Acoustic Horn Design – The Easy Way (Ath4)

Yeah, I just wasn't sure whether the delay caused by the diaphragm non rigidity (the more distant points always begin to move later than the voice coil) needs to be compensated in the phase plug anywhere below the first vibration mode (then is goes all south). Seemed to me as it should but it's probably only a small effect.
 
Last edited:
Ask yourself, how does one design a phase plug ( meant to equalize pathlengths and therefore delay issues relating to bandwidth problems ) when the diaphragm is violently rocking and parts of it are moving in anti-phase to others?



jgnuepx.jpg
 
Vibration modes of a simple plate with salt on it. Granted, it is neither a dome, nor a cone shape, but it does visualize the variety of decomposition of modes on a structure.



YouTube

A CD diaphragm is not a stretched membrane (2nd order DiffEq,) it is a free edge plate (4th order DiffEq) and the modes are completely different than a membrane.

When JBL first showed their results, I questioned them as the modes that I would have expected (similar to what is shown above) were not there and ones that I did not expect (rim modes) were. I went back (to Ford) and had our FEA people run a test to see if they agreed with JBL and lo-and-behold they were the same. Hence, if any of you think that you know what to expect but have not actually done the analysis then think again, because you are probably wrong.

And, of course, material and thickness and what-not all make a difference, but for everything else held constant as the diaphragm area goes up the mode frequencies go down.

The acceleration factor F/m turns out to be the largest for a smaller 1" driver than the larger ones (at least this is true of all the B&C drivers that I have used.)
 
Comparison of different diaphragm materials, including Klippel scanning of the diaprhagms.

The Use Of Beryllium In Transducers — VUE Audiotechnik

That appears to be a large diaphragm and up until about 10 kHz we don't see any appreciable deviation from rigid. For a 1" throat device with a smaller diaphragm up to 15 kHz a rigidity assumption is probably accurate enough.

The modes that JBL showed and I confirmed were not axisymmetric, like those shown. They were modes around the rim with very low displacements in the center. If the analysis was axisymmetric, then these rim modes would never even show up. With FEA you have to be very careful about how the analysis is done. Plates are by far the most difficult thing to do because of so many degrees of freedom.
 
Last edited:
Forget FEA for a moment... you can literally use a simple timing light to look for resonances in a driven cone, diaphragm, or in a suspension for that matter.



High speed camera technology has advanced so much that you could literally film the results of a diaphragm, cone or suspension, and cut out the middleman. ( FEA analysis time, hours and hours of simulation, plus wages )



Here's a high speed video of a valve spring in a V8 Dodge nascar engine driven at high rpm.



The time for thinking is long over, the time for doing is now.



YouTube
 
Forget FEA for a moment... you can literally use a simple timing light to look for resonances in a driven cone, diaphragm, or in a suspension for that matter.
High speed camera technology has advanced so much that you could literally film the results of a diaphragm, cone or suspension, and cut out the middleman. ( FEA analysis time, hours and hours of simulation, plus wages )
Here's a high speed video of a valve spring in a V8 Dodge nascar engine driven at high rpm.
The time for thinking is long over, the time for doing is now.
YouTube

I'm not sure what is your point. I could say "forget silly expensive cameras and setting up, just measure it with a microphone". Will quickly show resonances.

The point of analysis is to understand root causes, not simply to note final effects, and hence inform alterations to design.
 
Last edited:
I'm not sure what is your point. I could say "forget silly expensive cameras and setting up, just measure it with a microphone". Will quickly show resonances.

The point of analysis is to understand root causes, not simply to note final effects, and hence inform alterations to design.

I miss the point as well. If one has a part then scanning tools and such work very well (although these methods tend to be very expensive for DIY.)

But if one is in the early stages of a design and does not have prototypes, then FEA is all that one has. And it works very well in that setting.

In engineering work one needs to understand the tools and where they are best positioned to be useful. Not all tools work the same way in all situations.
 

Ro808

Member
2010-03-27 4:25 pm
Dr. Geddes presumably refers to the work of Fancher Murray at JBL, around 1988. He showed circumferential modes by using laser technology, used FEA to predict such modes and published a paper:

Murray, F., “MTF (Modulation Transfer Function) as a Tool in Transducer Selection.”
 
Last edited:
The modes that JBL showed and I confirmed were not axisymmetric, like those shown. They were modes around the rim with very low displacements in the center. If the analysis was axisymmetric, then these rim modes would never even show up...

How are these non axisymmetric modes excited?
I would have expected very little excitation and hence very low level.
"f any of you think that you know what to expect" indeed true!

Best wishes
David
 
How much the modes get excited is a different question from what they are. The JBL stuff shows that they are excited, but how that happens is unclear.

I don;t worry much about these details because in the big picture of making great sound loudspeaker systems, it isn't very important. Many years ago we did a blind study of peoples perception of different system designs and most importantly the same design with different drivers. The most important aspect of the results was the basic design, next was the waveguide design (if it had one) and the bass response. The specific drivers appeared to have little to no effect.

Later when we did comparisons of just compression drivers, these small resonance effects could be noted in the results, but were not major effects.

Basically my take-away over the years has been that when "good" drivers are used then the specifics don't matter much, it's the waveguide and the total system that matters most. So I don't pay much attention to these details, although I appreciate the manufacturers for doing them even if they are a small factor in the big picture.
 
How much the modes get excited is a different question from what they are. The JBL stuff shows that they are excited, but how that happens is unclear.

I don;t worry much about these details because in the big picture of making great sound loudspeaker systems, it isn't very important. Many years ago we did a blind study of peoples perception of different system designs and most importantly the same design with different drivers. The most important aspect of the results was the basic design, next was the waveguide design (if it had one) and the bass response. The specific drivers appeared to have little to no effect.

Later when we did comparisons of just compression drivers, these small resonance effects could be noted in the results, but were not major effects.

Basically my take-away over the years has been that when "good" drivers are used then the specifics don't matter much, it's the waveguide and the total system that matters most. So I don't pay much attention to these details, although I appreciate the manufacturers for doing them even if they are a small factor in the big picture.

I've mentioned this before, but I always found it interesting that people obsessed over the oblate spheroidal waveguides, but nobody paid attention to the Summa's cabinet. In particular:

1) I think the roundover may contribute to the speaker "sounding" smaller than it really is. IE, for such a big speaker, it really "disappears."

2) I think the ridiculously inert cabinet improves the sound.

For instance, my refrence speakers at the moment are Yamaha DXR 12s. They have a very nicely done waveguide and a very competent compression driver and the filter was done right.

But even at modest volumes, you can put your hand on the cabinet and feel that it's vibrating, and those vibrations will certainly radiate sound. The overall contribution will be difficult to quantify, because it's so much less than what the woofer and the tweeter are doing. But it will certainly have an effect.

At the same time, it's difficult to find a prosound cabinet that's truly inert. Nearly everyone is focused on keeping the cost and the weight down. Even the plywood cabinets I've seen are quite flimsy.

Andrew Jones, from Pioneer / Elac / Kef / etc mentioned something similar on Facebook the other day, basically that you could improve upon his speakers by using a superior cabinet.