Simulation Overview - Loudspeakers Design via Shape and Topology Optimization

[Cbdb2]

Yes. And the companies are not encouraging their engineers to work theoretically with the their problems. I was often asked why I had books open instead of just doing measurements.

Thats quite a paradigm shift. So it becomes trial and error. Is it because of the use of simulations? I learnt to describe circuits with equations and a pencil, I suppose they just teach the use of SPICE now instead, and the deeper understanding is gone.
Now we have AI and soon very few will have to understand how things really work.

 

[Cbdb2]

Thanks you for your time. Now the real question. You mention (like SPICE) you need accurate models to get accurate answers. What kind of models are we talking about? How accurate are the present models and how hard is it to get accurate models?

 

[René - Acculution.com]

Thats quite a paradigm shift. So it becomes trial and error. Is it because of the use of simulations? I learnt to describe circuits with equations and a pencil, I suppose they just teach the use of SPICE now instead, and the deeper understanding is gone.
Now we have AI and soon very few will have to understand how things really work.

Trial-and-error, starting too far downstream towards the final product instead from first principles (what are actually trying to achieve here?), a willingness from management in the beginning for the engineers to come up with novel solutions, engineers not really being equipped with the basics they once learned, going in circles, missing deadlines, having to revert back to old solutions. This is why the real innovation happens in smaller companies, or larger companies where certain people are allowed to do things their own way; the innovation typically comes from very few people, and while in the end it will be 'sold' as a teamwork achievement, that is not the case at all.

 

[René - Acculution.com]

Thanks you for your time. Now the real question. You mention (like SPICE) you need accurate models to get accurate answers. What kind of models are we talking about? How accurate are the present models and how hard is it to get accurate models?

Depends. Finite element, lumped, transmission line. Whatever works for the situation. Often a mixture of all it, combining lumped circuits for parts of the physics (not just electrical, but analogy circuits) with finite or boundary element models. Models are typically tricky when it comes to acquiring material data, but in general I will have a very good idea about how a product will behave before it is built. But I need some data, such as Thiele-Small for the drivers and so on. So it depends on which products/physics people are working with.

 

[witwald]

This is why the real innovation happens in smaller companies, or larger companies where certain people are allowed to do things their own way;

It really does seem that the first statement was well and truly contradicted by the second statement.

 

[René - Acculution.com]

It really does seem that the first statement was well and truly contradicted by the second statement.

Small companies where you are not bothered, or large companies where a small group or one individual is left unbothered, not dragged into the bureaucracy.

 

[JoseUJ]

Hi Rene, did you simulate or know how to simulate the heat dissipation of the loudspeaker?

For ex, if the speaker is inside a sealed box how does the speaker heat compounds with the other electronic items?

Thanks

 

[René - Acculution.com]

Hi Rene, did you simulate or know how to simulate the heat dissipation of the loudspeaker?

For ex, if the speaker is inside a sealed box how does the speaker heat compounds with the other electronic items?

Thanks

I have played around with some topology optimization of a heat sink for a driver. But not the heat generation itself. It can be done, and I think some companies have looked into it. I just haven't. The heat has to go somewhere, and there is radiation, conduction, and convection to 'choose' from. It is probably limited what you can do, but still it would be a good research topic, working towards possible engineering solutions.