Loudspeaker directivity by Roy Allison

[tysontom]

"It will be seen that the mechanical reactance caused by the air chamber behind the cone is three times the mechanical reactance resulting from the suspension system. Therefore, in the range where the compliances are the controlling mechanical reactances the compliance caused by the air chamber is the controlling compliance. (Hence a high alpha or Acoustic Suspension system). This expedient reduces the distortion caused by a nonlinearity of the suspension system."

From the RCA patent. Comment within brackets is from me.

Of course, I don't expect contradictory facts to change your opinion in the least,

David

David,

In the nearly sixty years since Villchur (Villchur -- two els) introduced the acoustic-suspension speaker, there have been dozens (perhaps hundreds) of critics and naysayers -- usually retired-speaker designers -- trying to discredit his accomplishment of the acoustic-suspension loudspeaker. You should feel honored that you are not the first to say that Olson and Preston anticipated the acoustic-suspension system! But as is usually the case, most of these criticisms are oversimplifications (a good example IMO is your statement above, "hence a high alpha or Acoustic Suspension system") and eventually fade away.

The problem with Edgar Villchur's patent was that he wrote the patent himself, and he didn't use a patent attorney because he couldn't afford one. He therefore didn't do due diligence in performing an adequate patent search. Villchur's patent could easily have been written in such a way as to protect itself against "prior art," had there been adequate patent research. As it happened, Olson and Preston's patent (#2490466) was missed by Villchur and his patent examiner during his search.

Villchur has said that shortly after AR was established in 1954 to build acoustic-suspension speaker systems, there were two licensees, KLH the Heath Company. It wasn't long before articles began to appear attacking the validity of the air-suspension design, and in one article ("Design of a Wide-Range Ultra-Compact Regal Speaker System") the authors -- engineers from Electro-Voice, Avedon, Kooy and Burchfield -- wrote:

"…On almost every point of performance the small cabinet is at a disadvantage … when it is said air suspensions are inherently more linear than mechanical suspensions a misstatement has been made…." This article went on to say that the non-linearity of the air was "an unavoidable consequence of the laws of physics."

In this same article criticizing the "non linearity of air," these engineers also described an air-suspension system that E-V had just introduced. When AR found out about this development, the company sued E-V for infringement of the patent, and the E-V attorney produced the RCA patent (that had been filed by Olson and Preston several years earlier) as anticipating the acoustic-suspension patent. The Olson-Preston claim was for a compliant-suspension speaker, but there was no claim for a system that had a free-air resonance below the optimum frequency, as with Villchur's invention. This is the characteristic of the acoustic-suspension system -- actually requiring a small enclosure. The essential proposal of the RCA patent was to improve speaker bass performance, in distortion and frequency response, by extending the region of mass control downward in frequency. Olson and Preston accomplished this by means of a compliant, folded-rim suspension that lowered the speaker's fundamental resonance frequency and allowed large, linear voice-coil travel. Olson and Preston plotted the reduced distortion for their speaker only for infinite-baffle mounting, and they showed optimum frequency response for the same mounting. They also pointed out that if their speaker were mounted in a small, enclosed cabinet, the resonance frequency could be raised to a point that was "still desirably low," and they showed a frequency response for such a mounting -- much inferior to that with the infinite baffle in low-frequency range and in correct value of Q, but superior to results with an open-back cabinet.

Where Villchur's air-suspension design relied on a cushion of air behind it, the Olson and Preston design was shown as being better off without the cushion of air, but capable of tolerating it for non-high-fidelity applications. There was therefore never a question of Villchur's speaker infringing on the RCA patent -- since the latter made no general claim for a low-resonance speaker in a small cabinet. However, the court in South Bend, Indiana ruled that Villchur's patent had been "anticipated" by the RCA patent. Because of a likely protracted legal fight -- and in view of what happened in 1954 (suicide) to Edwin Armstrong, the father of FM radio after his protracted fight (also) with RCA -- Villchur decided not to appeal the decision. The feeling was that AR could prevail, but it was not worth the tribulation to do it.

David, I am sure you've read Frederick V Hunt's Electroacoustics. My favorite quote is on page 87, in which Dr. Hunt states:

"As is often the case with ideas that appear superficially to be simple, a good many loudspeaker designers discovered that they had already made this invention -- after someone else had pointed out what the invention was. For example, there is a tendency to label as a forerunner of the base-reflex principle any prior-art loudspeaker in which an incidental hole or side opening in the cabinet can be identified. The primitive disclosures of Finch (Pat No. 216840) and Colby (Pat No. 371551) are of this sort. Merrill and Hays (Pat No. 669944) went a little further by exhibiting an enclosure containing holes intended to 'prevent the formation of a dead-air cushion against a receiver diaphragm.'"

This seems to sum it up pretty well.

--Tom Tyson

[soundminded]

David,

In the nearly sixty years since Villchur (Villchur -- two els) introduced the acoustic-suspension speaker, there have been dozens (perhaps hundreds) of critics and naysayers -- usually retired-speaker designers -- trying to discredit his accomplishment of the acoustic-suspension loudspeaker. You should feel honored that you are not the first to say that Olson and Preston anticipated the acoustic-suspension system! But as is usually the case, most of these criticisms are oversimplifications (a good example IMO is your statement above, "hence a high alpha or Acoustic Suspension system") and eventually fade away.

