Acoustic Horn Design – The Easy Way (Ath4)

[gedlee]

I don't believe anyone would prefer a dip in a blind test systematically

They don't. It's a personal opinion.

 

[diyiggy]

I indeed can not generalize it for all the speakers ! This thing is at sweet spot on a triangle where side are longer than the base which is not an ideal listening position !



So at 1 m a speaker should be flat between 1 k and 20 kHz (+ has the constant directivity goodies) ?

 

[mabat]

Back to the dome tweeter (T34-B) waveguide - I made the following one a little smaller (now it is 346 x 302 mm; the previous one was 450 x 390 mm).
The beamwidth actually widens with frequency a bit (although it is pretty constant for -6 dB). Makes me wonder if a slightly more beaming (and more loading) profile could/should be used in this case. I'll definitely give it a try.

(Again this is for an infinite baffle.)

 

[fluid]

storage energy is not where they perform best. Why not the peerless DX25 or DX32 instead ?!

I know you have trouble getting the model numbers of speakers correct but I think you are talking about the DA25 or DA32 as they use Corundum as the dome material.

Here is a comparison of the T34B and the DA25 from Hificompass as the graphs can be directly compared. I don't see a real issue with energy storage with either. DA25 left T34B right.

DA25 gets worse below 3K.

 

[solenskienr]

The beamwidth actually widens with frequency a bit (although it is pretty constant for -6 dB). Makes me wonder if a slightly more beaming (and more loading) profile could/should be used in this case. I'll definitely give it a try.

You could change back to an OS profile, with an opening angle 5-10degrees less than the conical angle

 

[diyiggy]

Yeap Fluid,


I mix up the DA25BG (aluminium) with the DA25TX (corundum as you stated out)


Edit : on the graph I find the left graph better for my tastes ! good from 2 k Hz ?!

 

[Ro808]

Back to the dome tweeter (T34-B) waveguide - I made the following one a little smaller (now it is 346 x 302 mm; the previous one was 450 x 390 mm).
The beamwidth actually widens with frequency a bit (although it is pretty constant for -6 dB). Makes me wonder if a slightly more beaming (and more loading) profile could/should be used in this case. I'll definitely give it a try.

(Again this is for an infinite baffle.)

Would you care to show the mesh of the tweeter-horn interface?

 

[mabat]

You could change back to an OS profile, with an opening angle 5-10degrees less than the conical angle

That's quite insightful!

I don't know exactly why, but with OS it gets worse around 10 kHz, even for a mild presence of it. This is for an opening angle 5 deg less than the conical angle (and it gets worse with increasing this difference). Maybe there's another way. BTW, non-axisymmetric shapes are far less sensitive to the "surround problem", it seems.

Anyway, I think this is a pretty good starting point.

 

[mabat]

In the original OS/CD version it behaves like this:

 

[mabat]

2" Aurasound NSW2-326-8AT, the drivers S. Linkwitz used in his Plutos, in the same waveguide (modified for a different throat diameter, of course).
It would make a nice contest - to improve this

; Aurasound NSW2-326-8AT

point  p1 -6.5 0  4
point  p2 0 16.5  2
point  p3 3 19.5  2
cpoint c1 17.69 0
cpoint c2 0 19.5

arc p1 c1 p2  1.0
arc p2 c2 p3  0.75
arc p3 c2 WG0 0.25

 

[mabat]

This is slightly better (reduced the value of "n"):

 

[mabat]

T34-B again, in a smaller waveguide (290 x 254 x 70 mm) with a little wider beamwidth (still only an infinite baffle):

I could live with this.

 

[mabat]

Should I make a public domain version for DA32TX00-08?

 

[mabat]

Anyway, the DA25TX (for which I have dimensions measured) is not as nice, seems to start beaming above 10 kHz in the same waveguide. It would probably require a different mounting interface and the same would probably hold for the DA32TX. The surround is convex here and not touching the wall of the WG.

 

[danibosn]

Hi


Marcel (or somebody else), please,


how do you get "Converted from PolarMap_H" graph.
And, is it possible to export data from Vacs and import to Vituix? If it is, how to do it?


Thank Danijel

 

[Boden]

Could you also plzze plzze try to sim the Peerless 830970?

Platon Measurements

 

[mabat]

It seems that the concave surround is a good thing. The following is the same as the T34-B above but with the surround made convex. Even though not a night and day in this case, the effect is definitely recognizable. So the recommendation at the moment would be to look for a tweeter with concave surround that could smoothly join the waveguide (which is apparently even more important), I just don't know many.

I'll also try to cover the surround with the waveguide somehow but I'm afraid that will not improve things that much. So far I haven't tried to model any ring radiators.

danibosn, Boden: I'll get back to you.

 

[diyiggy]

Anyway, the DA25TX (for which I have dimensions measured) is not as nice, seems to start beaming above 10 kHz in the same waveguide. It would probably require a different mounting interface and the same would probably hold for the DA32TX. The surround is convex here and not touching the wall of the WG.

Not sure as the DA32TX is falling faster off axis than its little sister ! In fact perhap with the same horn the 15k peak will looks good with the da32 ?!


Yeap a little disapointing the da25tx vs the T34...

 

[Ro808]

Regarding the T34-A; an (unmodified) OS-throat is probably not the best interface for the concave surround. It's basically a cavity that's extended axially by the OS-profile and likely the cause of sub-optimal performance in the treble range.

 

[fluid]

how do you get "Converted from PolarMap_H" graph.
And, is it possible to export data from Vacs and import to Vituix? If it is, how to do it?

You click/select the polar graph you want to convert in VACS and then choose

Graph - convert curves <-> contour from the menu that will generate another graph. To get those curves out to Vituix select all the individual curves at the bottom of the graph. Select the first one then press and hold shift and click on the last one, they will go darker when selected. Then go to export the data from the menu, should default to Amplitude and phase and choose a file name.

Each curve will be exported separately in a numerical sequence.

To get them into Vituix they need to be renamed to the Arta style or something similar. I made a script to rename them to a format that will work using advanced renamer.

The script was attached in this post if it helps.

Acoustic Horn Design – The Easy Way (Ath4)