[tubamark]

Seems like U's, H's and flat baffles are all in use by those on this forum.
Anyone out there (still) giving preference to W-frame alignments?
I can build a double-driver W, essentially within the same dimensions as a single driver H, and have double the displacement capability (for a given excursion). . .

I'm guessing that when SL evolved from the Phoenix to the Orion, he went with the double-height H simply because the full range of drivers could all fit into a simple tower, but he doesn't say for sure if there were any other reasons.

Theoretically, I could create a little cabinet 12" square that has the output of a 15" driver in an H frame (given identical excursion).  I like 'em compact.
There are always tradeoffs . . . so for those of you who have firsthand experience:  What are they, if any?

[JohninCR]

I have several W alignments.  One pair has 4 12's per side in a double ripole alignment, so they are only 12" wide.  The other is the straight Linkwitz W.  The benefits of a W are the mechanical force cancellation, and the ability to use push/pull.  There isn't another benefit unless you consider pure dipole a benefit, because given the same exterior dimensions you get the same performance with a single driver in a properly damped U-baffle.  Plus the U can be used higher because the resonances are damped.  The U can generally be made more compact, by shrinking the height and width since the outside of the driver frame is on the exterior.

The instance where I would recommend W manifolds is using them as bases for the main drivers.

[wikin]

Why not build a hybrid of w + u ? I'm in the process of building a w-frame using 2x15" Visaton BGS40 with the added longer tail on the rear, which would then be dampened. So I get mechanical force cancellations, reduced distortions due to the push-pull operation and increased efficiency. aa

[JohninCR]

Why not build a hybrid of w + u ? I'm in the process of building a w-frame using 2x15" Visaton BGS40 with the added longer tail on the rear, which would then be dampened. So I get mechanical force cancellations, reduced distortions due to the push-pull operation and increased efficiency. aa

Wikin,

Hybrids are great.  In fact, my 4 12's per side cabs are hybrids with an extra 8" to the rear.  You reminded my of another potential benefit of a W.  It lowers the Fs of the drivers with more air mass on the cone, and can flatten the bass response when taken to the extreme such as in a ripole alignment with very narrow cavities, reducing or eliminating the need for EQ.  This does come at a cost though, and that is reduced efficiency, not increased efficiency.

I find that OB low bass is easier to localize than monopole.  I believe this results from open alignments having highly directional bass.  Given a choice, I now always opt for dual cabs because of this.

Let me explain my first post a little further.  With OB bass driver sensitivity becomes a relatively unimportant factor.  What becomes important is limits and potential, because your limiting factor is driver excursion.  The reason a single driver U that is properly damped enjoys equal performance to a same sized dipole with double the drivers is that the U doubles "D", the rear wave propagation delay.  Doubling "D" gives you a full octave or +6db, which is the same as double the cone area (+3db) plus double the power to drive it to equal excursion (+3db).

[scorpion]

Tubamark,

Nothing at all wrong with W-baffles if crossed steep and at 100 Hz or under. Over they tend to sound 'boxy' and resonant in my experience. Look at Linkwitz' Phoenix, you must take care of the first resonance if you don't cross extremly steep.

/Erling

[Davey]

John,

The "U"-frame trades the symmetrical response for the other aspects you noted.  This is simply NOT "equal performance" as you've mentioned twice now.  It may even be superior, but that's another subject.  The "U"-baffle benefits have been espoused by John K. and well documented, but there is no free lunch.  Yes, there are some "advantages" (depending upon how you look at it) of the U-frame compared to symmetrical dipoles, but there isn't any equating.  It's apples and oranges because of the designed/expected polar response difference.

I believe the force cancellation aspect of the "W"-frame construction is a considerable advantage.  At least in my application.  However, for other users with different environments....floor construction, non-integrated mid-baffle, etc, etc,...it may be a needless complication.

Davey.

[JohninCR]

Davey,

Thanks for correcting me. By performance I meant spl capacity.  Regarding polar response, while that may be true outdoors, in-room the difference is less defined.  First, it's nearly impossible to build a perfect U, so you end up with a hybrid.  Even though you end up with a hybrid response with U's, the rear propagation delay is a finite and known.  Second, dipoles don't really achieve a dipole polar response in the low bass region in your room anyway, unless they are placed halfway into the room, but that would be atypical.  It's really not apples and oranges, more like peaches and apricots, but placement and room construction often have bigger audible impact. 

Even though JohnK has espoused U-baffles for some time, he is very technically oriented and most of us get lost in much of what he says, which probably a big part of why they aren't more popular, not to mention that up until very recently many other technical types were still disputing John's +6db claim.  Now that SL agrees, I think that is put to rest.  As far as dipole purism is concerned, pure dipole speakers don't exist anyway.

[Davey]

John,

Yep, I knew you knew what you meant, but it wasn't clear in your description.  The listening room is indeed a large variable and whether or not the trade-offs involved with one or the other works in a particular room is hard to determine.  The SPL capability is ultimately determined by the driver(s) itself (Sd*Xmax)...all other things being equal.  However, in this case they're not so it's more complicated. 

John K. feels the trade-offs with the U-frame construction are preferred relative to a symmetrical dipole.  He may be correct, but he makes it seem like it's a win-win situation on his webpage.

I believe the ultimate DIY dipole design would be a totally symmetrical configuration for the whole frequency range.  Something like maybe twin MTM baffles mounted back-to-back with edge diffraction control mounted above a symmetrical dipole woofer system.  I currently have built a prototype system just like that and am testing it.  It looks promising.

Davey. 