The problem with Edgar Villchur's patent was that he wrote the patent himself, and he didn't use a patent attorney because he couldn't afford one. He therefore didn't do due diligence in performing an adequate patent search. Villchur's patent could easily have been written in such a way as to protect itself against "prior art," had there been adequate patent research. As it happened, Olson and Preston's patent (#2490466) was missed by Villchur and his patent examiner during his search.

Villchur has said that shortly after AR was established in 1954 to build acoustic-suspension speaker systems, there were two licensees, KLH the Heath Company. It wasn't long before articles began to appear attacking the validity of the air-suspension design, and in one article ("Design of a Wide-Range Ultra-Compact Regal Speaker System") the authors -- engineers from Electro-Voice, Avedon, Kooy and Burchfield -- wrote:

"…On almost every point of performance the small cabinet is at a disadvantage … when it is said air suspensions are inherently more linear than mechanical suspensions a misstatement has been made…." This article went on to say that the non-linearity of the air was "an unavoidable consequence of the laws of physics."

In this same article criticizing the "non linearity of air," these engineers also described an air-suspension system that E-V had just introduced. When AR found out about this development, the company sued E-V for infringement of the patent, and the E-V attorney produced the RCA patent (that had been filed by Olson and Preston several years earlier) as anticipating the acoustic-suspension patent. The Olson-Preston claim was for a compliant-suspension speaker, but there was no claim for a system that had a free-air resonance below the optimum frequency, as with Villchur's invention. This is the characteristic of the acoustic-suspension system -- actually requiring a small enclosure. The essential proposal of the RCA patent was to improve speaker bass performance, in distortion and frequency response, by extending the region of mass control downward in frequency. Olson and Preston accomplished this by means of a compliant, folded-rim suspension that lowered the speaker's fundamental resonance frequency and allowed large, linear voice-coil travel. Olson and Preston plotted the reduced distortion for their speaker only for infinite-baffle mounting, and they showed optimum frequency response for the same mounting. They also pointed out that if their speaker were mounted in a small, enclosed cabinet, the resonance frequency could be raised to a point that was "still desirably low," and they showed a frequency response for such a mounting -- much inferior to that with the infinite baffle in low-frequency range and in correct value of Q, but superior to results with an open-back cabinet.

Where Villchur's air-suspension design relied on a cushion of air behind it, the Olson and Preston design was shown as being better off without the cushion of air, but capable of tolerating it for non-high-fidelity applications. There was therefore never a question of Villchur's speaker infringing on the RCA patent -- since the latter made no general claim for a low-resonance speaker in a small cabinet. However, the court in South Bend, Indiana ruled that Villchur's patent had been "anticipated" by the RCA patent. Because of a likely protracted legal fight -- and in view of what happened in 1954 (suicide) to Edwin Armstrong, the father of FM radio after his protracted fight (also) with RCA -- Villchur decided not to appeal the decision. The feeling was that AR could prevail, but it was not worth the tribulation to do it.

David, I am sure you've read Frederick V Hunt's Electroacoustics. My favorite quote is on page 87, in which Dr. Hunt states:

"As is often the case with ideas that appear superficially to be simple, a good many loudspeaker designers discovered that they had already made this invention -- after someone else had pointed out what the invention was. For example, there is a tendency to label as a forerunner of the base-reflex principle any prior-art loudspeaker in which an incidental hole or side opening in the cabinet can be identified. The primitive disclosures of Finch (Pat No. 216840) and Colby (Pat No. 371551) are of this sort. Merrill and Hays (Pat No. 669944) went a little further by exhibiting an enclosure containing holes intended to 'prevent the formation of a dead-air cushion against a receiver diaphragm.'"

This seems to sum it up pretty well.

--Tom Tyson

I think one of Villchur's problems what that he didn't fully understand his own invention. His thermodynamic explanation just doesn't fit because it is not an applicable description even though the laws of thermodynamics apply. As he was not a trained engineer or scientist as I understand it, this is not surprising. It took me nearly 30 years to understand it after it had been explained to me and I had seen it. The explanation for the air cusion is only partly correct and it is incomplete. The remaining part is the aerodynamic damping provided by the stuffing. Villchur's invention is a classic application of the mass-spring-dashpot equation used for so many things such as automobile suspensions for one. The mechanical tuning of this arrangement allows the designer to choose any arbitrary f3, any arbitrary cabinet size and adjust the cone/voice coil mass and damping material to achieve those results. Had the system been described that way, it would have been clear that Olsen's patent was off in an entirely different direction. Also the application of the restoring force over the surface area of the cone by the air pressure is entirely uniform eliminating the unevenly applied shearing forces that result from mechanical spring mechanisms. This lack of shearing force reduces the chances for cone breakup and doubling. The restoring force of the air displacement in compression and rarifaction is nearly ideally linear because air behaves as a perfect gas with respect to compressibility anywhere near atmospheric pressure and the percentage of compression compared to the total trapped air volume is relatively small. I think written on that basis he would have had a patent that would have stood up against challenges. Short of adding servo correction as well, the concept is still unbeatable because it inherently exploits the laws of physics which govern it. This is in direct opposition to ported systems where the port inherently makes the restoring force of the compressed air a function of frequency since the resistance to the flow of air in the port depends on what frequency it is tuned to. The linear 12db per octave falloff below resonance makes the system response practical to extend in a way not feasible with ported systems which have a 24 db per octave falloff.

[Zilch]

For a while the front wall was partially draped. No drapes extended behind the speakers or outboard of them. The side walls are lined with filled up bookcases. In no way does this resemble an LEDE arrangement or the Toole-favored NRC configuration.

