• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Simulation Analysis - Screws in Waveguide

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,780
Location
Oxfordshire
Some manufactures use thin cabinets and 'tune' their sound.
Indeed, I read all the BBC research dept documents on this back at the time.
It is much much easier and cheaper to do this sort of investigation today and much more likely to give accurate results!
Software has come a long way, but the best way to understand the expensive software is to have programmed it yourself
I am sure it is. It was with the simple software I was writing too, racing car suspension and transfer matrix vibration analyses. A lot of the effort went into trimming the data to get useful results with 32kbytes of RAM :)
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
721
Likes
359
Sure. I think for this forum it is a good idea to look at general topics: "Does it matter if you round cabinet edges?", "Do screws matter or not?", "Does cabinet vibration matter?" and so on. Anything where there are myth about general topics can often be closed down fairly quickly with simulations.
I don't know how to describe it but see if you understand me.....o_O What is the effect of using a driver at the back side of horn? What I mean is the sound does not travel inside (or front side) of the horn, but travel outside (or back side) of horn. What about not horn but other shape?

Another one is what is the effect of anything in front of driver or even envelope the driver with a few holes for sound to come out.
 
OP
René - Acculution.com

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
416
Likes
1,221
I don't know how to describe it but see if you understand me.....o_O What is the effect of using a driver at the back side of horn? What I mean is the sound does not travel inside (or front side) of the horn, but travel outside (or back side) of horn. What about not horn but other shape?

Another one is what is the effect of anything in front of driver or even envelope the driver with a few holes for sound to come out.
Sure, some phase plug/obstruction study. Sure. Do you know any specific drivers that use the first principle that you describe?
 

maxxevv

Major Contributor
Joined
Apr 12, 2018
Messages
1,872
Likes
1,964
If someone can draw an approximate geometry that would certainly be interesting (as long as Genelec does not mind).

That's not too difficult. I think we could work with the configuration that @Rick Sykora was doing with the DIY speaker experiment.
I could whip up a few models quite quickly if needed. But I need the models of the tweeter and woofer being used.
I previously did the original tweeter that Rick selected, perhaps I could combine with that with the Purifi woofer and do a few configurations.

Would need specifications for the bass port and casing volumes to get a reasonably accurate model. And also, make it an exercise worth the time involved.

I see that Comsol imports UG NX files directly, that could help smooth things out though.
 

BostonJack

Active Member
Editor
Joined
Jul 2, 2019
Messages
288
Likes
350
Location
Boston area, Cambridge, MA
You are very kind. Love to see more simulations. One I like to see are diffractions around the edges of a rectangular speaker.

I'm interested in the difference between small baffle edge radius (say 5 mm) and large radius (say ~20 mm), plus the angular (prismatic) shape done on say (to pick a non-random example) the DXT-Mon.

I've seen DIY testing of different angles in the prismatic case. I'll search for a pointer.

congrats on great work. I used to verify chips for Intel, Cisco, AMD using simulation techniques (discrete) so I am all about the simulations!

Jack
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
I'm interested in the difference between small baffle edge radius (say 5 mm) and large radius (say ~20 mm), plus the angular (prismatic) shape done on say (to pick a non-random example) the DXT-Mon.

In @Rick Sykora 's DIY Directiva thread you can find some simulations on speaker cabinets and their optimization.
It's not very systematic, since most of the simulations took place over a PM exchange, but it might be interesting as a first start.

See here, here, here and here.
 

TheWalkman

Senior Member
Joined
Jan 9, 2020
Messages
380
Likes
998
As someone who is becoming an old curmudgeon by the day, I suppose I should be concerned about the hair I now find growing out of my ears and how it impacts my listening experience.

For the sake of audio science, I should probably submit my noggin to Amir for testing, but his setup looks terribly uncomfortable and I’m not sure how to box up my cranium for the UPS guy.

In the meantime, I’ll turn the volume up a notch and have a second G&T to compensate for the hair-wave interaction.

Cheers!
 

DDF

Addicted to Fun and Learning
Joined
Dec 31, 2018
Messages
617
Likes
1,355
@René - Acculution.com thanks for sharing your expertise. This type of analysis is immensely powerful

ASR tests show that loudspeaker port pipe-resonance remains a significant source of colouration. Given that grills and diffraction are well studied and shared in the public domain, it would be novel to investigate if simple changes in port location or shape can result in rules to reduce their amplitude. Last study I saw on this was in the JAES out of Japan 25 years ago.

