Why horn-loaded loudspeakers are subject to design tradeoffs

[Chris A]

By the way, I might test my '79 Cornwalls for IMD to get a baseline on the K-33 in a vented box, since I'm not aware that any IMD measurements have been shared for this bass bin.

My thinking is that a rather good two-way can be made with a 12 or 18 inch woofer (one of the lower-THD types) in a suitably sized vented box and a K-510 or K-402 on top with perhaps a Faital Pro 2 inch compression driver (or something better). This might work better as a center speaker between my Jubs than a tri-amped Belle...it would be an interesting project to see if that works as well as the numbers/measurements might indicate.

Edited August 18, 2014 by Chris A

[Chris A]

I'll continue the discussion on IMD by discussing...

...the "derating" of higher output capability loudspeakers for home use in order to further reduce IMD (among other improvements in performance), and also...

...a discussion of HF drivers (and horns, as the case may be) and FMD effects. It's always been interesting to me why so many audiophiles revel in the "full-range driver" world--including but not limited to Nelson Pass, who developed his First Watt amplifiers and other electronics almost exclusively for this market.

Of the typical problems encountered with discussing this type of subject is the tendency to drive the conversations toward false dilemma and away from dialogue (and to the more commonly assumed interchange via debate) resulting in a zero-sum game outcome instead of mutual learning on the original subject. I understand the dynamics of why these detours occur--and they mostly stem from distrust or perhaps other affective needs of those that resort to debate instead. In this thread I'd rather continue dialoguing on design tradeoffs of horn-loaded loudspeakers, which is of course the title of the thread.

On the admittedly advanced subject of design tradeoffs in general, which I've found to be more than often dominated in practice by human affective needs in "achieving one's mark on design" rather than by achieving outstanding product outcomes, I've found that looking at successful design approaches and studying those is required reading in order to find successful future approaches.

That's why I examine PWK's, Roy D's, and others' horn-loaded loudspeaker design methods, patterns, and insights closely on their products' use, looking for design patterns and product design approaches that have worked and that continue to work. I believe that these individuals have produced successful designs. [by way of example: the same kind of technique is used today in design of complex defense systems/products and is iconically expressed in quoting from Clausewitz or Sun Tzu, etc. YMMV.]

All this is to say that there are a couple of design insights/approaches to horn-loaded loudspeaker design that have worked very well, one of which is "derating" loudspeaker designs (electrical and mechanical components) so that they perform robustly in service and break down rarely, even if abused (relative to market-competing products). PWK clearly did this with his personal designs and so do most designers of commercial products in the audio industry. This technique isn't used much by other manufacturers in the consumer industry, who seem to be driven almost exclusively by "look and feel" and low manufacturing costs, ostensibly in order to be able to maintain high profit margins on their products by producing cheap-to-manufacture products, then marking the sales price up--something Bose is noted for doing.

One of the foreseen benefits of horn loading drivers is the relative difficulty that consumers have in "testing to failure" the acoustic drivers and electrical networks relative to direct radiating driver designs. Khorns, Heresies, Cornwalls, La Scalas and Belles are noted for their toughness in service, and if you want to destroy a driver or network, it's going to be very, very loud for a long time before you succeed. This design approach cannot be overlooked

However, a more subtle effect occurs by horn loading that PWK recognized and wrote about in the 50s-80s: low modulation distortion inherent in horn-loaded designs, particularly woofers (i.e., NOT subwoofers), but also present in midranges and tweeters to a great extent. This allowed PWK to design a two-way Khorn initially, and was the design's hallmark until the recording industry increased the bandwidth of recorded music from 12 KHz to 20 KHz, thus forcing a three-way design of the Khorn in the late 1950s, and in subsequent company loudspeakers due entirely on the inability of single midrange-to-high-frequency drivers to cover this passband (subsequently overcome by compression driver manufacturers in the early 1990s with the introduction of titanium and beryllium diaphragms and neodymium-based rare earth magnets). It's notable that the design of the Jubilee (for home use) reverted back to the original two-way design that PWK wished to return to, and was accomplished by the design of two advanced horns by Roy D. for these new full-range two inch compression drivers: the K-510 horn first, to be followed very soon by the larger K-402 horn, which accomplishes the design goal of crossing the two-way speaker at a low enough frequency (about 425 Hz) to avoid diffraction effects from the Jubilee bifurcated horn bass bin that becomes apparent at higher frequencies.