I'll say it again: the room in the diagram is NOT LEDE; only the front wall and the sides at the speaker carousels are damped. The side walls from there to the listener(s) are highly reflective, more so than Howard's, even, and thus would generate a generous imaginary reverberant field when very-wide dispersion speakers were deployed therein....

[speaker dave]

David,

In the nearly sixty years since Villchur (Villchur -- two els) introduced the acoustic-suspension speaker, there have been dozens (perhaps hundreds) of critics and naysayers -- usually retired-speaker designers -- trying to discredit his accomplishment of the acoustic-suspension loudspeaker. You should feel honored that you are not the first to say that Olson and Preston anticipated the acoustic-suspension system! But as is usually the case, most of these criticisms are oversimplifications (a good example IMO is your statement above, "hence a high alpha or Acoustic Suspension system") and eventually fade away.

Olson and Preston describe a system with air stiffness 3 times the driver stiffness. Describing that as "high alpha" is exactly what it is. Interestingly Dickason quotes Weems and Small as defining Acoustic Suspension as a system with an alpha of 3 or greater. I prefer high alpha to acoustic suspension which is more of a marketing description than an engineering one.

The problem with Edgar Villchur's patent was that he wrote the patent himself, and he didn't use a patent attorney because he couldn't afford one. He therefore didn't do due diligence in performing an adequate patent search. Villchur's patent could easily have been written in such a way as to protect itself against "prior art," had there been adequate patent research. As it happened, Olson and Preston's patent (#2490466) was missed by Villchur and his patent examiner during his search.

I am not convinced that knowing about Olson and Preston's patent would have so easily allowed Villchur (sorry about the previous mispellings) to write a patent that would get around the prior art. To many of the essential claims are present: a system where the cabinet stifness dominates driver stiffnes and largely determines resonance. Replacing suspension nonlinearity with the more linear restoring force of the box and claiming a distortion reduction because of it (I agree that EV is off base on that count).

The essential proposal of the RCA patent was to improve speaker bass performance, in distortion and frequency response, by extending the region of mass control downward in frequency. Olson and Preston accomplished this by means of a compliant, folded-rim suspension that lowered the speaker's fundamental resonance frequency and allowed large, linear voice-coil travel. Olson and Preston plotted the reduced distortion for their speaker only for infinite-baffle mounting, and they showed optimum frequency response for the same mounting. They also pointed out that if their speaker were mounted in a small, enclosed cabinet, the resonance frequency could be raised to a point that was "still desirably low," and they showed a frequency response for such a mounting -- much inferior to that with the infinite baffle in low-frequency range and in correct value of Q, but superior to results with an open-back cabinet.

Where Villchur's air-suspension design relied on a cushion of air behind it, the Olson and Preston design was shown as being better off without the cushion of air, but capable of tolerating it for non-high-fidelity applications. There was therefore never a question of Villchur's speaker infringing on the RCA patent -- since the latter made no general claim for a low-resonance speaker in a small cabinet.

I'm not sure the courts would give credence to an argument about relative box size or correct value of Q. These are variables on a continuum. Isn't it difficult to patent a parameter that is significant over just a part of a possible range?

I agree that Villchur and AR changed the direction of speaker design. Olson and Preston may have had prior art but didn't have the vision to follow through. Along the lines of your quote from Hunt, I like the story of DeForest vs. Armstrong. DeForest as a dabbler added a third element to a light bulb and invented the Triode. He never understood its operation fully and labored under the false belief that the presence of gas was essential (it was detrimental). Armstrong came along and analyzed it thoroughly, improved its performance and developed the regenerative detector. They were in the courts for years. In the end Armstrong got the credit he deserved.

David

[speaker dave]

I'll say it again: the room in the diagram is NOT LEDE; only the front wall and the sides at the speaker carousels are damped. The side walls from there to the listener(s) are highly reflective, more so than Howard's, even, and thus would generate a generous imaginary reverberant field when very-wide dispersion speakers were deployed therein....

Okay, LELSDE (live end, live sides and dead end).

[soundminded]

Olson and Preston describe a system with air stiffness 3 times the driver stiffness. Describing that as "high alpha" is exactly what it is. Interestingly Dickason quotes Weems and Small as defining Acoustic Suspension as a system with an alpha of 3 or greater. I prefer high alpha to acoustic suspension which is more of a marketing description than an engineering one.

I am not convinced that knowing about Olson and Preston's patent would have so easily allowed Villchur (sorry about the previous mispellings) to write a patent that would get around the prior art. To many of the essential claims are present: a system where the cabinet stifness dominates driver stiffnes and largely determines resonance. Replacing suspension nonlinearity with the more linear restoring force of the box and claiming a distortion reduction because of it (I agree that EV is off base on that count).

I'm not sure the courts would give credence to an argument about relative box size or correct value of Q. These are variables on a continuum. Isn't it difficult to patent a parameter that is significant over just a part of a possible range?

I agree that Villchur and AR changed the direction of speaker design. Olson and Preston may have had prior art but didn't have the vision to follow through. Along the lines of your quote from Hunt, I like the story of DeForest vs. Armstrong. DeForest as a dabbler added a third element to a light bulb and invented the Triode. He never understood its operation fully and labored under the false belief that the presence of gas was essential (it was detrimental). Armstrong came along and analyzed it thoroughly, improved its performance and developed the regenerative detector. They were in the courts for years. In the end Armstrong got the credit he deserved.

