Home Speaker Asylum

General speaker questions for audio and home theater.


You are correct that offset, baffle width and driver's bandwidth/coverage play a role in diffraction - for flat baffles on rectangular boxes. And that it can be managed to some extent using those factors. I disagree that it can be made to disappear.

"you'll find that the responses in figure 15,16, and 17 (the rectangular box) have the best response. The sphere with offset tweeter (figure 6) is not a practical design and you didn't specify an offset driver which makes a significant difference as the centered driver on a more practical circular baffle (figure 8) is the worst case (which is what I was referring to). "

You have made a significant error about spheres and the drivers in the Olson paper. I suggest that you read the referenced 'OLD' paper by Olson, all of it, a couple of times. If you had read it you would not have made this rather revealing error.

The reported tests used the same very small high quality driver, to ensure that it was a pistonic source. The pictures of the sphere and the hemisphere throughout the paper are offset so that you can see the single driver clearly. So Fig 6 is not a special case with an offset tweeter to give that very smooth FR.

Put a single driver on a sphere - anywhere - and it will be centred on the sphere by definition (as would any infinitesimal dot be,) no!? And, you would measure it on axis, as with all the other shapes covered.

Fig. 6 - the sphere with a single - centred - driver, has the smoothest FR / best diffraction behaviour. Despite the driver being - inevitably - centred. So that rule about centred drivers falls away - in the case of a spherical enclosure only.

Because a sphere is not a circular baffle as it does not have any edge. It will have a very smooth BDS dip, as predicted by the spherical diameter formula, so that compensation for it is easy to build for, or you can forget about it, as my 2-way spheres do. Room-gain helps.

Even a truncated sphere with a small sliced (circular) baffle is measurably superior to a 90-degree edge box and audibly so.

In the late 1970s. Enclosures of the same Vb, same tuning, stiff laminated walls for the box (3mm hardboard pressure glued to the inside walls, which took days) same amount of BAF. Gave the box every chance by offsetting the driver (not done on the sphere) and making the baffle area narrower than that of the sphere, and maximally dissimilar dimensions, sharp edges but. The mass of the enclosures were matched using a concrete block glued underneath the box encl, so motor reaction effects were at least similar. Same screws, same plastic rawl plugs. The same WR 5inch Coral Flat 5 was used.

Measured FR? no contest. Listening? not in the race. Measurements done at ANU. More than 30 years ago.

Offsetting the driver on even the small circular baffle on the sphere did help a bit more, so the production model 2's had that as well.

Now, the audible superiority back then may have had an additional cause - that curved enclosures are quieter and are self damping. The test did try to minimise the differences there, and the enclosure talk of boxes just is a given.

Practical / real spheres then were truncated and thus did not have the predicted single/audible internal standing wave either. Just in case you want to bring that up. Neither did Gallo's spheres. Very good tight bass though.

It is plain to me that wide radius curved baffle edges are best. Probably not practical. Curved edges are better than sharp, and felt helps, a lot.

LBNL John Dunlavy 'respected' both the Model 2 Audiosphere with the single Coral Fl5, and the 2-way Model 3. Told me so himself.

Mind how you go.


Tim Bailey

Skeptical Measurer & Audio Scrounger

Edits: 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12 04/18/12

This post is made possible by the generous support of people like you and our sponsors:
  Audio Hoarder  

Follow Ups Full Thread
Follow Ups


Post a Message!

Forgot Password?
Moniker (Username):
Password (Optional):
  Remember my Moniker & Password  (What's this?)    Eat Me
E-Mail (Optional):
Message:   (Posts are subject to Content Rules)
Optional Link URL:
Optional Link Title:
Optional Image URL:
Upload Image:
E-mail Replies:  Automagically notify you when someone responds.