Acoustic reflectors and piston drivers

[markus76]

Here's some data for the equalized "double cone":





The problem with this device and/or driver is lots of resonances:



Not sure if this is worth pursuing.

[vacuphile]

Markus,


Have you done some distortion measurements as well, comparing the speaker with and without deflector? I did a quick test some pages ago that pointed into that direction, but it would be nice to have a second measurement from a different setup.

vac.

[ScottG]

Originally Posted by markus76

Here's some data for the equalized "double cone":





The problem with this device and/or driver is lots of resonances:



Not sure if this is worth pursuing.

It's remarkable for a single full-range driver, but I don't think it's ever going to be "it".

On the other hand I think some of the effort of the design could well be beneficial.

For instance the overall uniformity of pressure loss off-axis from 1-4 kHz is pretty good, not excellent (that 1.8 kHz artifact needs some work), but still pretty good when you consider that this range posses difficulties with larger drivers.

Below 1 kHz (and particularly below 700 Hz) you can improve the uniformity with a leaky vent on the same baffle as the driver but at its "rear" (considering it's a radial design). This will of course limit it's usefulness as freq.s decrease.

Above 4 kHz - well, that's always going to be a problem (at least with regard to a uniform drop in pressure off-axis, and overall linearity). I think you'll have to transition to another driver(s) to get something more uniform. (..even then though, the response is still pretty good on an absolute-level to more than 10kHz.)



I wonder what it looks like when you eq. it on the dominate axis you listen to?

[markus76]

I let a DCX2496 do the heavy lifting (878Hz, -9.6dB, Q1.8; 1.88kHz, 11.2dB, Q2.83; 2.44, -10.5dB, Q7.9) and then run it through Audyssey MultEQ which does a pretty good job at speaker correction. After that, the speaker sounds pretty good.

A smaller fullrange driver would probably help with the 1.9kHz region but then we're talking about a two-way design which brings in additional complexity.
There are also traditional two-ways that are even more omni: audio blog: Review of Polar graphs

[6.283]

Here are some omni insights hot off the press. Enjoy the lengthy text with your favorite beverage.

http://www.2pi-online.de/html/interim_conclusions1.htm

Oliver