[JohnR]

I was toying around with some baffle ideas, and wondered how well an "L" shaped baffle would work. I haven't seen it used so I did a couple of drawings to explain what I'm thinking of.

In the below, you have two point sources of opposite polarity at the corner of the "L" shape. This would mean that you have nulls in the direction of both the rear and side walls. I don't know whether there would some acoustic loading inside the "L" corner that would invalidate this simple analysis, but I am guessing that the nulls to the rear and side would reduce excitation of room modes in both of those directions. Moreover, the front lobe is angled in towards the listener so increasing the direct/reverberant ratio.



Here's my thought on how to put this into practice. The L-baffles are used for the subs (up to, say, 200Hz). The main panels are located further to the center of the room, but angled towards the listener. This makes the main panel and the subs roughly the same distance from the listener; each also has a lobe pointing at the listener. This arrangement puts the sub in the null of the main panel, so reflections from the sub panel would be minimized.



Any thoughts on this idea? Has anyone tried anything similar?

JohnR

[Viridian]

I'm using L baffles with very good results. I'll post pics later, going to the coast for the day. I didn't like the sound that comes from the JE Labs baffle with both rear braces.

[scorpion]

With regard to OBs one should not be narrow minded. I think L-baffles would be OK. Probably main baffle distance to wall should be more or less the same as the bass-baffles. But experimentation is encouraged.
Don't expect very solid advice, just go on do it ! It costs so little to test.

/Erling

[Viridian]

I'm deleting my posts because I realized that my application has nothing to do with what JohnR is up to. My apologies.

[JohnR]

Viridian, no worries - would be interested to hear what you're working on anyway.

scorpion, you are saying move the main baffles forwards towards the listener? Could be, could be. I will try the L baffle for the subs - I was wondering if anyone had tried it or had any thoughts on the reasoning behind it.

[BPT]

Try pointing the corner of the woofer "L" directly at the listener, so the two woofers are equidistant to the listening position. That way you will have no time smear between the two woofers to the listener.
Chris H.

[JohnR]

Hi, that's the intent, although it may not be so clear from the diagram. However, it's a dipole (of sorts), so there is inherently going to be a time delay / phase shift between the front and back waves.

[JohnR]

This is the overall schematic of my planned project (one channel). I won't go into a lot of detail just now - but feel free to ask if you are curious about anything. Overall my goals in sound reproduction are, I think, quite similar to those of JoshK. The approach I have ended up with is different, though.

[D OB G]

Hi JohnR,

I've tried an L-baffle as you have designed, at the angle to the wall you intend, using three full-rangers (coral clones) per leg of the L, right at the corner.  i.e. all drivers equidistant from the listening position.  Worked very nicely.  Three of the drivers in parallel operating below an inductor as per "an OB design".

I also tried an L with one leg shorter, no drivers, placed against the wall so the the long leg was at an angle, perpendicular to the axis to the listening position.  Drivers right at the corner of the L.  Obviously open-baffle, but only dipole on one edge.  Effectively a much larger baffle for the size! with greater path length differences to the edge (infinite on two sides).  The whole panel was sand filled, 75mm thick.  Great bass.

Anyhow, your design looks good to me, with the nulls pointing at the walls a possible advantage? (diffraction more complicated?).

David

[Viridian]

I did essentially what David described but only with one leg and the driver biased to the no-leg side and with a 12" HE coaxial driver. The L keeps you from getting cavity resonances but gives good bass response.

[tubamark]

. . . This would mean that you have nulls in the direction of both the rear and side walls. I don't know whether there would some acoustic loading inside the "L" corner that would invalidate this simple analysis, but I am guessing that the nulls to the rear and side would reduce excitation of room modes in both of those directions. Moreover, the front lobe is angled in towards the listener so increasing the direct/reverberant ratio.


JohnR

The Null(s) would NOT be oriented as you indicate.  There can only be one null plane.  It would form approximately as if you were to draw an imaginary third side (dashed line, if you like) to complete a triangle.

Basically, it functions as a very open U-frame configuration; The null forms somewhere around the rear plane of the folded baffle
It may be a bit different in practice, depending on how tall the panels are to the open "top".  You may even consider adding a triangular top if you make the sides particularly short.

Placing the woofers equidistant from the listening position really doesn't matter, unless you are running them very high-- above 200 Hz.  Bass wavelengths are far too long to cause comb-filter/phase problems when placed close together as you show.

-- Mark

[Rudolf]

The Null(s) would NOT be oriented as you indicate.

 
That's right.

There can only be one null plane.  It would form approximately as if you were to draw an imaginary third side (dashed line, if you like) to complete a triangle.

That's only partially right. ''Only one null plane'' is true only for wavelength which are long compared to the L-baffle geometry. For higher frequencies there will be a change from omnipolar to (sort of) cardioid and dipole (because of beaming) again. In case of cardioid there surely are different null planes.

[JohnR]

Hi guys - thanks! You got me thinking. I see now that the nulls are indeed not as I first thought I need to do some more study but I suspect that perhaps there are not distinct nulls with this geometry. On the null plane question, Linkwitz does show that moving the dipole source away from the center of an H does create nulls at angles - http://www.linkwitzlab.com/frontiers.htm#L. (Item 2 in the fifth figure.)

[tubamark]

. . .  On the null plane question, Linkwitz does show that moving the dipole source away from the center of an H does create nulls at angles - http://www.linkwitzlab.com/frontiers.htm#L. (Item 2 in the fifth figure.)

'Glad you are spending time with Linkwitz - He's THE man.  Bingo on the asymmetrical H-frame. However, even shifted to U-frame that if the aren't angled by very much at all.  As a practical matter (we're never in free-space; even outdoors the subs are still in half-space), it is roughly a plane.  If you put your SW at ~45 degrees, you should get better room interaction (a couple studies have been published on this) than if parallel/perpendicular to walls.

Yep, Rudolph is right about null shifts higher up . . . Since JohnR is in a subwoofer application, I didn't want to muddy the waters.  Definitely a concern if one expects/requires a full-spectrum dipole pattern.   

slightly OT: To experience what Rudolph is describing, careful listening (pink/white noise, outdoors, is ideal) while walking around a typical OB panel reveals tone shifts, and never a true full-spectrum null.  I have experienced a deep null from a small Magnepan (the little fullrange MMG-W) - the difference is unforgettable, once experienced.  I believe Linkwitz's Orion accomplishes this as well, except in tweeter range where baffle is too wide to generate HF null. This can be "solved" by using highly directional tweeters (panel tweeter or Geddes OS waveguide).  So, you're on the right track, JohnR!   'Sorry for the aside . . .

[Telstar]

This is the overall schematic of my planned project (one channel). I won't go into a lot of detail just now - but feel free to ask if you are curious about anything. Overall my goals in sound reproduction are, I think, quite similar to those of JoshK. The approach I have ended up with is different, though.

Very interesting. Looking for the build