Foam in the range of 30-40 ppi works best. I tried a wide range. more open doesn't do much and more closed looses too much sound. I would expect the gutter foam to be way to open.
The 4430 was a good step along the path from the early Altec's to what we are talking about here. It's all been a progression along the path of what designers have believe for almost 100 years was the best approach to the problem. The 4430 waveguide was intended to be axisymmetric which was new as most to that point were not. I like the axisymmetric approach. Keele did the horn the way he did because he did not know how to do one that was round (his studies of ideal curvature were in two dimensions so he just duplicated those curves in the X and Y with appropriate displacement to allow for the diffraction slot.) Had he known then what we know now I am sure they would have been OSWG.
I suspect that the M2 is an attempt to update the 4430 approach, just as we are talking about here. But could you imagine JBL doing an OSWG in it!, and using foam!! Marketing would never have allowed that. They had to find some other, newer way to do the same thing that would not look like they simply copied the design. In the real world its not about making the best its about looking like you make the best. It can't look that way if it looks like someone else's.
The 4430 was a good step along the path from the early Altec's to what we are talking about here. It's all been a progression along the path of what designers have believe for almost 100 years was the best approach to the problem. The 4430 waveguide was intended to be axisymmetric which was new as most to that point were not. I like the axisymmetric approach. Keele did the horn the way he did because he did not know how to do one that was round (his studies of ideal curvature were in two dimensions so he just duplicated those curves in the X and Y with appropriate displacement to allow for the diffraction slot.) Had he known then what we know now I am sure they would have been OSWG.
I suspect that the M2 is an attempt to update the 4430 approach, just as we are talking about here. But could you imagine JBL doing an OSWG in it!, and using foam!! Marketing would never have allowed that. They had to find some other, newer way to do the same thing that would not look like they simply copied the design. In the real world its not about making the best its about looking like you make the best. It can't look that way if it looks like someone else's.
Last edited:
Didn't realize we're in the same area. If there are any audio get togethers coming up, I can throw the Summas in my car and bring 'em out. I did this at the Mercer Island audio meet up, back when I lived in Washington. (I'm in San Diego now.)
Excuse me, but with all this chatting about geometry wg etc. I think you are missing main point. M2 are big format compression driver speaker without supertweeter. All design aspects are related to this. You cannot have 1,5 inch throat and good dispersion with slotless horn. D2 is superb driver its throatless design. The dispersion slot on wg is made to spread the resonance which is caused by.
You just cannot compare this to regular 1inch throat designs.
You just cannot compare this to regular 1inch throat designs.
From my recollection the throat area was a bit sensitive. Our first samples were a yucky brown material (reaction injection molding?) and didn't hold their dimensions well in that area. The blend between the back throat (a casting) and the front horn was a bit variable and always seemed to move the low frequency peaks and dips around. We did play with it until we got the best overall frequency response (polars weren't much effected) and the production horns were better with proper tooling.
Still, the bottom end of the horn has some roughness to its response, as mentioned by Earl. The Keele formula works primarily on contours that will give the polars and if you go for the ultimate in polars then the slot area shrinks (highest frequency intended to hit a horizontal beamwidth is the critical parameter). The only part that is exponential is the linking section from the compression driver exit to the slot.
Maintaining your polars to a high frequency would typically give a low cutoff flare rate with inadequate mouth area to support it, i.e. lumpy response.
Secondly, the acoustical roughness is matched by the impedance curve bumps so a typical passive network will add at least a little to the peak to trough spread. I mentioned at Lansing Heritage that active drive to it or even a partially negative output impedance would be a neat way to reduce the roughness.
My sense is that the JBL designers today are much less obsessed with polar curves and achieve a reasonable overall compromise in their horn designs. Less good in the polars but better in frequency response and with a flaster flare to hold the distortion down.
Probably the smartest approach.
David S.
Well, I'd be happy to host you, if you don't mind a relatively small spot- except that you have a serial killer for a moniker!!
Just kidding- I fear no fictitious character 😀
No hifi meets that I know of, in the near future. Would love to hear Summas, since I've utilized summa their concepts. I'm up in Tustin, if it's ever convenient, my schedule tends to be pretty flexible, for hosting, but travel is harder as I have young kids (6 and 10) and a wife that works retail (oddball hours).
