KLH 5 fatal design flaw

[Steve F]

I’ve wanted to talk about the KLH 5 for about 40+ years, but I’ve never gotten around to putting my thoughts down on (digital) paper. It’s an interesting speaker, pretty good-sounding, an excellent value in its day ($179.95 ea. vs. the 3a’s $250 ea. and the AR-5’s idiotic $175 ea.). I never thought its bass was as good as the 3a’s, but I’d have happily lived with a pair of KLH 5’s if AR hadn’t existed. 

But those dual midranges. Those dual midranges. The driver itself was a truly terrific driver. Wasn’t it essentially identical to the full-range driver in the various KLH table radios? It had amazing excursion for a small driver and could produce real bass on its own as used in the radio. One of audio history’s great drivers. 

But the KLH 5 was designed in the early 60’s, well before the time when driver placement and interaction was fully understood. I’m not criticizing the design decisions that were made at the time—the industry’s collective knowledge base and understanding simply wasn’t that advanced in 1964. However, in retrospect, those dual side-by-side midrange drivers are just so, so wrong.

First of all, as any serious student of speaker design knows, when two drivers are reproducing the same frequency range, they behave as if they were an oval driver of their combined dimensions as far as directivity is concerned. So two 3 ½-inch drivers side-by-side perform the way a 3 ½ x 7-inch oval driver would perform, from a dispersion/directivity standpoint. For the KLH 5, the 7-in dimension is in the horizontal plane, so its horizontal midrange dispersion will be as narrow as a 7-inch driver’s dispersion would be. Do the math: 13560 (speed of sound in IPS) divided by 7 = 1930Hz. That’s where the mids become badly directional—well short of the crossover to the tweeter.

These days, speaker designers use this knowledge intentionally, as in when they do an M-T-M “D’Appolito” design, vertically-arrayed. With the typical 4- or 5-inch mids being 12 inches apart (because there’s a tweeter in between them), the vertical dispersion is drastically limited after around 1100Hz, which is a good thing (to them): The design limits the vertical “scatter” and “splash” off of the ceiling and floor, while preserving wide horizontal dispersion (because the 4-in dimension governs the H dispersion). M-T-M proponents feel you get both good imaging (because the speaker’s output is more ‘focused’ at the listener’s ears) and good horizontal dispersion. This is precisely why early THX-certified speakers were M-T-M designs: to enhance dialog intelligibility by limiting unwanted, uncontrolled vertical “scatter,” while maximizing listener seating flexibility with wide horizontal dispersion.

But there’s another demon lurking in the side-by-side driver arrangement: The drivers’ output will overlap and interfere with each other, producing a “picket-fence” output of hot-cancellation-hot-cancellation-hot, etc. The two drivers’ outputs will alternately reinforce and cancel each other, depending on the angle, time and phase of their output. If you take a horizontal polar graph of a side-by-side arrangement, it looks like “fingers” rather than a smooth semi-circle. If you’re sitting in line with a finger, you hear it. If you move your head into a null, you lose it. Very disconcerting, and it’s why good speaker designers these days never put two drivers in the midrange side-by-side. You can stack them vertically, because the vertical plane isn’t as critical to the listener as the horizontal plane.

The KLH 5 is all “wrong” in the midrange. As a matter of fact, if you were listening in the near-to-critical field, you’d be better off with the KLH 5 placed on its side, so the two mids were vertical, not horizontal. That placement would throw all of the offending interference and narrow dispersion into the less-critical vertical plane and would optimize the horizontal plane.

Note that only 12 short years later, when the AR9 was being designed, engineers had gained an order of magnitude’s greater understanding about directionality, driver interference, etc. The AR9 was a remarkably advanced speaker, literally light years ahead of the KLH-5-6, OLA, 3a/2ax/11/10 Pi, etc. Light years. 

Also, it’s important to note that the KLH 5’s midrange problems will be far less noticeable in the reverberant far field than when listening closer up. Nonetheless, those dual side-by-side mids are a fatal design flaw. I’m surprised that no reviewer, audio historian or critic has mentioned them before, at least to my knowledge. That stone is now officially overturned.

Steve F.

[MOEB74]

I dont know anything about speakers, but I have a random question. Does the Model Fives midrange driver has visual excursion? Ive never seen it move at all...

 

Thanks

[JKent]

3 hours ago, Steve F said:

I’m surprised that no reviewer, audio historian or critic has mentioned them before, at least to my knowledge.

That's because they sound so great. And AFAIK no classic speaker was ever intended for "near field" listening.

