HFSS 13.02 is
running with Wine, but the output is erroneous ! I can hardly imagine
how a calculation can be completed but yield incorrect results: my
understanding
is that a library (or API implementation) either works or does not -- and to
tell the truth I did not expect such a complex piece of software as HFSS to
run on Wine -- but it does and rather nicely except to the point of analyzing
the
results.

Being new to HFSS but quite familiar with antenna design, I started with the
trivial test case of the dipole antenna: resonance frequency was close to the
expected value considering the dimensions, so the software must be working
properly since it computes admittances (see http://jmfriedt.org/S11both.pdf,
green curve)
which are consistent with the values given by (FOSS) NEC2 and GPRMax3D.
However all attempts to plot the radiation pattern failed (no plot at all, or
sometimes
inconsistent results of an isotropic radiator with no null), and radiation
efficiency
was consistently 0 (a poor result for a classical antenna), so I was of course
questioning my ability to understand the proper operation of the software.
After
attempting all possible solutions, I ended up installing a Win7 system on
VirtualBox
for testing, and to my amazement the software yields the proper radiation
diagram (http://jmfriedt.org/diagramme_3D.png) and slightly different
reflection
coefficients (http://jmfriedt.org/S11both.pdf, purple curve), the difference
being
emphasized on the impedance curve (http://jmfriedt.org/Z11_both.pdf).

I am of course not blaming the amazing emulator on this difference, but with
this
post
1/ may point out a subtle misbehavior of one of the computation library used by
Wine
2/ help others attempting to run HFSS on Wine question the results they
achieve.

These tests were completed with wine-3.0.2 as provided by Debian GNU/Linux
Unstable
at the time of writing this post, configured (winecfg) either as Win10 or Win7,
fresh
install with no other MS-Windows or proprietary library installed prior to HFSS
(clean
$HOME/.wine directory).

Not sure how I can help debug,
best wishes, Jean-Michel

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.