[Ant_222]

Hello,

I was researching (and thinking about) the physics of wave diffraction, especially when applied to that of low-frequency spherical sound waves around an OB enclosure, and found a program called Edge that can model either the baffle step effect for a closed-box loudspeaker or a frequency response of an open baffle (all in free air).

And the models in this program are quite close to my own calculations which led me to a design that has a strangely smooth response (according to Edge) which you can see on the attached pic.



Before getting a measurement mike and building this design, I'd like to ask you: Has anybody tried this program and compared its results with actual measurements in free air?

Thanks in advance,
Anton

[Rudolf]

Anton,
I believe you are well aware that this is a simulation of the baffle response only. Every peculiarity of the driver response (on infinite baffle) needs to be added on top of the simulation you have shown.

It is not too hard to come up with a nice response diagram in EDGE - if you only look on axis. But such a sim is useless until off axis simulations show a benign behaviour too.

This is a tiny list of correction values for mic distance and mic offset from the speaker center which helps you to get off axis simulations:


Let us have a look at those sims too. 

[Ant_222]

It is not too hard to come up with a nice response diagram in EDGE - if you only look on axis. But such a sim is useless until off axis simulations show a benign behaviour too.
<...>
Let us have a look at those sims too. 

Never thought of this. I wasn't optimizing for it. But here they are anyway.
Note that my initial simulation was for a far-away point, while these are for 1 m.

http://www.filefactory.com/file/b133fa8/n/0_7.zip

(Update: the link is broken because I used a dumb file hosting. Will update it in 10 hours)

Anton

EDIT:
I haven included anything steeper than 45 degrees.

Oh, and since you have asked me, you do agree that Edge is realistic?

[Rudolf]

Sorry, I could not download that .zip file. Do I need to "join in" at the filefactory?
And yes, as a baffle simulator EDGE should be as realistic as it gets. AFAIK the basic algorithm in EDGE is the same as in all better programs that sim baffle shapes. Once I compared the results of Linkwitz diffraction test http://www.linkwitzlab.com/diffraction.htm with EDGE and found the simulations quite realistic.

[Ant_222]

Sorry, I could not download that .zip file. Do I need to "join in" at the filefactory?

No, you don't have to register for that. But I just came to recheck my file and it reported that my "download slot has expired". FileFactory is very bad. Sorry for the broken link.

Can you suggest a good file hosting so I can upload my file there?

Anton

[Rudolf]

Can you suggest a good file hosting so I can upload my file there

I'm loading my own files to my personal web space, so I haven't looked outside. But I would like to keep that space to my own files only.

Why don't you want to show those sims as pictures on audiocircle?

Rudolf

[Ant_222]

Why don't you want to show those sims as pictures on audiocircle?

I didn't think it would be fine to post so much images. Besides, it was more convenient to pack them all into an archive instead of posting one by one.

I will post them today when I get home.

Anton

[Ant_222]

ANGULAR RESPONSE AT 1 METER:

0 Degrees: http://img46.imageshack.us/img46/476/97547458.png
10 Degrees: http://img688.imageshack.us/img688/1774/66058415.png
30 Degrees: http://img145.imageshack.us/img145/7459/64102040.png
45 Degrees: http://img689.imageshack.us/img689/1544/95166014.png

Anton

[Saurav]

There's a little 'lock' or 'freeze' button in the EDGE graphs, using which you can put multiple curves in the same image. Then you'll only have to upload one image.

And to answer your original question, I use EDGE a lot, as as Rudolf said, barring the issues of driver geometry and resonances between the baffle / magnet / cone, I've found the sims to be pretty accurate.

[Ant_222]

Thank for your feedback, Saurav

You said you are using Edge to help you simulate "resonances between the baffle / magnet / cone". I wonder, how can EDGE do it? Doesn't it model the baffle as infinitely heavy and rigid?

Anton

[Saurav]

Here's what I said:

as Rudolf said, barring the issues of driver geometry and resonances between the baffle / magnet / cone, I've found the sims to be pretty accurate.

Sorry I wasn't clear. What I was trying to say was, EDGE does *not* simulate the baffle / magnet resonance issues, it assumes ideal radiation in front and behind the driver. It does take driver directivity into account though. Also, what I'm talking about isn't the baffle itself vibrating. It's the air trapped between the baffle and the magnet. At least, that's what I think it is. With my current midrange driver, I see a big change in response if I mount the driver on the back of the baffle (which I assume allows the rear radiation to 'breathe' better). EDGE does not simulate that aspect of the performance.

Like any other tool, it's useful as long as you understand its limitations. There's no substitute for building and measuring your design

[Ant_222]

It does take driver directivity into account though. Also, what I'm talking about isn't the baffle itself vibrating. It's the air trapped between the baffle and the magnet.

I as understand, directivity in EDGE simulations results from:
1. The essential directivity of sound waves (with higher frequency),
2. The use of multiple point sources.

... Air between baffle and magnet? You mean between cone and magnet? But I think this volume does not depend on whether it's mounted on a baffle or an enclosure... The main sources of vibrations are mechanical vibration of the loudspeaker and the effect of sound waves reflecting from the baffle.

