If a horn is used between a waveguide and a direct radiator the priorities will be different. A smoothly increasing directivity can be useful and the trade off may be less obvious at the lower frequencies. Then there's loading and efficiency.
They may honk if not crossed as designed but when used within sensible constraints, can sound fine.
The system design vertical directivity is limited so should you be concerned with nulls outside this?
Wouldn't it be better to increase vertical spacing toward this limit and use it to produce a waveguide that performs better?
I personally believe there is a "sweet spot" where the vertical size (setting the lower limit) and the position of the nulls match. Too large a horn provides good pattern control but prohibits having close enough spacing so nulls cut deep into the pattern. Too small a horn has limited pattern control, so the beamwidth increases too far, even greater than that of a larger horn with taller coverage pattern. But just like the story of the three bears, there is a size that's just right for the desired beamwidth and crossover point.
Wayne:
Two questions. Take the four Pi for instance. Looking at measurements, it seems like the horn works down to and is crossed around 1600 Hz. Why don't you use a bigger horn that covers more area?
Secondly I'm wondering about the driver for midrange/bass and it's usage. There seems to be a consensus among DIYers that a JBL 2226 driver don't work very well above about 600 Hz. What's your take on that?
Two questions. Take the four Pi for instance. Looking at measurements, it seems like the horn works down to and is crossed around 1600 Hz. Why don't you use a bigger horn that covers more area?
Secondly I'm wondering about the driver for midrange/bass and it's usage. There seems to be a consensus among DIYers that a JBL 2226 driver don't work very well above about 600 Hz. What's your take on that?
You have to remember that this effect happens only over a very narrow bandwidth right at the crossover. You can't use it to "produce a waveguide that performs better".
The question is are the nulls wide enough so as to not be a problem. The lower the crossover the less of a problem it will be, but then the larger the waveguide needs to be. A smaller vertical dimension of the waveguide makes the crossover have to move higher and the narrower the nulls are for a given spacing. If you decrease the vertical dimension but do not raise the crossover then the device narrows (it doesn't widen until much lower in frequency) making the main beam narrower. This is simply not a problem for which there is a good solution, although, if you defeat the floor and ceiling reflection s like I do then it is not an issue (unless the beam gets very narrow, which need not be the case in a proper design.)
Crossover isn't that high. The crossover region is about 900Hz to 1300Hz, where both midwoofer and tweeter contribute sound. If I was going to describe the crossover "point" as a single frequency, I'd say it was about 1100Hz.
I choose the crossover using a method described in the link below:
In a nutshell, I select the crossover region where a few things are simultaneously accomplished:
1. The acoustic centers are aligned
2. The horizontal directivities match
3. The vertical nulls are outside the tweeter's beamwidth at HF (e.g. above about 3kHz)
I don't agree that this is any sort of consensus. In fact, I think one of the 2226 driver's greatest strengths is its ability to provide smooth response above that point. Just about any 15" woofer is good to 600Hz - The 2226 is set apart from many others precisely because of its HF performance.
I will add that there are some other "DIY darlings" that are used to relatively high frequency that don't work nearly so well. I've tried some of the boutique drivers that some DIYers love, mostly because they're made at job shops by passionate craftsmen, and I like working with those sorts of people. But frankly, the measurements of those boutique drivers are terrible.
It's all about cone and suspension damping, and I suppose that's where the economy of scale from the larger corporate shops matters most. They don't have to depend on vendor OEM cones and suspension, and can specify exactly what material and shape they want them to be. This is also true for motor structures, but I think the R&D facilities are more important there, not so much the manufacturing capability. A large shop can afford magnetic FEA software to optimize the motors, where the smaller job-shop cannot.
Another aside: I've tried using compression drivers at much lower frequency, like some DIYers do. I've found it sounds OK up to a few watts, but transients really stress the drivers when pushed too low. They definitely cannot be used at a continuous high power level, but even in a home hifi environment where that isn't likely to happen, the transients still sound "crunchy" to me when a compression driver is run too low, even on a suitably large horn.
So where I wan to crossover the woofer lower, I like to split the band with a cone-driven midhorn. I really like to split the band into three decades, 20-200Hz, 200-2kHz and 2-20kHz. With a cone driven midhorn, I can get really close to this ideal.
But for a DI-matched two-way loudspeaker, I tend to crossover around 1kHz where the three conditions I mentioned above are met, and use the process described in the link above to dial it all in.
Thanks for the explanation. I assumme this described experience of yours with drivers operating outside their comfortable zone is limited to 1" compression drivers?
A 2" compression driver like JBL 2446 or similar should be able to be crossed substantially lower. So why not use a 2" driver in a bigger horn for a two-way speaker?
No 2" compression driver can get to, let alone go above, 10 kHz. I have tested many 2" drivers and they just can't do it - unless the horn beams. That's because their power response is dropping fairly fast above about 7-8 kHz so in a constant directivity system you cannot do a 2-way with a 2" driver.
2" *exit* compression driver. 😉 (..there are a few 2" diaphram compression drivers with smaller exits.)
-this beg's the question though:
How important is any sort of "constant directivity" above 7 kHz?
That's a fair question. On par with: How important is it to take the crossover down below 700-800 Hz., which for Hi-Fi is easily done with a 1" driver. If the answer to the second is "It's not", then the answer to the first doesn't matter.