David

The point of Villchur's invention was not the box size as acoustic suspension systems can be built to any size. The point was to use the parameters of moving mass, springiness of air compression and rarifaction in a sealed enclosure, and the viscousity of aerodynamic drag alone as the three coefficients for tuning a loudspeaker driver/enclosure combination to optimize the parameters of Newton's second law of motion to achieve any arbitrary mechanical fr/resonance response rather than relying on any mechanical restoring force which is inevitably non linear. In fact the patent application should have described the invention in its purest form with infinite compliance and no mechanical restoring force whatsoever and then in one of the claims or a folloup patent added a small amount of mechanical restoring force and air leakage to compensate for and overcome the manometric effect of varying outside barometric pressure. This would have made it entirely different from Olsen's patent which relied on a combination of mechanical and air spring, made no mention of viscous damping as far as I know (I haven't read the patent) and would have established the idealized version as entirely novel.

A vacuum tube diode is not a light bulb even if it has a directly heated cathode that is a glowing filament.

Armstrong's major contribution that was not recognized and then stolen which led to his suicide was the invention of FM radio, the method of signal frequency modulation of an rf carrier and detection instead of amplitude modulation as had been the prior art. At the time he invented it, so called experts said it could not work....just the way they said the acoustic suspension loudspeaker would not work. Industries set themselves up for failure and extinction when the hubris of their success leads them to reject innovative technologies that surpass the one they are familiar and comfortable with. They often overwork their advertising hyperbole to the point where every little variant or minor tweak becomes a breakthrough so that when a real one comes along, it is dismissed until it fights to prove itself. Then when they can no longer ignore it, they often steal it. The longer the empty hyperbole goes on without genuine advance, the harder that industry hits its head against the brick wall at the dead end of its technology trying to break through. That is where the entire consumer audio industry is now IMO.

[Steve F]

OK, I must confess to being a sucker for rhetorical questions and always think that the poster of such material is being serious.

Anyway, Re: the Neil Grover LvR demos. Yep, the 3 and 3a could not have played as loud as the 10 Pi. Hey, thank you ferrofluid! "Better" for you in this particular discussion seems to primarily involve the ability to play louder. Jeeze, if that is the case the Klipschorn beats the pants off of everybody else.

However, if you subtract the max-output issue, subtract the midrange dispersion issue (between 500 and 1000 Hz), deal with the mid-bass suckout (Allison boundary effect that impacts normal listening rooms) issue, and the treble dispersion issue (the .75-inch AR dome tweeter still beats most contemporary 1-inch domes and of course the AR-3 dome), super-low bass reproduction (thank you Velodyne, Hsu, and SVS), I think we can see that really not all that much has advanced since the Villchur LvR era. Sure, some of the later designs were improvements, but if we use the LvR demos as a reference standard what are we to make of those later designs? Do you suppose they could have delivered results that were even more "realistic" sounding than the live performers?

Ferrofluid allowed the Model 11 to deliver flatter respose (not smoother, just flatter, with less of a high-range tilt downward) than the AR-3a. As best I can tell, that was about it, and also I am pretty sure the AR-3a (at least the earlier versions with the more loosely suspended woofer) could reach deeper into the low bass range than the 11.

Howard Ferstler

The 10 Pi was used in the Grover L v R demos, in the 2 Pi position, so that means the 11 could have been used as well.

The "11 Pi" (HF--that's shorthand for '11/10 pi'--don't take it literally) was so superior to the 3a it wasn't even close. Merely 'adding ferrofluid' to the tweeter for flatter response is not as insignificant as you imply. The 11 pi's flatter response corrected the main shortcoming of the 3a and transformed that speaker's sound completely. I submit that a correctly-functioning 11 is still a satisfying speaker even by today's standards (Dave--I'm not saying 'better than today's best,' just 'satisfying' by today's standards), while a 3/3a is just an "enjoyable" speaker today, but its shortcomings/colorations are plainly evident.

Remember, the "11 Pi" also had flush-mounted drivers, a non-acoustically-intrusive foam grille and therefore none of the 'near-field havoc' of the 3/3a. That in itself is a significant improvement.

And I would therefore submit that since the "11 Pi's" cabinet molding didn't introduce those disastrous diffraction effects like the 3a, the 11 Pi's 3/4" dome tweeter likely had better real-world dispersion than the 3a--in spite of the 3a's hard-dome tweeter being somewhat better when measured individually on a test baffle. In any event, the 11 Pi's flush-mounted 3/4" soft dome has better upper-octave dispersion (above 12-13 kHz) than today's typical horn-recessed 1" dome.

There is no similarity between the tonal character of the 11 Pi and a Klipsch. The 11 Pi sounds like a cleaned up 3a with the same great bass, the same (like it or hate it) widely-dispersed midrange, and a crisp no-treble-boost-needed high end. There is no significant difference between the 11 Pi's bass and the 3/3a bass. Perhaps the very first Alnico 12" woofers had a free-air resonance a cycle or three lower than the later ones, but not significant. The 11 Pi, latest 3a, and LST woofers were all essentially identical. If you or anyone else could hear the difference between the 3's bass extension and the 11 Pi's bass extension, I'll send you to Disneyland.

BTW, the LST was tried by AR in the early going of the Grover L v R demo. The LST couldn't do it--not because of any output or power-handling issues, but because its multiple MR and HF drivers betrayed its origin every time. Whether it was driver interference or a too-different radiation pattern from the live source or something else altogether, isn't known.

