Baffle diffraction step - is this right?

[buzz1939]

Thisarticle is incredibly useful and I've quite enjoyed going through it
http://www.t-linespeakers.org/tech/b...tep/index.html

But there's one part where I'm not sure they got the words right.

It states:

"As an approximation, the rise begins at the frequency whose wavelength is 1/8 the smallest dimension of the baffle. This dimension is typically the width of the loudspeaker since most are tall and narrow. Using the same 18" baffle as in the previous example, the response would begin to rise at [1/8 * (13560/18)], or 94 Hz. Also, the maximum amplitude is attained at a frequency whose wavelength is twice the smallest dimension of the baffle; in this case [2(13560/18)], or 1.5 kHz."

But of course higher frequencies have shorter wavelengths. So it seems like it should read:

whose wavelength is eight times the smallest dimension of the baffle

and

whose wavelength is one-half the smallest dimension of the baffle

Which makes sense to, that the loss from baffle diffraction has disappeared by the time you get to the point where the frequency is high enough that one wavelength doesn't reach from the center of the baffle to the edge, and all the energy will radiate forward.

If there is a problem, it's probably due to the way the formula is used - as that paragraph uses the rearranged version f = 13560/l and that's where it gets into trouble. Might be a bit clearer to use

13560/(18 * 8)

and

13560/(18 * 0.5)

[Svante]

Yes, that seems wrong. For a 547x1000 mm baffle the baffle step can look like any of these curves. The curves correspond to different placements of an 250 mm driver.

At 94 Hz the increase is some 2.5 dB for all placements, but at twice that frequency/half the wavelength, the curves differ by ~1.5 dB. The wavelength can be seen on the top of the figure.



Clearly, these rules of thumb should not be used for serious construction. OTOH, that is OT...

[CeramicMan]

A rise 'beginning' at 8 times the largest dimension of the baffle would make more sense, even if a different approximate value such as 4 or 16 would be more appropriate.

The largest dimension affects the lowest frequencies, and vice versa.... i think...

[buzz1939]

Thanks. Dave, who runs that site, says he's going to have a look at it and change it.

[Svante]

Originally posted by CeramicMan
A rise 'beginning' at 8 times the largest dimension of the baffle would make more sense, even if a different approximate value such as 4 or 16 would be more appropriate.

The largest dimension affects the lowest frequencies, and vice versa.... i think...

No, actually, for a baffle that is tall and narrow, it is the narrow measure that has the largest impact. Try The Edge i my signature and you will see.

[buzz1939]

I think the approximation issue is where the difference lies here. The simple answer is to just use the shorter dimension as a good approximation of the overall behavior. I could imagine that, as Svante suggests, it's the shorter dimension that has the most impact.

But I'd think a really detailed analysis would show the long dimension has some impact, along the lines CeramicMan suggests. Unless it's really overwhelmed by the impact of the shorter dimension - you'd think you would see some rise show up as the long dimension starts to shift some lower frequency energy forward.

Using the short dimension seems to provide most of the answer, though.

(As a newbie, my posts are moderated, so my last comment was posted before Ceramic's but showed up long after I posted it.)

[DSP_Geek]

Originally posted by Svante


No, actually, for a baffle that is tall and narrow, it is the narrow measure that has the largest impact. Try The Edge i my signature and you will see.

It turns out a tall narrow baffle ends up having a step about an octave below a square one with the same width.

I first heard this on a baffle where I'd calculated the step using only the width and heard a lump in the midrange, which was confirmed by simulations using The Edge. Lovely program, can't endorse it enough. Anyone who designs without it is stumbling in the dark.