I'm thinking in terms of level differences here. In that respect then something as low as 400 Hz might be important.
But,
..level differences above 6.5-7 kHz (listener-dependent) with only a modest angle for loudspeakers relative to the listener?
I think at the very least we can say it's *less* important above 7 kHz (..until about 13-14 kHz up, to perhaps 15 kHz before becoming far less of an issue again).
As a point of practice, I know more than a few systems where the owner uses a super tweeter firing reward above 7 kHz (usually closer to 9 kHz).. and a 2" exit compression driver with a horizontally "constant" directivity down to 500-600 Hz - and it seems to work very well perceptually.
Last edited:
I think the Klipsch Jubilee is an example where it seems to work very well and there isn't that much beaming.
Firstly, many seem to disagree that a 1" driver works well to 700-800 Hz. Isn't that pushing it out of it's optimum performance area?
I've little doubt that there's a clear advantage in lowering the crossover. Ideally, below the vocal area. Those speakers I've heard that has a very low crossover are simply better in some aspects to my ears. Sound is more coherent and the vocal is more faithfully reprodruced.
Hello,
How important uniform directivity being the subject and the Le Cleac'h horns having been mentionned I think I have to say few words.
1) few exception apart (e.g. based on interference between a rear wave and a front wave as the dipôles) no multiway system provides uniform directivity all along the frequency range 20Hz to 20kHz. The one using a direct radiating loudspeaker in the bass range and a waveguide in the upper frequency range all of them present a huge variation of the radiation angle beween 500Hz and 2000Hz.
2) the shape of round Le Cléac'h horn is not based on any esthetical choice, only parallel isophase wavefronts and a wavefront area obeing to the Webster's equation solution is the source of their shape (the roudover at mouth is not added but results naturally form the calculation. (BTW Webster's approach still prooved to be largely useful even if it is only an approached solution to the wave equation.
3) If you don't like round horns you may prefer other horns I used to design.
Horns
Horns
(All the Le Cléac'h horns have calculated with parallel isophase wavefronts and a variation of the wavefront area with distance to throat being a Webster's solution)
4) control of directivity became in the last 10 years a marketing issue for many companies (specially US companies). In very few cases (e.g. record studio monitoring) there is a need for constant directivity.
5) you cannot have both, smooth pressure field Inside a horn and constant directivity (despite claims ). You cannot have smooth polars and control of directivity. There is always diffraction problems associated with the control of directivity (this cause the non smooth polars and frequency responses at different angles).
6) for Hifi purpose still many people prefer unequalized sound than the equalization required by control directivity.
7) round Le Cléac'h horns should not be listen to with parallel axis. Their axis must crossed some distance ahead of the best listener place. Le Cléac'h horns should be listen to in near field to reduce the effect of the (untreated) room signature having a different spectral balance than the direct sound.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Last edited:
Thanks for the explanation. I assumme this described experience of yours with drivers operating outside their comfortable zone is limited to 1" compression drivers?
A 2" compression driver like JBL 2446 or similar should be able to be crossed substantially lower. So why not use a 2" driver in a bigger horn for a two-way speaker?
I have over a period of some four years now compared a wide selection of horns and drivers and as such found a reasonably absence of "free lunches" - as expected. First of all, in a high-end system I cannot see any two-way system as the optimum solution. The classical combo of a 15"/radial horn xoed at anywhere from 600 - 1.000 Hz does a pretty good job, but still, the compromises are far too big. IMO. Any 2" driver (or 1,4" for that matter) has no business over 7-8k regardless of what measurements etc. show. A substancial part of the output above this range will be breakups and distortion. Yes, some drivers/diaphragms work better than others but the laws of physics still apply.
At an early stage I had high hopes for a combination of JMLC-200T and Altec/GPA 288 - that combination worked like a charm all the way down to 350 Hz with 24 dB LR cutoff. The grunt and growl was fabulous, but above 7k approx. the beaming was far too big and above 10 k nothing much happened.
A horn systems worst enemy is stretching the working area of both the driver and horn in any direction. Shure, a 2" driver can work pretty well all the way down to below 500 Hz if properly xoed and with the right horn size, BUT, do not expect any sensible output in the treble region. I have tried, tested and measured this with such different drivers as the GPA 288, TAD 4001, JBL 2446, GPA 399 and ALE 7550.
Did you ever hear a well designed complete system using JMLC profile horns and high quality drivers?
In my opinion, all in all, there is no smoother horn profile to be found. I have here JMLC profile horns with the full rollback feature, in sizes from 40" (covering 350 - 3k easily) to the tiniest 6" working from 2k++ and upwards, again, tried tested & measured a LOT of combination and I can honestly say I have never heard better sound from any horn. Period. As long as you respect the frequency limitations and remember that a horn is basically a bandpass system with a fairly limited range per unit I have not been able to produce "bad" sound from any of the seven different size horns. Never.
Given the proper sized waveguide, I have not found any problems in home use. I wouldn't use a 1" driver down that low in a sound reinforcement situation - which is where the manufacturer issues a warning about going below 1 kHz. In home use you are never going to play a compression driver so loud that it would be a problem.
There is no extra space. All of these 2-way speakers aiming for constant horizontals with a waveguide are already at or beyond the vertical spacing "limit", depending on who you're asking about which speaker.
- Status
- Not open for further replies.