But they did try the LST and it was no go.

Steve F.

[Pete B]

Olson and Preston describe a system with air stiffness 3 times the driver stiffness. Describing that as "high alpha" is exactly what it is. Interestingly Dickason quotes Weems and Small as defining Acoustic Suspension as a system with an alpha of 3 or greater. I prefer high alpha to acoustic suspension which is more of a marketing description than an engineering one.

I am not convinced that knowing about Olson and Preston's patent would have so easily allowed Villchur (sorry about the previous mispellings) to write a patent that would get around the prior art. To many of the essential claims are present: a system where the cabinet stifness dominates driver stiffnes and largely determines resonance. Replacing suspension nonlinearity with the more linear restoring force of the box and claiming a distortion reduction because of it (I agree that EV is off base on that count).

I'm not sure the courts would give credence to an argument about relative box size or correct value of Q. These are variables on a continuum. Isn't it difficult to patent a parameter that is significant over just a part of a possible range?

I agree that Villchur and AR changed the direction of speaker design. Olson and Preston may have had prior art but didn't have the vision to follow through. Along the lines of your quote from Hunt, I like the story of DeForest vs. Armstrong. DeForest as a dabbler added a third element to a light bulb and invented the Triode. He never understood its operation fully and labored under the false belief that the presence of gas was essential (it was detrimental). Armstrong came along and analyzed it thoroughly, improved its performance and developed the regenerative detector. They were in the courts for years. In the end Armstrong got the credit he deserved.

David

You have been completely on the money and unbiased here with your comments

David, I agree with all of your points.

Professor Wadsworth at WPI pointed out to me that large VC overhang was a key

feature of Villchur's design and it seems odd that more emphasis was not placed

on this in the patent. Wadsworth felt that this was more important than the spring

linearity as far as Villchur's design was concerned.

[speaker dave]

The point of Villchur's invention was not the box size as acoustic suspension systems can be built to any size. The point was to use the parameters of moving mass, springiness of air compression and rarifaction in a sealed enclosure, and the viscousity of aerodynamic drag alone as the three coefficients for tuning a loudspeaker driver/enclosure combination ...

Size was certainly an issue as otherwise the system would be indistiguishable from an infinite baffle of which prior art obviously exists. The essence of acoustic suspenion is the relative stiffnnes of box vs. woofer. In most woofer box papers this ratio is refered to simply as Alpha. This can also be looked at as Vb vs. Vas where the first is box volume and the second is the compliance of the woofer expressed as an equivalent volume. The claim for acoustic suspension is that Vas is large relative to Vb, hence the box stiffness dominates. If the box stiffness dominates then you can claim, as Olson and Vilchur did, that it, as an air spring, is more linear and distortion is reduced.

A few comments on that: first, although the marketing notion is that the woofer has been made very compliant and the box dominates, this alone would not get us from a typical 1950s infinite baffle system to an AR-1. Typically a 12" woofer in a 4 to 6 cubic ft box would have more magnet and a lighter cone than would suit a smaller cabinet, even if all stiffness were removed from the woofer. For the AR-1 to work considerable mass was required and an adjustment to the magnet strength. This is not a criticism of the AR-1 but a clarification that high compliance alone was not all that was entailed. To Villchur's credit he actually considered all the major parameters in concert to create a woofer box system that had a flat in 2 pi response. This is a mindset change over previous design that assumed that a light cone and the biggest magnet possible were the ideals to pursue.

Secondly, suspension linearity is not necessarily the evil problem that we think it is. In my experience motor linearity (simplified to voice coil overhang) is the major determinant of LF max output or distortion. It is common for advanced designers to balance a certain nonlinearity of spider against magnetic drive falloff. That is, a less linear spider will frequently give a more linear overall effect. This has a lot to do with DC offset or jump phenomonon. Refer to an early paper by T.H. Wiik of SEAS if you want to read more on that, it is a fascinating effect and cure.

Finally, viscous damping is not the primary mechanism for achieving the system Q that you want. A sealed system will have the combined effects of viscous drag and electrodynamic drag from the voice coil and magnet. Typically these are refered to as Qm and Qe which combine (inverse of the sum of the inverses) to give Qt. The preference is that Qe supply the greater drag because that gives the best "efficiency constant" (compromise between efficiency, bass cuttoff and box size). A Qm of about 7 is usually assumed which would leave us to need a Qe of about .78 to achieve a Butterworth allignment.

Pete B. are you out there?

David

[speaker dave]

Pete B. are you out there?

David

Oh, you replied while I was typing!

Thanks

[speaker dave]

The "11 Pi" (HF--that's shorthand for '11/10 pi'--don't take it literally) was so superior to the 3a it wasn't even close. Merely 'adding ferrofluid' to the tweeter for flatter response is not as insignificant as you imply. The 11 pi's flatter response corrected the main shortcoming of the 3a and transformed that speaker's sound completely. I submit that a correctly-functioning 11 is still a satisfying speaker even by today's standards (Dave--I'm not saying 'better than today's best,' just 'satisfying' by today's standards), while a 3/3a is just an "enjoyable" speaker today, but its shortcomings/colorations are plainly evident.

Steve F.

