I will think I will have trouble picking a favourite among similarly sized Kugelwellen, tractrix and JMLC.
I will think I will have trouble picking a favourite among similarly sized Kugelwellen, tractrix and JMLC.
Way more important is the combination of different sized horns and to reach an acceptable midbass solution. I forgot about midbass and one year in I am still struggling with different midbass horns.
Kugelwellen profile is very near of Le Cléac'h profile for the same cut-off.
Tractrix is different and cannot naturally roll back at the mouth as the Le Cléac'h and Kugelwellen horns.
But the main difference is in the way the horn provide its acoustical load to the driver. Le Cléac'h and Kugelwellen will provide a more resistive load than the tractrix (this one having an exponential horn loading behaviour). This means that you can use them on a larger interval of frequency.
Best regards from Deauville, France
Jean-Michel Le Cléac'h
I hate to burst a few of the assumptions that are being used here but classic compression drivers hardly have a flare rate as you are stating them. The original JBL, Altex and WE drivers were far from exponential in expansion. What you truly have is either two or three points in the compression driver that would match the rate of expansion of an exponential flare. The phase plug section and the throat section are conic in section and only at the end points would the rate of change match an exponential expansion rate. Perhaps with the newer no throat compression drivers this has finally been manufactured with a true flare rate but I haven't disassembled any of those drivers to analyze the true rate of change.
I assumed wrongly that tractrix was close to hyperbolic but it is in fact Kugelwellen and JMLC that is closer to hyperbolic, right?
Well, good then because I sculpted a kugelwellen horn for the JA6681B's exit angle and it sound vvery nice. If this is almost JMLC I am all for it. Again, I would probably like both KW and JMLC profiles and have difficulty picking a favourite.
I do not like exponential horns, in any size. Maybe for a hornloaded subwoofer, this I have not tried yet.
That's good and desired info. While the conic sections are not exact, you appear to be saying that they intend to approximate an exponencial flare?
I'm assuming this intent based on this exerpt from JBL history link above:
While the full LeCleac'h profile is not the same as an exponential, it appears that it can be loaded at one (i.e., T=1).
I don't know how the modern wide angle horns or waveguides acoustically load the driver, and/or whether the modern phase plugs of JBL match the wider flares for *no* discontinuity in flares.
Pooge,
Yes you are basically following what I said correctly. I have actually physically disassembled some compression drivers and what was done in the past was that the sections of the phase plug and the throat section of those drivers is made up of truly conic sections. What they had done was to take two or three points along an exponential expansion rate and set the end dimensions of the conic sections to match the exponential expansion rate. This is as close as they get to making the driver exponential in the least. So when I see all this talk of having to match the horn to the expansion rate of the compression driver I have to ask what people think they are actually accomplishing?
If you truly were to match the throat or expansion rate you would have to build a conic horn, that is the only thing that would match with a compression driver. Now that being the reality it is obvious that with the results attainable with a compression driver and horn that this matching is not in-fact as critical as people are saying here, at least as far as the air loading of the diaphragm is truly concerned.
I am only familiar with the JMLC waveguides or horns from this website. If it is indeed a hyperbolic expansion that would be something that I have used for the past 35 years as a normal expansion rate, but at the same time you must realize that a hyperbolic rate of expansion does include exponential as one of its solutions. So when I see a comment that someone likes hyperbolic but hates exponential that makes no sense at all.
Lynn,
I am somewhat surprised by the T value that you are saying you want to use for your horn design. To me it would appears that you will have a very long and narrow expansion in the initial section of the horn and that would seem to me that you are making a very directional horn in the upper frequencies. Perhaps you want to do this to bring up the upper frequency response on axis to counter some of the mass roll-off of the driver you prefer? Personally I am not as concerned with the loading of the diaphragm and the uncorrected axial output as it is so easy today to make those types of corrections. I would be more interested in the dispersion angles and the smoothness of the polar responses across the board.
Yes you are basically following what I said correctly. I have actually physically disassembled some compression drivers and what was done in the past was that the sections of the phase plug and the throat section of those drivers is made up of truly conic sections. What they had done was to take two or three points along an exponential expansion rate and set the end dimensions of the conic sections to match the exponential expansion rate. This is as close as they get to making the driver exponential in the least. So when I see all this talk of having to match the horn to the expansion rate of the compression driver I have to ask what people think they are actually accomplishing?
If you truly were to match the throat or expansion rate you would have to build a conic horn, that is the only thing that would match with a compression driver. Now that being the reality it is obvious that with the results attainable with a compression driver and horn that this matching is not in-fact as critical as people are saying here, at least as far as the air loading of the diaphragm is truly concerned.
I am only familiar with the JMLC waveguides or horns from this website. If it is indeed a hyperbolic expansion that would be something that I have used for the past 35 years as a normal expansion rate, but at the same time you must realize that a hyperbolic rate of expansion does include exponential as one of its solutions. So when I see a comment that someone likes hyperbolic but hates exponential that makes no sense at all.
