Wide baffle projects (ala "Acapella WB, Poor Man's Strad")...

[fjhuerta]

I don't recall having seen one here. We all build (well, except for Dan and Jon!) slim columns / bookshelfs. High WAF, I know.

Lately, I've been browsing through Troels Gravesen's web site, and his wide baffle projects caught my eye. They look interesting, but I'm wondering if a design with basically zero BSC could be built.

This has a couple of benefits - no BSC and no baffle diffraction (that, and my wife liked the look of his PMS). He mentions some people believe "pinpoint accuracy" is sacrificed, but that then again, the whole "accuracy" thing only applies to synthesized music. Fair enough.

I'm considering building a set of wide baffles as a proof of concept, using the W3-1335SB as a midrange, a pair of RS150's as woofers and a SEAS 27TAFN as a tweeter. I'd love to build a speaker with the minimum baffle loss possible, in order to use smalller inductors and have a high sensitivity.

Do you think I could get away with zero BSC with a 55 cms. wide baffle? So far, I can see that using Edge, I can get to 160 Hz without loss. At 100 Hz, I'm down -2.1 dB, and at 50 Hz, -4.5 dB. Can I consider room gain will cover those losses?

[bob barkto]

In one of his plots he shows some room gain starting from ~100hz with his crossover in the PMS version.
He predicted enough for a reasonable f3 of around 40hz I think.
So you may need some compensation depending on what you think is adequate.
Typically you'll see most room gain well below 80hz.

I've been thinking of building a variation of his design.
Have the drivers but no cabs yet. I'm pretty sure a test cab is in order for this as I haven't heard this design before. Weather permitting I'm a few months off.

[Saurav]

http://www.htguide.com/forum/showthread.php4?t=30421

[Bear]

I will say that this particular effort is one of the most beautiful DIY cabinets I've seen. If I had the space for two 20" wide speakers, this would be a fun summer project.

Bill

[HareBrained]

Very interesting idea. I think I'd step up to the RS180 simply because it works in the same size enclosure and it'll give you a little deeper FR. It seems the norm to gently curve the baffle to eliminate the "edge" of the speaker. The amount of room gain, and it's starting point, is going to be dependent on the size of the room, but generally speaking, it won't cover those loses. But a subwoofer would.

I'm considering a similar project. I recently built a half wall in the basement to hide some plumbing, and covered it in bead-board wainscoting. Well, WAF was incredibly high. That got me thinking, if I were to build a half wall in the TV room, there'd be a place to stand the TV and I could put speakers into the new half wall.

[Curt C]

One possible caveat with wide baffles is a peak and dip in response not present in a typical narrow baffle.

While the wide baffle moves the baffle step down about an octave, it also develops a diffraction artifact squarely in the midrange: A pretty substantial peak and dip at roughly 600 Hz and 900 Hz respectively with about 3 dB of swing. This will be difficult to correct with a passive crossover, as norcad noted in his design thread.

I didn’t put the curve up on the plot below, but modeling suggests that baffles up to 18” wide will exhibit only the bump, and not the dip, and under the conditions noted below, and provides most of the benefits of a 24” wide baffle. Obviously this will vary somewhat with the design edge treatments, driver dispersion characteristics, etc..

C

[Winter]

Curving the baffle sides back is important. If one is are using diffraction modeling software, try modeling 6" 45 degree wings on the baffle, i.e. a 15.5" flat front with a 6" 45 degree wing on each side (close to 24" wide). The diffraction curves will improve over just a 24" flat baffle.

[fjhuerta]

All this has made me even more curious about wide baffles!

I wish I could use the RS180s. Actually, if I like the wide baffles, I'd build a second version with the RS225 (RS225, W4-1337s, 27TAFN). The first version would use the RS150, W3 Titanium, 27 TAFN).

Curt, I spent some time modeling the baffle and indeed, I couldn't get rid of that dip. Off axis it's minimized (15°) and FR is quite flat. I wonder if power response would be good enough to ignore the dip.

I'm still wondering whether at least *some* BSC should be added, too.

I'll probably build a test baffle in the following weeks. Those speakers look intriguing.

[Winter]

Curt C,

The reason for the large diffraction dip at 950 Hz for the 24" wide baffle is that the driver is located equal distance to 3 edges, 12" to each side and 12" to the top/bottom edge. The effects are cumulative. Move the driver up or down and the peak & dip will decrease in magnitude. The same effect would occur for the 10" wide baffle at a higher frequency if you located the driver 5" from the top edge.

I did not realize this either till I stumbled on the effect while modeling 4 drivers in a vertical array using Jeff Bagby diffraction modeling software. Also, on another note, the software limits the maximium chamfer or rounded edge one could model to 4.3", which definitely has an effect on the diffraction.


Two interations of Jeff Bagby's programs:

http://www.pvconsultants.com/audio/frdgroup.htm

http://audio.claub.net/software/jbagby.html#FRM

[fjhuerta]

I find that I can't get rid of that dip, no matter what... on axis.

Off axis, response is very, very flat. Flatter than I recall on most of my designs.

It's possible that power response will be good enough that this won't matter. Or design axis could be at 15°.

[Winter]

Javier,

With the following parameters using Bagby's diffraction modeling program:

Baffle height: 40"
Baffle width: 24"
Edges: 4.35" Round
Drivers Sizes: 6.5" x 4
Piston: 5"
Next Sample: Quadruple
Next Edge: Complex
Next Source: Directional
Next Shape: Quadrangle
Next Type: Separate
Axis Angle: 0 (On-Axis)

Driver Center Height From Bottom on Baffle, dB peak/dip, modeled on-axis

6.5 Inch Driver, 8" center from bottom 1.75dB peak / -0.4 dB dip
6.5 Inch Driver, 15" center from bottom 2 dB peak / -1.15 dB dip
6.5 Inch Driver, 22" center from bottom 2.5 dB peak / -0.35dB dip
6.5 Inch Driver, 29" center form bottom 1.75 dB peak / -1.1 dB dip

Be aware of the units of scale used on the sound level axis (dB, vertical axis), so a small dip is not reviewed as a large depression. Our graphs frequently use 5dB graduations on the vertical axis.

[norcad]

The diffraction "problem" I had with my wide baffle speakers, was almost eliminated when I put the midrange on top. Now the tweeter has this dip/peak combo, but its so low in freq that it doesnt matter when I cross at 3khz. Look at the latest measurements in my threath and you will see that the frequency response is within +/-1dB from 200-17000hz.

So there is no reason not to build wide speakers!
Well, the diffraction isnt anyway...........

[Curt C]

Curt C,

The reason for the large diffraction dip at 950 Hz for the 24" wide baffle is that the driver is located equal distance to 3 edges, 12" to each side and 12" to the top/bottom edge. The effects are cumulative. Move the driver up or down and the peak & dip will decrease in magnitude. The same effect would occur for the 10" wide baffle at a higher frequency if you located the driver 5" from the top edge.

I did not realize this either till I stumbled on the effect while modeling 4 drivers in a vertical array using Jeff Bagby diffraction modeling software. Also, on another note, the software limits the maximium chamfer or rounded edge one could model to 4.3", which definitely has an effect on the diffraction.

Oh, I knew, it... :B
I chose it intentionally to demonstrate the issue, as a mid would normally fall about 12" down from the upper edge. As norcad and you have indicated, the situation improves significantly as the driver is moved up closer to the top or bottom edge of the baffle. Thanks for making that point clear, as I was remiss in my post to not mention it.

C