BuzzforB-
My woofers are BD BD15, my tweeters are modified diaphragm Selenium D2500Ti (I like their embossed phragm, ring neo, and the small assembly size), my rearfiring woofers are JBL 2206h just as in the Disneymod speakers (I need to find a silly title for them, taking suggestions), the rear tweeter is
Building A Budget (Super?) Tweeter Reaching the stratosphere! Article By Jeff Poth
Ah! That reminds me of an article about 3 woofers. Nice. If I dont like the altec setup, I will try either the 2235 or AE td15 next. The 2235 can go really low in a raather small box, but loses in the efficiency dept. The td15's seem to lie somehwere in between the two, considering all models. Lately I have considered dual TD12's, using the M and H. It would defintely allow for more horn/waveguide choices. Interesting about the Selenium
John's not a fictitious character, and you have to wonder about someone who idolizes one, fictitious or not. 🙂 (Talking Heads?)
I have seen that movie a couple of times and I am not convinced that he actually was a serial killer, he may have just been fantasizing it. At any rate I really loved that movie.
Hi Dave
I am not sure that I agree that "maintaining polars to a high frequency" does require a slow flare rate. In my waveguides the flare rate is very rapid and the polar's hold to well above 10 kHz. The classic beliefs that come from Webster's theory (exponential etc.) do not hold when one uses waveguides. It is the failing of the exponential profile that causes the high frequency collapse - it is not true in general.
I don't see why you can't have smooth response AND good polars. The DSP that they seem to be using is not much different than what I use in terms of smoothness. Seems to me that both are obtainable a the same time.
As to "holding the distortion down" Alex Voisvillo is very much of my mindset on this - it's not a factor - so I doubt that is the reason for a fast flare rate.
I don't see why not. If 1 inch is the best option then using something else is wrong. It is certainly a valid comparison. In a two way like the M2 efficiency of the HF is not an issue because it has to be padded down to match the woofer, so a larger diaphragm is not required.
Earlier we were talking about giving credit where credit is due.
I remember 10 or 12 years ago, when I first became interested in the catenary flare profile, doing a lot of research and finding catenoid horns were popular in another field. Seems like ultrasonics, maybe. I also remember reading about an early adopter of catenoid horns, someone that promoted the flare shape long before any of us here were using them.
Does any of you know that researcher's name or the reasons he preferred catenoid horns? Might be an interesting factoid, one that I have long since forgotten.
I remember 10 or 12 years ago, when I first became interested in the catenary flare profile, doing a lot of research and finding catenoid horns were popular in another field. Seems like ultrasonics, maybe. I also remember reading about an early adopter of catenoid horns, someone that promoted the flare shape long before any of us here were using them.
Does any of you know that researcher's name or the reasons he preferred catenoid horns? Might be an interesting factoid, one that I have long since forgotten.
It's definitely a good flick- and the line between his fantasy and reality is blurred. You might also like "The Machinist".
Well Earl, although i tend to agree with your sensibilities, i was able to audition a pair of M2s a few weeks ago at an industry event and i gotta say....best speaker i've ever heard hands down.
I am referring to the Keele style of horn. The 2344 held its 100 degree beamwidth to 16 kHz. This required a diffraction slot of about 3/4" wide. The slot width determines the slot area and with throat area, slot area and separation distance defined then you have predetermined the expansion rate. I'm sure your wave-guide is controlled by the same physics (of diffraction, not of area growth). As it appears to be deeper with a wider throat diameter I don't see it having the same high frequency dispersion as the 2344.
As you know, horn design is all about compromises. You have chosen to give up good on axis response to have smooth response off axis. Others have made different choices. I haven't seen your polar curves presented in a form comparable to the JBL horns but I will say that I have never, before or since, seen any loudspeaker unit with a consistancy of polar curves, Octave after Octave, to match them. That was my point, the current JBL horns aren't nearly as good in that regard but the designers have probably steered towards a more balanced set of compromises.
David
They got lucky
Truthfully, I bet the sum as would sound pretty dn awesome with the JBL drivers. This is something we have not spoken of. It's hard to argue with the quality of the components.
