Why buy an expensive EM simulator?

[irfan1]

If you feel offended by my post I am deeply sorry. I have no intention to insult the work of people.

 

[fekete]

I wasn’t offended. I don’t work and never worked in companies that develop code of any type. But I worked on some small EM simulation codes as a student and I know how difficult is to write an acceptable quality one even for very simple cases.
In many postings you will see partisan messages like: Software X is the best, Y is crap.
I don’t like that; all have some good and bad points. Decent competition among software companies is good for the customers. That is why you can find good products at lower prices and even the expensive ones improve continuously. For example HFSS had to find a better user interface (version 9) after the strong competition from MWS, Microstripes and others.

What I am looking for is some concrete examples of success stories with lower cost software, or the contrary, situations when only the luxury software could do the job.

 

[Debeli]

2 examples (pro et contra):

Tools: WIPL-D 5.1(ca 20000 $), HFSS 9.0(ca 45000 $)

1. Problem: Printed Wideband Antenna Design
Model definition time: WIPL-D - 4 hours (experienced user); HFSS - 2.5 hours(mid-level user)
Simulation time : WIPL-D (20 points 1-20 GHz, 13000 unknowns) - 7 hours; HFSS - 12 hours
Result matching to measured : good, slight advantage for WIPL
Time neccesary to make major change in the design: WIPL - D -1 hour; HFSS -15 min
Winner: WIPL-D

2 Problem: Multi-pole Dielectric resonator filter, 1 GHz
Model definition time: WIPL-D - 20 hours (experienced user); HFSS - 3 hours(mid-level user)
Simulation time : WIPL-D (21 points, 10000 unknowns) - 8 hours; HFSS - 6 hours
Result matching to measured : Spar WIPL - fair, HFSS - good, frequency accuracy - both good
Time neccesary to make major change in the design: WIPL - D -3 hours; HFSS -30 min

Winner: HFSS

Winner is for me tool that I would take for another similar project.

Similar experiences?