Intermoduation non-harmonic frequency products in a FFT plot

Edited August 18, 2014 by Chris A

[Chris A]

Continuing the discussion on derating with loudspeaker designs, and considering the Klipsch Jubilee corner horn...

Additional benefits of using a "derated" loudspeaker design for home use is illustrated most clearly via:

• extremely low modulation, compression, and impulse distortion across the entire audible band (31-16000+ Hz) even though it is only a two-way design
• the elimination of the complexity and cost of an additional tweeter and crossover network, along with the corresponding elimination of time delay issues that must be matched to within 1/2 inch of driver delay (actually within 1/10 of a wavelength at crossover, within ~7 microseconds of being in-phase of driver/horn phase and path length variations), and matching relative SPL polar outputs in both horizontal and vertical directions in crossing from midrange-tweeter horns/drivers. All this is enabled due to the reduced output SPL required of the home version of the Jubilee relative to its commercial brother - the Jubilee 535 three-way design
• controlled directivity of vertical and horizontal polar output to 90 degrees from lower midrange up through the highest audible frequencies
• virtually infinite life of components due to their derated design input electric power and output SPL, along with virtually no driver compression due to thermal or acoustic effects

The results speak for themselves. One point should be made: the quality of the high frequency compression driver will control the overall fidelity of this design due to the current titanium-diaphragm compression driver chattering at or above 10 KHz and poor inherent damping of these type of 2-inch compression drivers.

The only real downside of this design is size of the enclosure and the realization that placing these loudspeakers in room corners wll increase their lf extension by almost one octave: something that the consumer should consider carefully when placing them in-room. The results sonically speak for themselves: it's been said that this design can outperform loudspeakers costing even 10x more.

Edited August 18, 2014 by Chris A

[tromprof]

The last couple of pages have been particularly interesting.

Thanks Gentlemen !

[CECAA850]

Looks like the dog is in the sweet spot. He's no dummie.

[DrWho]

I'm not sure I'd pick Klipsch as the poster child for sufficient engineering derating. The pro stuff has always been so easy to blow up - not to mention Klipsch speakers tend to sound extra nasty when driven near the limit (I think that's true of all their lines actually - not just pro). Sure, there is merit to saying don't run above the red line, but when talking derating, Klipsch has always been near the bottom of the chart for comparably priced/spec'd speakers. JBL is near the top.

Ironically, where the Klipsch stuff gives up first is in the xovers...and then quickly followed by the compression driver blowing up from the lack of low frequency protection. Some of those reactive voltages inside the xovers can get quite insane...and those ceramic resistors are so easy to crack and fall apart from normal handling.

Edited August 18, 2014 by DrWho

[Chris A]

Thanks for the clarification...the definition of derating used above isn't in design derating of parts but in consumer use of commercial product lines, where the limits of the hardware used are far, far lower on average in my experience and in actual use than commercial. 

If you're blowing up commercial stuff at home then my guess is that you know your local police patrolmen by first name

EDIT:I don't share your assessment of JBL's consumer-based products, but JBL really hasn't been in that market since the 80s. Foam surrounds and polymer laminate and painted drivers aren't very robust.

Edited August 18, 2014 by Chris A

[DrWho]

Just to be clear, I was referencing the JBL professional lineup and comparing to the Klipsch professional lineup - which is the only place that the SPL limitations of a system should even be approached.

Unfortunately increasing the power output of a speaker isn't just a cost exercise - there is almost always a performance tradeoff.

The classic example is that of voice coil power handling. To get higher power out of a voice coil you need more surface area - and that means a larger diameter. For the same number of turns, this means you have a heavier moving mass and this reduces your high frequency mass corner. There is some extent to where you can increase the strength of the magnetic field, but drivers are already running well into saturation. The only way to get more Bl is to make the gap smaller....good thing a smaller gap helps with the heat a little bit too (metal conducts heat better than air).

To make the gap smaller, the voice coil position needs to be better controlled. Some of this is a cost exercise in manufacturing tolerances, but the majority of it is handled with suspension geometry and/or stiffness. Higher stiffness raises the LF corner. More elaborate geometries tend to increase the Q of the suspension resonance - eventually causing a passband suckout. You can damp the suspension, but then this increases your suspension losses (at some point requiring more voice coil power handling and thus negating the benefit).