There are more complex ways to mitigate these resonances but almost all manufactures don't bother due to cost (product or design) or knowledge:
- Technics method of a lossy baffle at input side (risks raising even order distortion due to non symmetrical flow)
- Kef's lossy port section at the pressure maximum
- strategic holes drilled in ports (trial and error)
- resonators hanging off the side like a muffler (big n nasty)
- good old straws (reduces maximum SPL)
 

Biblob

Addicted to Fun and Learning
Joined
Sep 13, 2018
Messages
633
Likes
600
@René - Acculution.com thanks for sharing your expertise. This type of analysis is immensely powerful

ASR tests show that loudspeaker port pipe-resonance remains a significant source of colouration. Given that grills and diffraction are well studied and shared in the public domain, it would be novel to investigate if simple changes in port location or shape can result in rules to reduce their amplitude. Last study I saw on this was in the JAES out of Japan 25 years ago.

There are more complex ways to mitigate these resonances but almost all manufactures don't bother due to cost (product or design) or knowledge:
- Technics method of a lossy baffle at input side (risks raising even order distortion due to non symmetrical flow)
- Kef's lossy port section at the pressure maximum
- strategic holes drilled in ports (trial and error)
- resonators hanging off the side like a muffler (big n nasty)
- good old straws (reduces maximum SPL)

So the common themes seem to be

- Edge diffraction (working on it)
- Enclosures; structurally especially
- Bass ports

All very much suited for some general analyses. Thanks for your input (and the very nice reception over all from people).
Regarding the last two points, DIY-Audio user Augerpro made a big shootout using different box materials and also is testing port designs. He measures frequency responses as well as distortion. Maybe it can be of use :)
Thanks for spreading these awesome facts!

In the links below you can find information about it:
https://www.diyaudio.com/forums/mul...-construction-methods-shootout-thread-24.html

https://www.somasonus.net/box-construction-methods

(forgive the non hyperlinks, doesn't work on my phone)
 

pierre

Addicted to Fun and Learning
Forum Donor
Joined
Jul 1, 2017
Messages
962
Likes
3,048
Location
Switzerland
If I may ask. I did a lot of shape optimization long time ago. That would apply very well to horns and other waveguides.
gradient are more complicated to compute (wrt the shape) but it is easy if the shape is parametric for example ( a horn could be modeled as 2 splines and then the dof are only the parameters of splines). You would optimize for directivity for example and solve the direct problem at each step.

the jbl m2 horn would be a good example.

do you know if it has been done before?
 
Last edited:
OP
René - Acculution.com

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
416
Likes
1,221
If I may ask. I did a lot of shape optimization long time ago. That would apply very well to horns and other waveguides.
gradient are more complicated to compute (wrt the shape) but it is easy if the shape is parametric for example ( a horn could be modeled as 2 splines and then the dof are only the parameters of splines). You would optimize for directivity for example and solve the direct problem at each step.

the jbl m2 horn would be a good example.

do you know if it has been done before?
Some work has been done in that field, and shape optimization and waveguides are a great match. It can be done via polynomials or splines, or via more free form strategies. Some recent work has been done by Bezzola (https://www.researchgate.net/public...s_for_Acoustic_Elements_in_Loudspeaker_Design) and l looked at even more diverse cases (https://www.researchgate.net/public..._Topology_Optimization_of_Loudspeaker_Drivers).
 

Coach_Kaarlo

Active Member
Joined
Apr 27, 2020
Messages
196
Likes
222
Location
Sydney
The Yamaha NS-5000 has some interesting design features which have been developed using similar software tools and processes (in my understanding). And Yamaha has access to far more in-house resources than many other audio manufacturers @René - Acculution.com what is your view of some of these engineering approaches - novel or innovative? Interested in your informed opinion.
 
OP
René - Acculution.com

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
416
Likes
1,221
The Yamaha NS-5000 has some interesting design features which have been developed using similar software tools and processes (in my understanding). And Yamaha has access to far more in-house resources than many other audio manufacturers @René - Acculution.com what is your view of some of these engineering approaches - novel or innovative? Interested in your informed opinion.
Well, there is a bit to unpack here... Specifically for this loudspeaker there are some interesting technology going on, and they actually illustrate the functionality via animations from simulations. Nice. But some companies are even better at showing off their simulation work, such as KEF, Harman/Samsung US, and a few others, and make YouTube videos and present at different venues. Now, the innovations that we see are typically not born out of simulations; the resonator/"meta"-material stuff behind the drivers are first investigated via analytical work (it is often fairly basic) and later validated via simulations, which also makes it easier to sell to management as you can show these nice looking animations, and they can also be used directly in sales material, and to educate fellow engineers.

