[JeffB]

I just wanted to see if my understanding of this is correct.
dipole peak frequency is
p = C/(2d), where C is the speed of sound and d is the distance between 2 sound sources.

If I have an 18" wide baffle, then my dipole peak should be about
peak = (1126 ft/sec * 12 in/ft) / (2 * 18 in) = 375.33Hz.

The response rolls off at 6db per octave below the dipole peak.
47Hz is 3 octaves lower.

So if I had a 15" driver on this baffle, and the sensitivity of the driver was 96dB/1watt, then output at 47Hz would be 78db/1watt
and output at 23.5Hz would be 71db/1watt?

In order to get a flat response, the woofer will need to cut output at 6db per octave as frequency increases.  If the cut is started at 47Hz, then there will be a flat response from 47Hz to 375Hz at 78db/1watt.
In order to mate this with a 3" 87db/1watt speaker, then (87db - 78db) = 9db, would require about 8 watts to the 15" for every watt fed to the 3" ?

[rjbond3rd]

MJK mentions that it will be 18db/octave:

"For an OB speaker system the bass will roll-off at a rate of 18 dB/octave, this is the sum of a 12 dB/octave roll-off for the driver and a 6 dB/octave roll-off for the open baffle. "

http://www.quarter-wave.com/OBs/OB_Theory.html (first article, page 2).

[Davey]

Yeah, that's only if you get down into the second-order roll-off area of the driver itself.  Then it becomes 6db/octave plus 12db/octave which yields the 18db/octave ultimate roll-off.

Jeff,

Remember that the dipole peak created is (theoretically) +6db relative to the nominal sensitivity.  So, in your example, your dipole peak would create an equivalent sensitivity of 102db/1watt.

Cheers,

Dave.

[scorpion]

Jeff, if you had a circular baffle with 18" diameter (equal distance to the edge everywhere from the center), and placed the speaker in the center, the formula would work out like: 1126*12/(2*9) where 9 inch radius is the effective distance between the sources. So for say a rectangular baffle it will be approximatively the baffle width to divide with.

John Kreskovsky has it beautifully on his web: http://musicanddesign.com/Dipoles_and_open_baffles.html .

You can simulate rectangular and other size baffles also well with the Edge program available here: http://www.tolvan.com/index.php?page=/edge/edge.php .

For reallife simulations with different loudspeakers there are programs available from Martin J King and John Kreskovsky that will tell at least almost the truth of the performance of a given speaker on an Open Baffle.

/Erling

[JeffB]

Ok, this has been very helpful so far.
I have never heard of the 12db driver roll-off.
With a little googling I found this statement

"a driver in an (truly) infinite baffle or in a sealed enclosure has a rolloff of 12dB/oct. below the resonant frequency".

So does this 12db driver roll-off only come into play below Fs on an open-baffle?

[JeffB]

So my dipole peak is 1 octave higher than I computed.
But the dipole peak is 6db greater than nominal, which I didn't account for.
And the 12db driver roll-off starts at Fs?
I would have to look up later, but I think Fs on my speaker is around 36Hz.
Thus, I still have the same 47Hz at 78db/1watt ?

[scorpion]

Yes, you are on the right track. 6 dB dipole roll off and 12 dB speaker fs roll off,
summing to 18 dB eventually.

/Erling

[JeffB]

Now, I am playing with Edge, and the open baffle compensation f1 and f2 settings have me confused.
I entered the following data.
x: 457mm,  y: 900mm
corners: 4, speakers: 1
sources: 100, source density: 1
shape: circular, size x: 381mm
mic dist: 3m
open baffle: checked
15" speaker drawn near center bottom of baffle
If I use f1:30Hz and let program suggest f2: 153.9Hz, then edge shows frequency drop off from about 700Hz to 30Hz from +11db to +3db.  There is definitely correction applied, but the fall-off looks rather steep to me, but the over-all graph looks "balanced" for lack of a better term.
If on the other hand I enter f1: 1hz and f2: 270Hz, then I get a flat bass response from 20Hz to 700Hz at +11db that is +5db above baseline.
This second choice looks more desirable as a starting point for a cross-over to a 3" driver around 700Hz.
Or am I completely off-base here.

[scorpion]

Edge is just a baffle effects simulation tool for comb filtering. No room effects like floor bounces are taken care of. And ideal piston working speakers are supposed. But it is great to calculate the dipole peak with.

If you would like to be able to simulate more realistic with your chosen speakers the ABC-Dipole programme is now free to download and play around with.
You will find it here: http://musicanddesign.com/A_B_C_Dipole.html .

/Erling

[johnk...]

Jeff,

The dipole peak is Fp = C/(2d) , the  dipole = monopole (fe) frequency is given a C/(6d). So at fe the dipole will have the same sensitivity as the same driver in an infinite baffle assuming 2Pi sensitivity is quoted.

Best regard,

John  k...

[johnk...]

Let me add that the basic dipole response (assuming a perfectly flat source, is identical to a 1st order HP filter with 10dB gain and with corner frequency the same as the dipole peak. if you superimpose that on the drivers 2Pi response you basically have the dipole response, assuming driver directionality isn't a factor at higher frequencies.  This is all discussed on ABC Dipole, which as pointed out, is a free download from my site.
http://www.musicanddesign.com/A_B_C_Dipole.html

I don't know if the software works withnewer version of Excel, but the chm Guide should be fine.