I understand your distinctions. Interesting about the LST,

David

[soundminded]

Size was certainly an issue as otherwise the system would be indistiguishable from an infinite baffle of which prior art obviously exists. The essence of acoustic suspenion is the relative stiffnnes of box vs. woofer. In most woofer box papers this ratio is refered to simply as Alpha. This can also be looked at as Vb vs. Vas where the first is box volume and the second is the compliance of the woofer expressed as an equivalent volume. The claim for acoustic suspension is that Vas is large relative to Vb, hence the box stiffness dominates. If the box stiffness dominates then you can claim, as Olson and Vilchur did, that it, as an air spring, is more linear and distortion is reduced.

A few comments on that: first, although the marketing notion is that the woofer has been made very compliant and the box dominates, this alone would not get us from a typical 1950s infinite baffle system to an AR-1. Typically a 12" woofer in a 4 to 6 cubic ft box would have more magnet and a lighter cone than would suit a smaller cabinet, even if all stiffness were removed from the woofer. For the AR-1 to work considerable mass was required and an adjustment to the magnet strength. This is not a criticism of the AR-1 but a clarification that high compliance alone was not all that was entailed. To Villchur's credit he actually considered all the major parameters in concert to create a woofer box system that had a flat in 2 pi response. This is a mindset change over previous design that assumed that a light cone and the biggest magnet possible were the ideals to pursue.

Secondly, suspension linearity is not necessarily the evil problem that we think it is. In my experience motor linearity (simplified to voice coil overhang) is the major determinant of LF max output or distortion. It is common for advanced designers to balance a certain nonlinearity of spider against magnetic drive falloff. That is, a less linear spider will frequently give a more linear overall effect. This has a lot to do with DC offset or jump phenomonon. Refer to an early paper by T.H. Wiik of SEAS if you want to read more on that, it is a fascinating effect and cure.

Finally, viscous damping is not the primary mechanism for achieving the system Q that you want. A sealed system will have the combined effects of viscous drag and electrodynamic drag from the voice coil and magnet. Typically these are refered to as Qm and Qe which combine (inverse of the sum of the inverses) to give Qt. The preference is that Qe supply the greater drag because that gives the best "efficiency constant" (compromise between efficiency, bass cuttoff and box size). A Qm of about 7 is usually assumed which would leave us to need a Qe of about .78 to achieve a Butterworth allignment.

Pete B. are you out there?

David

"And I would therefore submit that since the "11 Pi's" cabinet molding didn't introduce those disastrous diffraction effects like the 3a, the 11 Pi's 3/4" dome tweeter likely had better real-world dispersion than the 3a--in spite of the 3a's hard-dome tweeter being somewhat better when measured individually on a test baffle. In any event, the 11 Pi's flush-mounted 3/4" soft dome has better upper-octave dispersion (above 12-13 kHz) than today's typical horn-recessed 1" dome."

This business about difraction is rediculous. What do you do about the diffraction of high frequencies created by music stands and adjacent musical instruments when people are playing actual musical instruments especially in an ensemble? One audiophile whose house I visited didn't want to turn up his subwoofer level because he didn't want to "excite room resonances." It made me wonder what he'd do if his kid wanted to play a real musical instrument and they brought a piano into his listening room. What would he do about room resonances then? Audiophiles are a curious bunch even if they make a living out of designing equipment. Reflections off the the outer molding on an AR3a can easily be eliminated by taping some sound absorbing matrial to it like the felt carpet underlayment they used to use that is very similar to the AR9's acoustic blanket. Not that it would actually matter much.

A large box is not the same as an infinite baffle and will not behave like one if the mass of the cone is light enough. If it were, there would be no reason to have built large horn or bass reflex systems, they could all have been infinite baffles and drivers could have been optimized for them. Infinite baffles can work quite well. The only driver on the market I'm aware of designed specifically for infinite baffles is one made by Dayton.

The predominant mode of non linear woofer distortion is cone breakup due to unequal stress. The breakup modes are exactly the same as a bessel function which describes the way a drum head will vibrate. Second and third harmonics are the most characteristic. As I said previously, this is largely due to the lack of strength of the cone material in shear and so torsional force due to unequal mechanical restoring force from the spider and outer suspension create a force gradient that is both transverse and circumferential to the cone. This flexing is eliminated in an AS design because the force is applied evenly over the surface of the cone and there is no force gradient. Also the stress per unit area on the cone is much lower, the force not being concentrated an any points on its surface especiallly at its perimeter where the torsional gradient force would be at its maximum compared to the center.

I have thought long about KK's passive model and still disagree with it. The stored potential energy in the inertial mass of the cone when it is displaced will tend to reconvert to kinetic energy at the system resonant frequency, not the driving frequency. Therefore the underdamped moving mass will generate non harmonically related motion which by default is intermodulation distortion, not harmonic distortion (all non harmonically related non linear distortion falls into the category of IM and noise.)

I view the electrical resonance problem and mechanical resonance problem as separate and superimposed. Of the two, the mechanical problem is much harder to solve, the electrical problems being completely and easily dealt with by any of several means, the use of an active equalizer being the cheapest, most flexible, and easiest to use. It is even more effective when used with active crossover networks. If time delay of propagation actually means anything (I'm not convinced it does) that can be dealt with most effecively at the active signal level using digital time delays as well. Active filtering can also be used to resolve those aspects of the mechanical resonance problem that remain after the best mechanical solution is applied and far more effectively than by use of passive crossover filters at the high power level.

I'm still waiting to hear how a vacuum tube diode is the same as a lightbulb.

[Steve F]

I understand your distinctions. Interesting about the LST.

