Back to presence and rationality. There are many reasons to build an omni based on the old hat trick with a cone or hyperboloid reflector:
- Full range drivers have benefited considerably from recent innovations in driver technology. So there was some noteworthy progress the past one or two decades. Today, the DIY market is full of high quality FR units (and sometimes you have the impression that DIY magazines don't review anything else anymore).
- It can be build easily and cheap (even with a passive x-over, but I won't bother).
- The reflectors are readily available.
- You save a cross over with its lobing in the all important midrange altogether (it is planned to cross over in the transition zone at ~200Hz).
- It is quite easy to keep structures small compared to wavelength. The full range driver is acoustically small on the low end of its operating range. With increasing frequency the sound emerges more from the middle of the cone. Now the tip of the reflector, which is acoustically small in the mids and low highs, becomes the primary sound source for the listener.
- The horizontal polar response is theoretically perfect (although kind of forced by the reflector).
- The design offers widest possible horizontal dispersion. According to my experience, "narrow" dispersion controlled or constant directivity from typical speakers with waveguides is inferior to dipolar and omni polar speakers for recreational listening in acoustically live sounding but otherwise simple rooms of “regular size” (long sentence but I would like to emphasize that I have no problem with waveguides or speakers using them. I just have no application for them).
The questions before measurements and listening tests can provide evidence, are:
- Is the vertical polar response good enough, can it keep up with the horizontal dispersion and response ?
- Are there any unplanned problems or side effects other than diffraction issues from the tube (transition from 4Pi into 2Pi) and the reflector (transition from 2Pi to 3/4 Pi) ?
- How is the imaging of the speakers ?
- The speaker will sound too bright if the on-axis response is flat. What equalization is required ?
From experience it is clear that SPL output capability matters a lot in order to get a live feeling. Too little SPL, no live feeling; a simple relationship once you have understood the significance. In this case I have to sacrifice some output capability in order to be able to drive the other two parameters dispersion (wide) and directionality (constant but more or less absent) to their "extremes" because I want to learn about them.
If the results are satisfying then there are ways to increase the output capability e.g. by driver choice and x-over point. But until then I am using lowest cost but high quality drivers (although this has nothing to do with the Greek austerity package). The FR is a Vifa 9BN 119/8.
It does not make a lot of sense to use a heavy cone - stiff compliance - powerful BxL - and so on - sub woofer type of driver in conjunction with such a filigree full range unit. Therefore, it’s going to be a real classical woofer for the first three octaves. In order to balance the power handling of the units, an 8" woofer should be sufficient (taking equalization for a CB into account).
All this translates into the following...
Primary design goals:
- widest possible horizontal dispersion
- constant horizontal directivity