In previous posts (#006 and #008) I have shown how you can make virtual prototypes for loudspeaker drivers. Now, I have also made a tweeter. This geometry can be used for investigating the effect of e.g. ferrofluids, porous materials, former holes, and opening up or closing off the air volume behind the dome, but also for doing shape and/or topology optimization; something I don't really see loudspeaker manufacturers venturing into. So let's give them a push ;-)
The tweeter geometry is, as was also the case with my woofer, made from memory of many different designs, and is not at all meant to be perfect in any way. I just need something more than a piston, that has a somewhat plausible outline, and a response that doesn't look too bad. Also, manufacturability is not an issue for now; any solution techniques used with this geometry will transfer directly to more accurate designs. For now, it looks like this:
There is the option to open up the cavity, and I will probably add some ferrofluid later. I have a pretty good idea of material properties, but the magnet system is not known, so I just put in some values. A frequency response simulation was run in COMSOL Multiphysics, and things seem to be connected correctly. I will not go into detail with the actual simulation setup; by now, there should be plenty of blog posts and other literature on this topic elsewhere.
It is always a good a idea to animated all DOFs, as this will often reveal mistakes. Here, I animate pressure and displacement, and I see that there is wave propagation, and mechanical boundaries have also been applied correctly.
I often just sit and watch animations and try to relate what I see to fundamental principles, that have to apply for the current case. And at the same time I think about misconceptions or confusions that I have had myself, or claims I have heard from colleagues, which do not fit with these principles. One thing that I see many engineers get wrong is how pressure relates differently to displacement/velocity/acceleration in different situations. To get rid of these confusions, analytical and or numerical methods are worth wild. Our tweeter example can used for this purpose. So for example try and ask yourself the following: How does pressure phase in front of the dome correlate with pressure phase behind the dome (at low frequencies)? Using intuition to solve physics problems is bound to fail at some point, and I am probably annoying to some colleagues, as I shoot down many ideas based on intuition alone quite a lot, unless of course the physics is shown to match the idea.
At some point in time, maybe next year as I am fully booked with moving to a new house, presentation commitments, and guest lectures, I will apply optimization methods to the tweeter, and hopefully some other applications also. With having drawn my own set of woofer and tweeter, I don't have legal issues with any companies, which will make everything must simpler.
Btw., this week I realized that I had some misconceptions about wave propagation in tubes, duct modes, and evanescent vs propagating waves, and will also try to find time to touch upon this topic and how it relates to/differs from modal analysis in a later blog post.