Lynn,
I am somewhat surprised by the T value that you are saying you want to use for your horn design. To me it would appears that you will have a very long and narrow expansion in the initial section of the horn and that would seem to me that you are making a very directional horn in the upper frequencies. Perhaps you want to do this to bring up the upper frequency response on axis to counter some of the mass roll-off of the driver you prefer? Personally I am not as concerned with the loading of the diaphragm and the uncorrected axial output as it is so easy today to make those types of corrections. I would be more interested in the dispersion angles and the smoothness of the polar responses across the board.
Yes. Some of this is an exercise in mental masterbation, as all the expansions essentially converge at the throat, and the throat is not that long. However, if given a choice of drivers, would one want to use a driver having a throat with an Fc of 160Hz in a 340Hz to 600Hz horn if a driver with a 340Hz or 550Hz driver, respectively, were available, all else being *equal*? Different drivers get different results. What are the characteristics of a particular driver that achieves the most favorable results on a given horn?
Yes, it's a continuum of solutions. While T=0 gives you the highest loading in the bass with higher distortion, and be crossed closer to Fc, it would constrict the polar dispersion up higher as compared to a T=1 that would require a xover farther away from Fc.
I don't want to speak for Lynn, but my impression is that one of the reasons is as you suggested, and that he wanted to use a ribbon tweeter anyway. Further, he stated that CD is not a top priority for him. He also suggested that this horn was the biggest horn that was practically available to him. To extend the HP xover down as far as possible and/or minimize group delay, a lower T would help get an extended flat response and better resistive loading
Pooge,
As far as which compression driver to use on which horn you can come at that two ways. One way is to look at this as a purely mathematical problem and try to match the rate of flare in the driver with the horn. The other way is to mount the drive on a horn and actually measure the response, it will often diverge significantly from what you would expect. Trial and error in effect unless someone else has already analyzed that particular combination.
I am not one who would try to get a compression driver to operate at the lowest frequency possible, those trying to hit 150hz with a compression driver are of a completely different thought than where I am at or anyone I know doing anything professionally with compression drivers. My time in pro audio was during a time of change away from those very large format Altec and JBL drivers to more compact smaller drivers. At the same time we went to smaller exit drivers we added mid-range horns to get out of the compression drivers higher. I much prefer a mid-range cone driver on a horn over a compression driver operating in that lower range.
As far as which compression driver to use on which horn you can come at that two ways. One way is to look at this as a purely mathematical problem and try to match the rate of flare in the driver with the horn. The other way is to mount the drive on a horn and actually measure the response, it will often diverge significantly from what you would expect. Trial and error in effect unless someone else has already analyzed that particular combination.
I am not one who would try to get a compression driver to operate at the lowest frequency possible, those trying to hit 150hz with a compression driver are of a completely different thought than where I am at or anyone I know doing anything professionally with compression drivers. My time in pro audio was during a time of change away from those very large format Altec and JBL drivers to more compact smaller drivers. At the same time we went to smaller exit drivers we added mid-range horns to get out of the compression drivers higher. I much prefer a mid-range cone driver on a horn over a compression driver operating in that lower range.
Hi Jean-Michel
If I wanted to go with a JMLC J321 would this allow a compression driver that will give reasonable treble without a separate tweeter, and using a bass reflex enclosure. And if so what T value.
Mid-Bass
Man Oh Man, isn't that just the truth (?) !!
I've been working on the mid-bass range for about three years now.
It would be nice to see a thread along the lines of "ideas for mid-bass horns"
Man Oh Man, isn't that just the truth (?) !!
I've been working on the mid-bass range for about three years now.
It would be nice to see a thread along the lines of "ideas for mid-bass horns"
To the dismay of some readers, I am not interested in constant directivity. There. I said it. Really, I mean it.
Why do I have this attitude? Simple. I have my favorite speakers that I've known over the last forty years. Not one of them was constant directivity. I frankly don't think it's that important.
What I do think is important is flat, peak-free response over a 40 to 50-degree horizontal arc (covering the high-quality listening area), a rapid decay-to-zero that is free of resonance (in the same listening area), adequate dynamic range and headroom, and efficiency that is high enough for DHT-triode amplifiers.
OK, rant over. In more detail, what matters to me is the quality of the first-arrival sound in that 40 to 50-degree arc, and the overall power spectrum into a sphere. I don't care about radiation into a 90-angle angle at all. Why?
It's not hard to know where the first reflections are coming from. The first will be the floor bounce, in the 1~2.2 mSec interval, depending on driver height from the floor, and to a lesser extent, listening distance. Depending on ceiling height, the next will be the ceiling reflection. The rear and nearest-side walls will come next, followed by a double reflection from rear&floor and nearest-side&floor. There are many more reflections after that.
Most of these reflections are not within the 90-degree cone of a loudspeaker optimized for constant directivity. That point bears repeating. The first rear wall reflection is not within the 90-degree cone. The nearest-side wall reflection is also not within the 90-degree cone. The double-bounce reflections are not within the 90-degree cone. For that matter, the great majority of room reflections do not come from the 90-degree cone.