Now, where simulations really can LEAD the innovation is topics such as design of composite materials (layers of anisotropic materials),
Untitled.png

and formal mathematical optimization (parametric, shape, topology),
Untitled2.png

where you simply have to let the computer run (with you guidance) and see where you end up, iterate from there, contemplate, and so on. I work with these techniques with some companies that have enough foresight to see that is where the industry is heading anyway, and so better partner up with someone who already has experience in the field than try and build it in-house.

When it comes to these new innovations in large companies, it is perhaps worth noting it is typically the same 1-3 people coming up with them, though they may be 10-20 times that in the respective department. I have worked in several companies (not just loudspeaker) and a major issue is that the average engineer has let go of a lot of knowledge acquired in uni; complex numbers are an issue, signal processing is not well-understood, basic acoustic and solid mechanics posed problems, structuring reports, the scientific approach in general, working from first principles, and so on. This can give an unfortunate dynamic in a company, especially if management is not investing in year-round knowledge and competence sustaining/building, and that creates a divide between those who are constantly chipping away at fundamental and new skills, and those who have mentally replaced "I am not using what I learned" with "I don't need to use what I learned, because my fellow engineers are not using it either". I have worked with a coach some years back to figure out how to lift departments, looking into biases and fallacies, but it became too much of a hazzle. But this is a whole other can of worms...
 
Last edited:

Scgorg

Active Member
Joined
Jun 20, 2020
Messages
129
Likes
420
Location
Norway
If I may ask. I did a lot of shape optimization long time ago. That would apply very well to horns and other waveguides.
gradient are more complicated to compute (wrt the shape) but it is easy if the shape is parametric for example ( a horn could be modeled as 2 splines and then the dof are only the parameters of splines). You would optimize for directivity for example and solve the direct problem at each step.

the jbl m2 horn would be a good example.

do you know if it has been done before?
While I haven't gotten around to reading this paper yet, the abstract seems to indicate it being something along the lines of what you wish (and not linked by René already): https://www.aes.org/e-lib/browse.cfm?elib=21009
 
Last edited:
OP
René - Acculution.com

René - Acculution.com

Senior Member
Technical Expert
Joined
May 1, 2021
Messages
416
Likes
1,221
While I haven't gotten to reading this paper yet, the abstract seems to indicate it being something along the lines of what you wish (and not linked by René already): https://www.aes.org/e-lib/browse.cfm?elib=21009
Nice. They are doing things very close to what I would do. Waveguides are easier than e.g. compression drivers where you can have colliding boundaries and is much more resonant, but they illustrate the functionality very well.
 
Last edited:

MrPeabody

Addicted to Fun and Learning
Joined
Dec 19, 2020
Messages
657
Likes
942
Location
USA
Dr. Rene, thank you very much for doing something very interesting and enlightening. You are making an extremely fine contribution to ASR, and I am very much looking forward to your findings with the other things you mentioned.

As you allude, it is quite apparent that with respect to acoustics especially, there is no other approach or method that compares to simulation. One of the members here, @ctrl, has done some very excellent work with his simulations, showing some of the effect of baffle edge diffraction and also waveguides. I expect that some of what you do will overlap with some of what he has done, and that there will be a positive synergistic effect. I look forward to this.

One thing I thought I would note in passing is that the frequency you used to assess the effect of the screw in the waveguide is high such that for many of us, it is moot because we can't hear it. Even when I was in my early thirties I could not hear above 15 kHz. I could only assume that I was able to hear 20 kHz at some younger age, but cannot say for certain. At this frequency, the wavelength is, I think, right about 1.7 cm, which means that the diameter of the little screw was slightly more that 1/3 of the wavelength. As such I am surprised that there was any effect at all, especially since it seems to me that the height of the screw is likely the more important dimension, and the height of the screw is less than 1/10 of the wavelength. Given that both dimensions, especially the height, are less than the wavelength by a significant amount, I expect that the effect of the screw will be nearly as great at 10 kHz as at 20 kHz. Obviously you cannot spend too much effort on the one question of the effect of the screw in the waveguide, but it would be interesting to see what the effect is at lower frequency (5 kHz would be good, for a tweeter frequency), and if the effect is still there at this frequency, to determine how much lower in height the screw will need to be in order for the effect to decrease to less than .1 dB, or if it is no longer significant at this frequency, to determine how much greater in height the screw needs to be in order for the effect to return to a level of significance. That said, it goes without saying that other people like me will take advantage of your kindness with this kind of question, "What happens if you change X a little bit, and then if you change Y a little bit, and heh, I'll wager that you change both X and Y together ..." There will be no end to this kind of thing, and you may find it practically necessary at some point to simply ignore these kinds of requests without so much as an acknowledgement!
 
Top Bottom