Sidenote:
Anybody played with JBL WTi MkII car woofer.. It clearly shares at least a similar shell with its more notable brothers.
The "physics" of Webster and those of my designs are NOT the same. That is really the point. While you seem to have a good understanding of most things in loudspeaker design your understanding of horns is out of date. My waveguides dramatically outperform those of the 4430.
I keep going back to the 18Sound XT1086, because I think it illustrates these two philosophies well.
The 18Sound XT1086 has a diffraction aperture at the throat, and the horn is not axisymmetric.
The waveguide in my reference speakers *is* axisymmetric, and has no diffraction slot. (Gedlee Summas.)
If you look at the polars of the two waveguides you notice two things:
1) the polar response of the Summa is more consistent horizontally than the polar response of the XT1086
2) The frequency response *in front* of the XT 1086 fits into a tighter window than the frequency response of the Summa
pics or it didn't happen:
^^^ Geddes waveguide from the Abbey, basically the smaller version of my reference
^^^ 18 Sound XT1086
I owned the Summas for over five years before the light bulb went off over my head, and I realized that "on axis" for this speaker means "off axis" for most speakers. The listening axis isn't in front of the speaker; it's off to the sides. And the crossover is optimized for that.
That diffraction aperture in the XT1086 flattens out the response in front of the speaker, but the beamwidth isn't as consistent. The measurements of the XT1086 are very good, and I hope this post doesn't come off as an 'attack' on the design. As I see it, these are just two different approaches. The beamwidth of the XT1086 varies more than the OS waveguide. The consequence of this is that the lower frequencies will beging to dominate as you move off axis. In the OS waveguide, the response falls off more consistently as you go off-axis*, so the timbre of sounds will be more consistent. IE, it will basically just get quieter. IMHO, this is one of the reasons that these speakers are imaging champs.
It's a subtle difference I think; one might look at the measurements of the OS waveguide, and seeing that big dip in front of the waveguide just dismiss it.
Not to get too far off topic, but I think the choice of compromises has a lot to do with it being 2013. In 1993 it wasn't easy to fix the frequency response problems of horns. While equalizers were readily available, even the most affordable measurement gear cost upwards of a thousand dollars. Nowadays I can measure my loudspeakers using Arta (free if you don't save anything), and Dayton mic ($15). I can correct the response using MiniDSP ($99.) It's never been cheaper or easier to EQ a speaker I'd say.
But you can't EQ directivity 🙁
* it's really difficult to discuss "on axis" and "off axis" with these waveguides and horns, because the definition of the term "on axis" is truly open to debate.
Last edited:
Depends on how much you're willing to EQ, and where your listening axis is.
I wound up going down this rabbit hole when I was trying to figure out how to get my mids to play higher on a Unity horn, and discovered that narrowing the wall angles raises the output level of the horn on the low end of the compression drivers output.
As Badman noted earlier in this thread, that's not just due to narrower wall angle; it's also due to the fact that narrower wall angle requires a deeper horn, and a deeper horn will load lower. So if the mouth size is identical but the beamwidth of the waveguide is narrower, the 'gain' at low frequencies will be higher. (Because the narrow angle horn is deeper, hence the quarter wave frequency is lower.)
But narrowing the wall angle introduces a ton of other challenges, which is one of the reasons that constant directivity horns generally had two stages, as illustrated in the Peavey white paper linked earlier in the thread. Also, the Peavey white paper makes a great case for OS waveguides, even though it doesn't mention them by name!
John
I'd just like to point out that an ESP12 is "like" an Abbey, but it is not the same, and its not a Summa. The axial "hole" is much greater on the ESP12 than it is in current models. You are correct that "on-axis" is off the "listening axis" by about 22 degrees.
"You can't EQ directivity" is a really important point.
I'd just like to point out that an ESP12 is "like" an Abbey, but it is not the same, and its not a Summa. The axial "hole" is much greater on the ESP12 than it is in current models. You are correct that "on-axis" is off the "listening axis" by about 22 degrees.
"You can't EQ directivity" is a really important point.
Last edited:
- Status
- Not open for further replies.