With my current midrange driver, I see a big change in response if I mount the driver on the back of the baffle (which I assume allows the rear radiation to 'breathe' better). EDGE does not simulate that aspect of the performance.

Interesting. Do you mean it is normally mounted on the front side? What's the difference in the response? And how thick is your baffle (compared with the loudspeaker's diameter)?

Anton

[Saurav]

You can read all about it here:

http://www.htguide.com/forum/showthread.php4?t=33242&page=2&pp=35

Post #58 has Rudolf's theory on what was going on with my measurements.

[Rudolf]

Interesting. Do you mean it is normally mounted on the front side? What's the difference in the response? And how thick is your baffle (compared with the loudspeaker's diameter)?

Anton,

the following is just an example of what can happen. Most every driver will react differently - so you can´t take general conclusions from this.
Mounted in the usual way the Visaton FRS 8 will show a nasty resonance to the rear at 2 kHz:



Mounting the driver to the back improved things remarkably:



I turned the already completed baffle 180° and mounted the driver to the flat side, with the taper acting as a kind of waveguide:



This led to a much better response all around. You can see the effect of the waveguide in the beaming of the response to the front (Winkel vorn):

[Ant_222]

Rudolf, thank you very much for sharing your experience. I see that you're using a small-diameter loudspeaker in a thick baffle, which makes the positive effect of flaring the hole very prominent.

But I did not quite understand what you did in case 3. Did you taper your "already completed" rectangular baffle? You wrote you turned it 180 degrees, but around what axis?

The addition of a taper made the response smoother because it significantly increased the range of paths from the back side of the driver to the microphone — as I think.

Anton

EDIT: May I ask you what hardware and software you use to measure the frequency response?

[Rudolf]

I see that you're using a small-diameter loudspeaker in a thick baffle, which makes the positive effect of flaring the hole very prominent.

My measurements don't show any comparison which could show "the positive effect of flaring the hole". What I show is the difference between a front mounted driver having a taper on the back of the cutout and mounting the driver from the back.

But I did not quite understand what you did in case 3. Did you taper your "already completed" rectangular baffle? You wrote you turned it 180 degrees, but around what axis?

Sorry, I obviously didn't use the proper expressions. What I wanted to say was that I turned the (tapered) baffle 180° around the vertical axis to use the driver cutout, that I had already flared at the rear side, as a sort of waveguide in the front.

May I ask you what hardware and software you use to measure the frequency response?

Those measurements where taken with JustOct. You can`t buy that program - you need to subscribe to a german website to use it. I have changed to ARTA now for most of my measurements. Hardware are a USB Soundblaster or the sound chip in my notebook. Microphone is a Monacor ECM-40.

[Ant_222]

My measurements don't show any comparison which could show "the positive effect of flaring the hole". What I show is the difference between a front mounted driver having a taper on the back of the cutout and mounting the driver from the back. [...]
Sorry, I obviously didn't use the proper expressions. What I wanted to say was that I turned the (tapered) baffle 180° around the vertical axis to use the driver cutout, that I had already flared at the rear side, as a sort of waveguide in the front.

Now I hope I get your point:
Picture 1: Driver mounted from the flared end on the non-tapered baffle.
Picture 2: Driver mounted from the flat side on the non-tapered baffle.
Picture 3: Driver mounted from the flat side on the tapered baffle.

...Then your measurements show that it's better to mount the driver on the flat side of a flared cut-out.

Concerning your experiments, did you use sine signal to measure the frequency response?

Anton

[moij]

Hello

When building OB speaker I find that Edge simulation of OB is not very accurate.
I had U-frame baffle with effective width/hight 1500 mm and 18" woofer. With measurement I get first dipole null on about 250 Hz, with Xlbaffle.xls about 240 Hz, but with Edge first dipole null will be at 400 Hz.
What can be the case, is Edge really so inaccurate or I had made error somewhere?

[Ant_222]

When building OB speaker I find that Edge simulation of OB is not very accurate.
I had U-frame baffle with effective width/hight 1500 cm and 18" woofer. With measurement I get first dipole null on about 250 Hz, with Xlbaffle.xls about 240 Hz, but with Edge first dipole null will be at 400 Hz.
What can be the case, is Edge really so inaccurate or I had made error somewhere?

Edge can only simulate flat baffles. Converting a U-shaped baffle to a flat one is not as simple as calculating its "effective" width and height. You'll have to derive an equation, in polar coordinates, describing the dependency of the shortest path from the back of the speaker back to the speaker plane on the front side. With flat baffles this is the same as the equation of the shape, but with all other types it is more difficult, and Edge will not do this for you.

Anton

P.S.: Did you get a high-Q (0.7..0.8) 18" woofer?

[moij]

Detailed conversion to flat baffle is not easy but even if measuring  "effective" width and height with tape can not be error so big - nearly 2 times. Also I get very close result with microphone and dimensions measurement used for Xlbaffle.xls.
Also question can be: If I use same 1500x1500 mm baffle and speaker in the center in Edge and in Xlbaffle.xls why first dipole null is 240 Hz in Xlbaffle and 400 Hz in Edge?

Yes I had 18" with Qts 0.7 -  http://www.lautsprechershop.de/pdf/selenium/selenium_wpu1805_x.pdf