So where did this magic 90-degree number come from? I have no idea, except as a residue of sound-reinforcement practice. It's much wider than the high-quality listening area at the back of the room, and yet, it's not wide enough to cover the first group of reflections that arrive at the listener. Most of the early reflections are well outside the constant-directivity cone ... while the woofer is not particularly directional in the sub-500 Hz region.
More seriously, constant-directivity horns or waveguides typically have a very sharp drop in loudness at the edge of the pattern. In sound-reinforcement applications, the hard edge to the coverage pattern can be an intentional and desired property to prevent mike feedback, and assists in planning arrays for most uniform audience coverage.
For domestic applications, though, the hard edge is not an advantage. In the region outside the 90-degree cone, the direct-radiator woofer output can be as much as 10 to 15 dB higher than what's coming out of the horn or waveguide. This can be checked by walking in circle around the loudspeaker when music is playing. What do you hear at the sides? Woof, woof, woof. That's what's going into the rear or side wall reflection, a very bass-heavy version of the first-arrival sound.
Non-constant-directivity horns like the LeCleac'h have a very soft edge to the forward pattern. This means when you walk around the speaker, you can still hear the horn, even when you're far off-axis. This also means that there's no sharp spectral discontinuity between the direct-radiator and the horn as you walk around the speaker; this is very different than, say, a conical, where the horn seems to blink off when you walk outside the 90-degree cone, but the woofer is still rumbling away.
Taking a step back, what happens when we compare the direct-arrival spectra of conical & direct-radiator woofer to the overall power into a sphere? If the direct-arrival is equalized flat (following standard practice), the spectral power into a sphere has a sharp discontinuity at the woofer/horn crossover, since the horn only covers a small portion of the sphere. Put another way, most of the reflections have a sharp drop in the mids and highs, beginning at the HF crossover frequency.
This aids intelligibility, useful for HT dialog, but also reduces the ambient impression, which is a serious problem for playback of classical music.
Last edited:
The last coat of paint goes on my new midbass horns (and more back chamber tweaking) this weekend maybe I'll post pictures and details. It's a short hypex 75 Hz horn loaded with JBL 2225 and good in my room from around 80 to 1K. I originally built it for the Altec 515G but found the JBL was much better (cleaner) above 300 HZ. I'll call the thread "Beyond the Fart Boxes"
Probably came from the paper of of Kates, "Optimum Loudspeaker Direction Patterns"
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&frm=1&source=web&cd=4&cad=rja&ved=0CD4QFjAD&url=http%3A%2F%2Fdecoy.iki.fi%2Fdsound%2Fambisonic%2Fmotherlode%2Fsource%2FOptimum%2520loudspeaker%2520directional%2520patterns_Kates_1980.pdf&ei=CeiPUsK4OrGA2QXJyIGQAQ&usg=AFQjCNGd5Vv2KSkoqOo7F2t_5uq0tkUZmQ
I notice that the Kates paper is based on listening in an anechoic chamber ... which is not a realistic assumption for 99.9% of listeners. In the following paragraph, the presence of room reflections are used as an argument for narrower directivity, neatly side-stepping any discussion of real loudspeakers made with combinations of direct radiators and horn/waveguides.
I think Linkwitz is on more solid ground, with loudspeakers that are uniform in directivity over the working bandwidth, thanks to dipole radiation.
I think Linkwitz is on more solid ground, with loudspeakers that are uniform in directivity over the working bandwidth, thanks to dipole radiation.
Last edited:
Likely done to study the direct wave for localization/imaging rather than to judge timbral accuracy. You have to reduce variables sometimes to gleen what you're looking for.
In a room, these directional patterns would reflect from opposite walls for long delays, an optimal reflection scenario according to Geddes, which he documents on his site.
Hear, hear ...
If the quality of the direct sound is not there, then the horse has bolted ... trying to fix things after "bad" sound has been projected into the room is almost guaranteed to fail, your ears will always tell you that you have compromised sound, no matter how you try to fiddle with or adjust the tainted sound now bouncing around ...
The spectra of the reflection from the opposite walls is of considerable interest. Arguably, that's what audiophiles are adjusting when they rotate their speakers ... the spectral contribution from the rear wall versus the opposite walls. If a loudspeaker has sharp edges to the directivity pattern, this rotation would have to be done carefully.
Part of the discussion about bass horns might be more about changes in directivity versus frequency than the distortion and headroom of the bass horns. Direct-radiators are nearly omnidirectional by 300 Hz, a big change from what the MF/HF horn is doing at 1 kHz.
I don't agree. They (I, at least) turn them to lock in imaging dictated by the direct wave, as in Kates. The fact that it also benefits reflections a la Geddes is a bonus. I do agree that sharp edges to the directivity pattern is undesirable. I don't think anyone would disagree with that. However, Geddes argues that his controlled directivity is required to take full? advantage of Kates' toe in. But I have never heard a conventional speaker that didn't benefit from toe in.
Last edited:
How is the LeCleac'h horn any different than the constant directivity waveguides near the crossover region? Both are radiating into an angle defined by the horn.
In terms of radiating patterns, the only difference between the LeCleac'h and a constant directivity horn is the HF direcitivity, say above 1 or 2 kHz.