The technical argument supporting the "all wrong" mids reminds me of the hot air balloonists who were blown off course and found themselves near a building with some people on the roof. One balloonist shouted out; "Where are we?" and the rooftoppers yelled back "You're in the gondola of a hot air balloon!"

The balloonist turned to the passenger and said "Oh. We're in Redmond, WA and that's Microsoft HQ." His companion asked "How do you know that?" "Well," he answered "like their tech support, the information they provided was technically accurate but practically irrelevant".

-Kent

[Steve F]

They do sound pretty good. But as detailed, technically-aware speaker aficionados, with an eye on history as well as engineering, and an understanding of when/how this all happened, we find a discussion of this type interesting nonetheless. At least, I certainly do. The 3a is similarly poor when measured/listened to too close.

The 'first arrival' contingent of the speaker universe maintains that the instantaneous anechoic response of the speaker is what first impresses upon your hearing, from which you form your impression of tonal quality, clarity, etc. This happens--according to them--regardless of listening distance, because 'first arrival' is always the shortest distance to your ears, so you'll hear 'first' whether you're 4 feet or 12 feet away from the speaker. By that belief system, some of any speaker's sonic signature for tonal balance, clarity (which includes phase interference and overlapping drivers), distortion, etc. will come from their first arrival sound.

Roy Allison did not believe in first arrival. He felt that far-field power response swamped all the first arrival sound and what people reacted to was far-field power response. Toole et al. believe the opposite. Andy Kostatos/Henry Kloss didn't put it into those explicit verbal terms, but they were first arrival/on-axis guys. The KLH 5's on-axis response was ok. The side-by-side mids would really become uncontrolled off-axis, because the distance to the listener's ears would be different for the two drivers, and that's where the time/phase problems would arise. Directly on-axis (like if the pair was toed in as was a not uncommon use case in a 1960's stereo system), the dual mids in a single speaker were the same distance from the listener's ears--regardless of how far away the listener sat.

The KLH 5 did indeed sound very good. But it didn't sound fabulously great and crystal clear, completely revealing of the smallest inner musical details. The "good not great' aspect of its sound was at least partially due to the side-by-side mids. And other things as well, for sure.

Steve F.

[Pete B]

Do any KLH 5 listeners find that the midrange is transparent in a way similar to headphone listening?

I have never heard them.

[Analogman]

On 11/14/2017 at 7:57 AM, Steve F said:

I’ve wanted to talk about the KLH 5 for about 40+ years

Bored?  Just feel like popping into the KLH section to slam and troll on what is thought by some to possibly be the best sounding (unsung) medium format KLH of them all?

(note KLH fans, I said SOME : - )

And not even a good slam that makes any sense, none whatsoever 

"fatal design flaw"

What a ramble based on nothing substantive or objective (and forget about anything even remotely resembling "technical") 

Try taking a look at the network schematic if you are able to understand what they did with the 5

Utter gibberish on every level

Sorry friend, but that's the truth

And by the way, I'm here, stumbled across and made the mistake of reading this travesty because I am looking for a pair of "5"s

Craig

[Martin]

As the happy owner of two pairs of Fives, each pair used in different rooms, I can report many  hours of listening to them with absolutely no awareness of any flaws, no fatigue and in fact, have no criticisms of any kind.  I'd say Henry got it just right, with two sweetly singing midrange drivers, which to me, sound better than the 2aX midrange and just as good as the AR 5 midrange.   To each his own. 

[Neward Thelman]

Martin - Stevievie's mostly full of crap. The only correct part in his post is that two drivers operating over the same frequency range will cause comb filtering - but it's not always guaranteed that the result will be bad. Also, two horizontally mounted drivers will interact in the same way - even if they supposed to be operating over different ranges - such as a midrange and tweeter - when they reproduce the same frequencies. That happens in the crossover area.

Lot's and lot's of classic speakers used horizontally placed drivers, but still produced excellent results. To name just a few, most of the classic Acoustic Research models - including the AR 3 series and the LST - all of the JBL 43xx series [such as the 4310, 4311, L100, etc.], the Rectilinear 3, and countless others. 

While it's true that vertical mounting's better, it's just not a concretely sure fact that horizontal mounting will automatically produce bad results - as the speakers I've just named demonstrate. 

There no "fatal" flaw in the Model 5. It remained in production for a span of 2 decades and has remained a favorite right into the present, acquiring admirers and a devoted following. Contrary to Spoviee's self-inflated pronouncements, the midrange is one of the Model 5's best qualities, and it's what drew listeners and sold the speaker from the beginning. Indeed, it's the rich, tube-like midrange that characterizes all of Henry Kloss's KLH designs and sets it apart from all of its competitors.