Discussion Forum

If you are doing a frequency sweep, then there is an assumption of linearity. When linearity holds, you can sum up individual parts of several sources, and you can change the phase (and magnitude) of each of them as you please, so just apply a unit input and adjust in post-processing. But, if the sources have an associated impedance, then that needs to be included for each study. So, if you apply a displacement/velocity/acceleration condition on one active source, then all other sources (passive) can be hard walls, which is anyway the default. For other source types, there is more to it. This is equivalent to Thevenin/Norton approaches to electrical sources.

Thanks! Indeed, the other sources are hard walls and I have no issues summing the pressure fields from individual sources, but summing the exterior fields (efc1.pext by default, to be clear) does not work properly. I can do BEM or extend the computational domain in FEM but both options are pretty much slower than a smaller FEM domain with EFC

Thanks! Indeed, the other sources are hard walls and I have no issues summing the pressure fields from individual sources, but summing the exterior fields (efc1.pext by default, to be clear) does not work properly. I can do BEM or extend the computational domain in FEM but both options are pretty much slower than a smaller FEM domain with EFC

What is the problem with the exterior field? It has phase information in it.

That's a good question. Summing the exterior fields just seems to produce very incorrect results, although it doesn't look totally random either. What's curious is that the dynamic range of the radiation pattern seems to increase from around 60dB to 130dB, if that helps at all guessing what I might be doing wrong

That's a good question. Summing the exterior fields just seems to produce very incorrect results, although it doesn't look totally random either. What's curious is that the dynamic range of the radiation pattern seems to increase from around 60dB to 130dB, if that helps at all guessing what I might be doing wrong

Probably best you just upload the file.

Nevermind! After I managed to convince LLMs that superposition should apply, they pointed out the obvious flaw: I was using Lp_pext instead of summing pext and converting to SPL :)

Good :)