David

Victor Campos (who led the Grover L v R project at AR) told me about the LST not cutting it at a BAS meeting some years ago. He also said that they used the 10 Pi in its 2 Pi position--even though that made it essentially an 11--because Marketing wanted the 'credit' for the demo to accrue to the top-of-the-line model. Perfectly understandable.

Steve F.

[genek]

Victor Campos (who led the Grover L v R project at AR) told me about the LST not cutting it at a BAS meeting some years ago. He also said that they used the 10 Pi in its 2 Pi position--even though that made it essentially an 11--because Marketing wanted the 'credit' for the demo to accrue to the top-of-the-line model. Perfectly understandable.

Wasn't the whole idea of the 10Pi's crossover adjustments to be able to make it behave as if it was in any pi mode you wanted it to be in, regardless of what mode it was actually placed? Or maybe I missed the concept altogether.

[genek]

I am not convinced that knowing about Olson and Preston's patent would have so easily allowed Villchur (sorry about the previous mispellings) to write a patent that would get around the prior art. To many of the essential claims are present: a system where the cabinet stifness dominates driver stiffnes and largely determines resonance.

It's been years since my "Patents 101" training at my first engineering job, but we were taught that the idea of finding prior art is not to "get around it," but to cite it and then describe what you are doing that is either different or adds to it. So if any of the essential claims of EV's patent were contained in Olson and Preston's, EV was claiming the wrong things. What he needed to claim was whatever his "aha!" lightbulb idea was that resulted in the AR-1 that Olson and Preston had failed to recognize and apply in whatever product designs resulted from their patent.

In his 2005 Stereophile interview, EV commented on the fact that the official who handed down the ruling was not technically adept, but the reality is that most patent disputes are ultimately decided by such people. Citing prior art and describing how you are improving it in layperson terms creates a warm, fuzzy feeling, while omitting it creates the negative impression that you're trying to pull the wool over someone's eyes.

[soundminded]

Wasn't the whole idea of the 10Pi's crossover adjustments to be able to make it behave as if it was in any pi mode you wanted it to be in, regardless of what mode it was actually placed? Or maybe I missed the concept altogether.

That is correct. It basically equalized the lowest bass to adjust to how much acoustic reinforcement from adjacent walls it received. Evidently in the setting it was used it, that level was identical to AR11's fixed FR.

The shortcoming with LST that IMO made it unable to perform sucessfully in the LvR demos was lack of midrange drivers on the front baffle. Had the midranges been arranged similarly to the tweeters, I think it would have done even better than 10 pi or 11 because its geometric propagational pattern, its directivity or lack of it would have been more similar to most musical instruments than for the 10 pi or 11.

I noticed in the AR3 demo with the guitarist that the speaker actually sounded slightly brighter than the guitar which surprised me considering my prior (and subsequent) experience with AR3 elsewhere. I was just about directly on axis having had an aisle seat. I think the more off axis I would have been, the more the guitar would have sounded similar and then brighter than the speaker. That's what I would expect. I wasn't aware until I read it hear that the treble had been boosted to make the speaker sound flat. Nothing wrong in that but it was never the way I'd seen it used in other places.

[speaker dave]

This business about difraction is rediculous. What do you do about the diffraction of high frequencies created by music stands and adjacent musical instruments when people are playing actual musical instruments especially in an ensemble?

I hadn't commented on diffraction, but since you ask...rather than wild conjecture, let me suggest an experiment. Have a friend hold a ruler with its long edge on the surface of you favorite system (single tweeter, in this case). The ruler is perpendicular to the surface and will create a reflective barrier. Play pink noise and have that person slide the ruler a little closer and a little farther from the tweeter. I guarantee you will hear a comb filtering effect that will go up and down in pitch with as the distance to the tweeter is varied. Now the remaining question is what reflection levels and what time delays are audible? Does the direction of arrival of a reflection have and effect? Let me recommend Toole's book, which has over a chapter on the subject, with his and others findings.

A large box is not the same as an infinite baffle and will not behave like one if the mass of the cone is light enough. If it were, there would be no reason to have built large horn or bass reflex systems, they could all have been infinite baffles and drivers could have been optimized for them. Infinite baffles can work quite well. The only driver on the market I'm aware of designed specifically for infinite baffles is one made by Dayton.

Yes, most baffles aren't truly infinite as that tends to take up a lot of space. A large enclosure is typically called an infinite baffle in our industry. I suppose you could qualify that as a low alpha system or one that had a small effect on the free air resonance of the driver.

The predominant mode of non linear woofer distortion is cone breakup due to unequal stress. The breakup modes are exactly the same as a bessel function which describes the way a drum head will vibrate. Second and third harmonics are the most characteristic. As I said previously, this is largely due to the lack of strength of the cone material in shear and so torsional force due to unequal mechanical restoring force from the spider and outer suspension create a force gradient that is both transverse and circumferential to the cone.

You really are confusing breakup modes with nonlinear distortion elements. The primary elements of driver distortion are coil length vs. gap length (Bl vs. x), spider and surround nonlinearity and magnetic hysteresis effects. With those alone you can do a pretty realistic model of woofer nonlinearity. Klippel is very good in this area, if you want to read something.

I really don't understand the rest of your argument, sorry.

Regarding the light bulb: a bit of hyperbole on my part. You are correct in that he added a third element to a diode, not that you couldn't read a book from the light of some bright emitter triodes.

