What you show are basically two different OSSE waveguides, there's nothing exponential about it -
The left one in your picture has a conical extension (cylindrical in this case), that has been the main subject of the last several pages of this thread...
It can work very well, as has been shown, but at the moment it's not clear exactly how it's possible and what are the requirements.
The left one in your picture has a conical extension (cylindrical in this case), that has been the main subject of the last several pages of this thread...
It can work very well, as has been shown, but at the moment it's not clear exactly how it's possible and what are the requirements.
Yes, I`m reading it through and trying to keep up, excellent thread!The left one in your picture has a conical extension (cylindrical in this case), that has been the main subject of the last several pages of this thread...
It can work very well, as has been shown, but at the moment it's not clear how/why.
Here is a second sketch to illustrate what I mean.
Left pic, seg1-cylinder...seg2- OS-SE
Right pic. seg1-conical...seg2 OS-SE
Regarding directivity (polars, diffractions, etc.), you can model/simulate both with Ath+ABEC/AKABAK right away, that's easy and it should answer all you questions. As for the "loading" aspect, that's what still leaves some questions open...
It's so complex topic that there are really no short answers. It seems that the actual drivers can play an important role and they should be included in the model with all their complexities (which is not readily available, unfortunately).
It's so complex topic that there are really no short answers. It seems that the actual drivers can play an important role and they should be included in the model with all their complexities (which is not readily available, unfortunately).
Last edited:
Exponential horns are best regarding acoustical input impedance at the throat. This concerning also the length of the horn - their dimensions are smaller for the same low frequency cut-off. Normalized acoustical impedance, A + jB, looks like this:
(fo is the horn critical frequency, the horn does not transmit power below fo, this applies to infinite horn. For a finite length, there is always some transfer below fo)
Conical horns are one of the worst types of horns.
A great deal of the reason that the ATH horns work so nicely is that there's a smooth transition from the diaphragm of the tweeter to the horn, and a smooth transition from the mouth of the horn to the room. Check out the pics of the "cutting edge" aka "CE" waveguides for an example.
Conical horns feature neither, and should be avoided IMHO. I think people build conical horns because they're easy to build and that's about it.
EDIT:
Note that some of the ATH waveguides, as well as the oblate spheroidal waveguides look similar to conical. Devil is in the details, the throat and the mouth geometry in particular.
You can evaluate this yourself with ATH and ABEC of course, and tweaking the models is a great way to learn what works and what doesn't.
A great deal of the reason that the ATH horns work so nicely is that there's a smooth transition from the diaphragm of the tweeter to the horn, and a smooth transition from the mouth of the horn to the room. Check out the pics of the "cutting edge" aka "CE" waveguides for an example.
Conical horns feature neither, and should be avoided IMHO. I think people build conical horns because they're easy to build and that's about it.
EDIT:
Note that some of the ATH waveguides, as well as the oblate spheroidal waveguides look similar to conical. Devil is in the details, the throat and the mouth geometry in particular.
You can evaluate this yourself with ATH and ABEC of course, and tweaking the models is a great way to learn what works and what doesn't.
BTW, Faital HF1000 has a cylindrical (0°) exit section. Normally, I would probably avoid such driver, but it could be interesting to try to insert different lengths of a straight tube between the driver and a horn...
