Unity horn build

Probably that's why I have way better HF when I move 1-or even 2 meters a side then when I sit in a front of speaker.

The idea of the increased flare at the mouth is to reduce edge diffraction. Ideally the mouth termination would be curved but that is tricky with large ply horns. The hi-fi horn designers, where large round overs are less impractical suggest that ideally the radius of the mouth termination should be equal to one wavelength of the lowest frequency. Others consider that a radius of ¼ wavelength is OK. One would expect that the mouth geometry would affect the lower and mid-band of the horn more than the higher frequencies.

The differences in the high band can be due to the different degree of coherence presented by the wave front at the compression driver’s exit. This depends on how the phase plug geometry has been optimised.

At very high frequencies a high-frequency driver can change its radiating figure quickly depending on frequency. It is possible for it to change from a plane wave, to a ring radiator or to a point source whose effective dimension is smaller than the throat. These changes are reflected in the wave front curvature, which can lead to driver>throat mismatching, and can be a source of internal diffraction and distortion as well. Changing the drive unit could significantly alter the amount of fingering a given horn exhibits.

If you set up a grid across the mouth of the horn and measure the pressure in each cell, the pressure variations show up like the room mode patterns, with high and low pressure regions that change with frequency.

Keith Holland (Southampton Uni) did some work on obstructions within the horn and if I remember correctly something on the left hand wall creates an artefact on the right hand side. With a unity style horn the mid apertures will be a source of problems.

As with most design problems, it is a case of deciding which features and compromises best fit your priorities.

I would imagine that an array of unity horns would behave in a similar manner to a multicellular horn. It would be expected therefore that there will be a narrow lobe at the frequency where the width of the array is about one wavelength. Above that the pattern will widen and at high frequencies there will be some fingering

Just because your suggestion might not be the correct one in this specific case does not mean that it is stupid (unless it is just stupid generally ). Diagnostics at a distance is always a bit hit and miss. Please share, it might be of use to other forum members.

steve, been reading this page about mouth termination

Clifford A Hendricksen "The Manta-Ray Horns" (co-authored with Mark Ureda), Journal of the AES, vol 26 no. 9, September 1978.

thanks DJK, managed to download the manta-ray paper as well as a couple of other interesting looking ones that reference it.