[JohninCR]

Davey,

Strictly in terms of size and SPL, the U is king.  For low bass they seem to sound a little better than dipoles in a typical placement with a boundary not too far in back, but dipoles sound much better than U's with placement near the side walls.  I think it's something related to reflections that makes drums sound more lifelike.  Listening position in relation to the wall in back can also create a lot of variability in bass response with U's, especially with concrete walls like I have.

WRT your current project, how do you address the midrange prominence well off axis?  I'm experimenting with dipole midrange waveguides in an effort to achieve constant directivity and balanced off axis response.  I'm taking Dr. Geddes constant directivity approach with his Summa, but going dipole to include directivity in the bass region.  Some of my test waveguides are showing very good promise as long as I stay clear of horn loading.

[johnk...]

...not to mention that up until very recently many other technical types were still disputing John's +6db claim.  Now that SL agrees, I think that is put to rest.  As far as dipole purism is concerned, pure dipole speakers don't exist anyway.


...but dipoles sound much better than U's with placement near the side walls. 

(JPK) That first comment is interesting. When I first introduced my U-frame system SL and I had a rather heated, on going discussion over at madisound on that.  He wasn't buying 6dB and his response was an analysis of the U-frame to support his point and further discredit my position in his frontiers section. How he tweaked his model to show his position I don't know. But  you know how models are, garbage in, garbage out.  I had performed my own analysis and made measurements confirming my position of 6dB with correct damping, consistent with Backman's model.

With regard to dipole sounding better against side walls than U's, that may well be, though I think such a generalization may be premature. The thing with dipole is that when placed against a side wall then you can be pretty sure that only axial (front to back) modes are excited.  That is not the case with U's. But it's not clear that that will always result in better sound. However, when a dipole woofer is placed out in a room, as it would be with an integrated system, and the system axis is angled towards the listener so that the dipole axis is not aligned with the room front to back axis, the dipole will excite other modes as well. There are some interesting AES papers on the subject by Kates, Dipole Loudspeaker Response in Listening Rooms (2002), and by Salmi, Dipole Source Placement in a Room (1992).

Constant directivity is becoming a new catch phrase. But what is the correct directivity pattern? What is best above the Schroeder frequency is not necessarily the best below it. Theory and modeling are starting to come together and indicate that cardioids are generally better below the Schroeder frequency (perhaps not againsit a side wall  ). Above the Schroeder frequency room treatment has a big effect on what sounds best. And then there is the power response issue which suggests that the the directivity index should change as well, for example as a speaker system encounters re-enforcement form a ground plane.

Lots of things to keep of busy. But like many areas of science, current trends change and if you look deep enough into the past you may find a lot of this is just cycling around. You can rest assured that the wheel was invented many times.

[JohninCR]

Hi John,

Yeah I heard about the debate, but that was before my time.  To me the +6db seemed simple because D is doubled, so SL own Fequal formula proves it nice and simple.  The part I never really caught on to until recently was the negative GD/lumped mass behavior thing, which makes damping a requirement for proper operation.  I wonder how many understand that they can damp the rear section of a compact H frame and increase bass with a free 50% increase in D.

Could damping only the rear pathway(s) of a dipole W prove beneficial?

[Davey]

John,

There won't be any midrange "prominence" if the baffle is designed correctly.  Prominence would be caused by too wide a polar response which adds to the reverberant field and makes it too "bright" even though the on-axis response looks good.  This is commonly caused by too wide a baffle in "conventional" dipole designs.

Constructing a symmetrical dipole is an interesting challenge because the question of how best to handle the edge diffraction, possible resonances between baffles, etc.  The two (if there are two) baffles could also be non-flat, curved, or straight with the edges brought together so they touch.  Any edge diffraction would cancel effectively in that case. 

Initially, I have constructed something straightforward with two 10" wide baffles mounted back-to-back and am experimenting with closing the sides (but not the top/bottom) with a half-section of PVC pipe.

Davey.

[JohninCR]

John,
There won't be any midrange "prominence" if the baffle is designed correctly.  Prominence would be caused by too wide a polar response which adds to the reverberant field and makes it too "bright" even though the on-axis response looks good.  This is commonly caused by too wide a baffle in "conventional" dipole designs.
Constructing a symmetrical dipole is an interesting challenge because the question of how best to handle the edge diffraction, possible resonances between baffles, etc.  The two (if there are two) baffles could also be non-flat, curved, or straight with the edges brought together so they touch.  Any edge diffraction would cancel effectively in that case. 
Initially, I have constructed something straightforward with two 10" wide baffles mounted back-to-back and am experimenting with closing the sides (but not the top/bottom) with a half-section of PVC pipe.
Davey.

Davey,

Are you going to stuff the heck out of that interior.  I have some PVC pipe and have often thought about using it to try both bipole and double driver dipole at the same time.  Easy for addressing edge diffraction, and easy to build.  I wouldn't want to put a divider to separate the drivers due to reflections, but I am concerned about the rear radiation interfering with the other driver's operation in the higher frequencies.  It seems like I remember reading that less than 1/4 or 1/8 wavelength separation between drivers is needed for isobarik to function properly.

Regarding polar response, can you really get a smooth transition from dipole cancellation at the sides to controlled dispersion due to driver directivity without using waveguides?  Isn't that why people are waveguiding the tweeters, but just not doing so on the midrange because they would be too large?   Maybe I just haven't tried a small enough flat baffle, and these big waveguides I'm doing are a waste of time for addressing polar response.  The increased dynamics (despite avoiding horn loading) and more balanced response +/- 20 degrees of axis are very welcome, but I was after balanced response 360deg around the speaker, but directional unlike omni.  Am I just going at it the hard way?