Open Baffle Midrange Peak?

[croat47]

Hello,

I can not seem to post an image. But, I modeled the following on The Edge and Xbaffle.xls:

406mm wide
914 mm tall
81mm driver centered on width, located at 711mm.

With The Edge, I get a peak in the mid range (100-1300Hz). It goes up ~2dB by going to open baffle. Is this due to the reflections off the back wall? Something else?

In an Audio Kinesis' blurb from another thread about why their bipole is shaped the way it is, it seems that the rear wave cancellation/dip is explained (seen at 1800Hz in The Edge's model).

In Xbaffle, no midrange peak is seen, but a dip around 175Hz is seen.

What is the difference in the two models that creates this "discrepancy" between them?

[Rudolf]

Originally posted by croat47
With The Edge, I get a peak in the mid range (100-1300Hz). It goes up ~2dB by going to open baffle. Is this due to the reflections off the back wall? Something else?

The 2-3 dB emphasis in peaks and dips is due to the backside radiation of the speaker coming to the front. Since EDGE does not care about any walls or the floor, it can not simulate any reflections off any walls.

In Xbaffle, no midrange peak is seen, but a dip around 175Hz is seen.

The baffle model of Xbaffle is not as sophisticated as that of EDGE. But Xbaffle accounts for the reflections off the floor and off the back wall and nearest side wall. As with any speaker, the room related effects by far outweight the baffle related ones and tend to mask them in your diagrams.

[croat47]

Thanks!

So, for a best guess on what a particular driver would do in a given baffle/room position, Xbaffle is a little better approximation?

I am trying to determine a rough point for using a monopole bass driver as compensation for baffle step loss. Based on Xbaffle, it looks like rolling in around 250 for that width.

[Rudolf]

Originally posted by croat47
So, for a best guess on what a particular driver would do in a given baffle/room position, Xbaffle is a little better approximation?

I am trying to determine a rough point for using a monopole bass driver as compensation for baffle step loss. Based on Xbaffle, it looks like rolling in around 250 for that width.

IMHO the best strategy would be to first determine the optimal driver position on the baffle and the dipole loss frequency with EDGE (try off axis microphone positions too to see what happens there), and then the best baffle position in room with Xbaffle.

Look what happens if you move the upper right corner of the baffle 15 cm to the left. For a 3" driver the 40 cm baffle is too wide by far.

[john k...]

Worth looking at.

[croat47]

Thanks again!

Rudolf: I did as you suggested in The Edge. I did not re-center the driver. Significant reduction in the peak and null! Then I read the link that John K said was worth reading.

John K: thanks for the link. Certainly worth reading. To Rudolf's point: 40cm is too large based on the sim and on your discussion about the directionality. Does a driver become directional at a point where it's effective diameter is 1/4, 1/2, or 1 lamda?

Then, would one use a baffle that creates a situation in which the driver is operating as much as possible in the omni-directional range? Let's say I used a tweeter: would I target the crossover to a point below where the midrange becomes directional?

If I am in any way moving in the right train of thought: how would I determine an optimum offset within the baffle? This would change the actaul baffle width a bit as one shifted to either side.

Or: would one decide on the best circular baffle, and create an equivalent rectangular baffle per another discussion on John K's site?

Good stuff.

[pacificblue]

Originally posted by croat47
how would I determine an optimum offset within the baffle?

That is the place, where the distance from the driver to each edge is as different as possible and where the distances are no entire multiples of each other. The 'golden mean' (~1:1,618) is a good starting point for experiments.

[Rudolf]

Originally posted by croat47
Does a driver become directional at a point where it's effective diameter is 1/4, 1/2, or 1 lamda?

Between 1/2 and 1 lambda. This depends on the actual radiating area of the cone, which may be smaller at higher frequencies.

Then, would one use a baffle that creates a situation in which the driver is operating as much as possible in the omni-directional range?

Directionality does not depend on the baffle, but on the cone size. Baffle size commands the smoothness of the on and off axis radiation.

Let's say I used a tweeter: would I target the crossover to a point below where the midrange becomes directional?

By and large: Yes.

[croat47]

Originally posted by Rudolf

Between 1/2 and 1 lambda. This depends on the actual radiating area of the cone, which may be smaller at higher frequencies.


Directionality does not depend on the baffle, but on the cone size. Baffle size commands the smoothness of the on and off axis radiation.


By and large: Yes.

So, where is that high frequency shift from 1/2 to 1 lambda?

I think I understand that the directionality is based on the driver. Maybe I did not pose the question well. You have shown me that the smaller baffle works better with respect to the peak/null behavior. John K shows that increasing the baffle size above a certain point magnifies the effect of directionality in a given model of a driver of arbitrary size ( I think that is what it shows .)

That driver becomes directional at say 1/2 lamda. I would want to keep the baffle small enough to prevent the magnification of that. Or, I would want to cross to a tweeter before that directional behavior kicks in, yes? Referencing John's site, when he shows the increased size of a circular baffle: the null after the dipole peak gets deeper and moves to 2K from 1K. So, going too big would drive a null as well as then require (in this case) a too-low crossover for typical dome tweeters.

In my case, the driver by calc and by published response becomes directional around 2200Hz. So, crossover to the right tweeter could help on any baffle.

But, what would be an "optimum" baffle size? You pointed me at 25cm. Was that arbitrary, or is there some rough math that pointed you there?

How far off track am I?

[Edited to switch "1K to 2K" to "2K to 1K"]

[Rudolf]

All baffle effects are generated by a driver radiating at least 180° (half space). When the driver becomes increasingly directional (which is not a singular frequency but a range), those baffle effects are generated less and less.

So for baffle effects we concentrate on frequencies below the range where the driver is directional. It has been shown that for a most even horizontal polar response a baffle should be no wider than 2.2 times the effective cone diameter. 3 times does not hurt much, 4 or 5 times does.

What happens, if the baffle is too wide? For some frequencies the off axis response will be louder than the response on axis. This may look like reduced directionality to you, but it is a kind of frequency dependent off-axis-beaming. No good thing to have.

For practical reasons a wide baffle can not be avoided all the time. In this case it helps to move the driver sideways on the baffle. This will straighten the horizontal response in a certain direction, but worsen it in other directions. At least you can optimise it for your listening angle.

What would be a good strategy? My 7" midrange dipole driver with an effective radius of 13 cm is mounted on a 30 cm wide baffle, offset 5 cm from the middle. To one side radiation stays within 2 dB from 0° to 30°. Thats the side where I listen. The other side is worse.

The midrange driver becomes directional above 1 kHz. So at 2 kHz I cross to a 1" dome tweeter on the same baffle. A 2.5 cm driver on a 30 cm wide baffle?

Yes, because the tweeter has a small waveguide making it directional above 3 kHz. So it "does not know" much of the baffle width.

Hope this helps.

Rudolf