The higher excursions also require a stiffer diaphragm and that is addressed with the type of material and its thickness. Again you run into mass corner tradeoffs as well as lowering the natural Fs of the system. Another problem with stiffer diaphragms is that they ring easier due to the higher mechanical Q. I know Berrylium is a popular choice, but it is not the best damped option. Other secondary effects include intermodulation distortion from out of band ringing (triggered by the distortion in the system). The larger the diaphragm, the closer those resonances get to the audio band and the easier unwanted frequencies modulate into the audible band.

With the larger diaphragms and larger excursions, you have way more pressure modulation happening between the diaphragm and the phase plug. The non-linearity of the air in this region is a major source of distortion in compression drivers. At some point you gotta move the phase plug further away to lower the distortion, and then you sacrifice HF performance (both polars and distortion due to reduced coupling). You are also limited by the acoustic resonances that suck out the HF information over narrow bands. What looks like a peak in frequency response in the old K55 drivers is actually a phase plug resonance create a big notch filter - thus making the natural acoustic rolloff appear to be a spike. (Ironically people try to "cure" that with another notch filter in the xover).

So yes, by having an increased power handling speaker and coasting along at low levels, we have reduced the nonlinearity from voice coil heating and chances are our suspension is more linear at the lower excursions. However, we're picking up other tradeoffs in the system that may not outweigh the advantage.

Of course all of this (minus the phase plug) is true for any type of speaker and is a rather textbook understanding, but for a given application you're always better off designing the tradeoffs to what you actually need rather than overkilling some aspects and letting the dust just settle wherever.

One of the reasons I've preferred the Klipsch PA speakers is because they're designed to sound better within the limits of the system. I'm also not an SPL junky and don't struggle with fader creep. And because I don't have issues with fader creep, people's ears aren't "turning off" and things actually sound louder even though they're not...and I can take advantage of the cleaner sound at lower SPL's. I've started measuring the crest factor of my mixes and it turns out I'm usually pushing over 40dB for rock and roll (compare that to your recordings, hah). At my church where things are more classically influenced we're often over 60dB. At some point you do need that crazy extra headroom, but at that point I think it's more important to have a system that clips pleasantly since those peaks are such short duration.

[ClaudeJ1]

Wow. Where is mr klipsch's yellow button when you need it?? Experts galore and no common sense in sight......sigh!

LOL. My comment about the point source was designed to be a joke of sorts. You know, like a Flux Capacitor in "Back to the Future"

[Khornukopia]

 

 

The fact of the matter is that you can use a horn as a funnel, so they are better. Who can argue with that? Fact!

 

 

Maybe now I can think of these as a horn-loaded headset.

 

 

 

Just trying to be funny. 

[twk123]

Think about it like this: you need to fill a large area with sound. What is the best way to do it? Horns, obviously. High output with a moderate amount of power. But when you get into smaller rooms, there are other options which sound every bit as good. Now, you're probably going to need 200 wpc, but that kind of power is relatively inexpensive these days -- the likes of QSC and Hypex deliver amazing sound for the money

 

Like this?

Edited March 20, 2015 by twk123

[ClaudeJ1]

 

Wow. Where is mr klipsch's yellow button when you need it?? Experts galore and no common sense in sight......sigh!

LOL. My comment about the point source was designed to be a joke of sorts. You know, like a Flux Capacitor in "Back to the Future"

 

Actually, Danley comes closest to a point source.

[ClaudeJ1]

On 7/12/2014 at 6:19 PM, Max2 said:

I think its time all the quasi-techno engineers build something better. Time alignment, peaky cycle ranges, yeah i guess my K-horns have them. No low end with my LaScalas's, got that issue too. Just build a horn cab close to same size, same good looks and better performance for the K and L and then we will have something. Until then....

It's already been done, but the Khorn has been the real Benchmark for decades.

[Travis In Austin]

13 minutes ago, ClaudeJ1 said:

It's already been done, but the Khorn has been the real Benchmark for decades.

Did you find a flux capacitor and Go Back to the Future?

 

This thread is 6 years old.

[ClaudeJ1]

1 minute ago, dwilawyer said:

Did you find a flux capacitor and Go Back to the Future?

 

This thread is 6 years old.

No, but it was part of the daily digest that someone (the Autoformer guy who recently discovered Khorns) alerted me to. Remember: "The excellent is new forever".................Ralph Waldo Emerson.

[Travis In Austin]

On 3/23/2015 at 9:11 PM, ClaudeJ1 said:

LOL. My comment about the point source was designed to be a joke of sorts. You know, like a Flux Capacitor in "Back to the Future"

I though you might of actually found one.