Geddes on Waveguides

Most compression drivers are tested on a terminated plain wave tube and not mounted on a horn. At least that is what I know from the CD designer that I know. Only after the driver has been built and tested will it be mounted on some horn and tested for on axis response and polar response, nothing really to do with the driver development though.

If only all gave plane wave tube curves on their data sheets. You can tell ones that don't when they show a nice flat high end, when the pwt or constant directivity measurement would show a knee in the passband where the response starts to curve downward by another 6db per.
 
Last edited:
These are the raw sensitivities of the B&C 15NBX100 and DE500 in the project posted above, around 20° off-axis. It's without foam plug. Earl is of course right. In this project the woofer is quite attenuated in a passive version (0dB@50Hz ... -10dB@500Hz) - the goal here was to have really a full-range speakers, so you have to do this to reach the low frequencies on it's own. I agree that it's probably not a good compromise. Well, it works.

BTW, thank you Earl again for all this work. It works and it works very well!

The program for the polar maps is my own. I do all the crossover work with it.
 

Attachments

  • BCs.png
    BCs.png
    22.1 KB · Views: 618
Can you explain a little bit why do you not recommend this, if we are considering scenario with a passive crossover, please?

Let me soften my statement to: "I would prefer to not lower the efficiency of the woofer." I can see that there might be good reasons to do that, but my preference is to not do so. I could envision changing this position.

I know you will be disappointed but it's just some interpolation of the input data. So the more angles you measure the more resolution you have. I don't remember exactly how it's done and I know it is not really correct but it helps me to see some things more easily anyways. It is still nice enough tool for me :)

Polar maps are an essential tool in good loudspeaker design. I will not take any design serious that does not show this kind of data. (Un-normalized as you and I do. Normalization obscures almost everything that one is looking for.)

That said, there are good and bad ways to do this. What happens when you interpolate polar data is that it tends to average out the diffractions because they are dips that change in frequency with angle - kind of like arcs in the response (the classic example of this can be seen in my simulation of a vibrating cap in a sphere in polarmap on my website.) Resonances, which do not change with angle get shown quite readily even with interpolation. This is why in your example the resonances are shown quite effectively, but the diffractions are not seen.

My technique avoids this problem since no interpolation is done in an angular sense. The radiation "modes" are extracted and then the polar response is reconstructed at whatever angular resolution one wants. There is some interpolation in frequency but none in angle (we have 1000's of frequency points, but typically only a few angular ones.) This technique is the same as that Klippel is currently using in his new system. He does this in three dimensions and I only bothered with two, but the techniques are otherwise identical. (And both derive from the work of Prof. Weinriech at Michigan in the 50's.)
 
Let me soften my statement to: "I would prefer to not lower the efficiency of the woofer." I can see that there might be good reasons to do that, but my preference is to not do so. I could envision changing this position.

Similar idealism about preserving bass efficiency here, but as you say, there are good reasons to try it- maybe you want a dB or two extra bump around Fc or want to manage amplifier load, or the padding can smooth the impedance to simplify XO design.

Pads integrate well into series crossover design, too, which some prefer (regardless of perceived/analyzed value of different XO topologies), always nice to have options.

In this case, I'd rather have a better top end, I generally prefer smaller diaphragmed drivers crossing higher for that extension/smoothness, but I'm still in my 30s, it may be less of an issue for our older friends with more HF losses in hearing.
 
IIRC (it's been a long time) adding series resistance to a woofer raises the Q, which usually requires a larger box.

-it's particularly useful for free-air/dipole midbass designs where the low-pass uses a cheaper inductor (smaller gauge windings with more resistance) to net a bit more gain lower in freq.. :) (..starting of course with a more efficient driver with a Qe typically around .5.)