David

[speaker dave]

It's been years since my "Patents 101" training at my first engineering job, but we were taught that the idea of finding prior art is not to "get around it," but to cite it and then describe what you are doing that is either different or adds to it. So if any of the essential claims of EV's patent were contained in Olson and Preston's, EV was claiming the wrong things. What he needed to claim was whatever his "aha!" lightbulb idea was that resulted in the AR-1 that Olson and Preston had failed to recognize and apply in whatever product designs resulted from their patent.

In his 2005 Stereophile interview, EV commented on the fact that the official who handed down the ruling was not technically adept, but the reality is that most patent disputes are ultimately decided by such people. Citing prior art and describing how you are improving it in layperson terms creates a warm, fuzzy feeling, while omitting it creates the negative impression that you're trying to pull the wool over someone's eyes.

You are probably right about the warm and fuzzy approach. I have only been through the process once and remember that the patent examiners are generalists and brought up possible "interference' patents that didn't make sense while omitting ones that might have been closer.

I'm sure that Villchur could have taken a different tact. The problem was that the claims he wanted to make: a compact enclosure and lower distortion because of it, were already claimed by Olson and Preston. What else would he claim that would allow him to protect the product he really wanted to make?

David

[genek]

I noticed in the AR3 demo with the guitarist that the speaker actually sounded slightly brighter than the guitar which surprised me considering my prior (and subsequent) experience with AR3 elsewhere. I was just about directly on axis having had an aisle seat. I think the more off axis I would have been, the more the guitar would have sounded similar and then brighter than the speaker. That's what I would expect. I wasn't aware until I read it hear that the treble had been boosted to make the speaker sound flat. Nothing wrong in that but it was never the way I'd seen it used in other places.

AR's instruction sheet for setting the level controls of the 3a and 5 cited the need for speakers to roll off in normal use because of excessive HFs in then-contemporary recordings, which they clearly considered a bad thing. The recordings for the LvRs were made by AR; we can probably presume that they would not have deliberately made a recording with the same bad characteristics, and if they made a recording they considered "accurate," they would have wanted to tune the speakers to be that way as well.

I occasionally run across a recording that seems to be "duller" than most (almost exclusively classical DDD CDs), and setting my speakers as AR recommended for "flat" (mid and high up full and treble at +3dB) usually sounds just about right to me for them.

[genek]

What else would he claim that would allow him to protect the product he really wanted to make?

What did AR do in the AR-1 that had not been done by the holders of the prior art patents? If there was nothing significant, then considering the kind of market splash AR made, you would think that Olson, Preston or their assignees would have challenged AR, or would at least have tried to step in and demand that AR's licensees and anyone else who wanted to use the technology start paying royalties to them instead once AR lost its protections based on the argument that their patent was the prior art. They would already have had one patent judgement backing their claim.

[Steve F]

The patent discussion strikes me as slightly beside-the-point. Whether or not the AR patent, as written by EV, was or was not defendable or accurate didn’t seem to affect the way AR came to market with its products, nor was it the cause of their long-term success or failure.

Regardless of Olsen or anyone else, it was Villchur and his then-diminutive 25”-long AR-1 that factually materialized on the scene in 1954 and turned the speaker market on its head overnight with low distortion 30 Hz bass. EV had the foresight to turn the device—whether technically a defendable, exclusive “invention” by AR or not—into a viable, game-changing product.

Further, EV didn’t appear to have the marketing “killer instinct” that later executives at other electronics/speakers companies had, so it is open to conjecture whether or not AR under EV would have been “more” successful had the original AS patent been written differently. I’m not going to try to get into people’s thoughts and motivations 56 years after the fact, but the speculation by others is fascinating, and it will continue to be interesting reading.

But the fact of the matter is that by 1966—a full 12 years after the company’s founding, and many years after the AS principle was in full use by arch-rival KLH—AR’s market share was an astounding 33%. Clearly, AR was a tremendously successful company, and the reasons for their subsequent demise 10-15 years hence has been well documented here, and had nothing to do with how the original patent was handled, one way or the other.

Steve F.

[Zilch]

Olson = 12/6/49:

http://www.freepatentsonline.com/2490466.pdf

AR invalidation = 12/28/62:

http://bulk.resource.org/courts.gov/c/F2/3...08.13822_1.html

*********

Rejection of LST for the later LvR demos is documented somewhere, 'cause I knew that.

[Give me two weeks and I'll find it.... ]

[genek]

The patent discussion strikes me as slightly beside-the-point. Whether or not the AR patent, as written by EV, was or was not defendable or accurate didn’t seem to affect the way AR came to market with its products, nor was it the cause of their long-term success or failure.

Yes and no, I think. It certainly didn't materially affect the impact that the introduction of AR's products had on the state of speaker design. But as EV originally wanted to sell his invention to someone rather than run his own company, the invalidation of the patent may have played a role in his continuing to manage and grow the business for another five years. Had the AR patent held up, it could potentially have enabled him to sell off the company and its IP and get out of the building without having to stay there to build the market share up as high as he did.

[speaker dave]

I doubt if anybody today would be able to audition a pristine AR-3a and compare them fairly to much of anything, since the drivers (particularly the tweeter) will have deteriorated to the extent that the system would not sound the way it did when first built. Those who hear current AR-3a systems are probably experiencing "updated" or "modified" versions that are the result of restoration projects, and so we have no idea of just how good or bad they might be, compared to today's best systems.

Howard Ferstler

There you go, the ultimate let-out clause. If you hear a pair of AR3as and they aren't the best system of all time, then they have clearly declined with age or were modified along the way.

Other reasons for disapointment:

you don't appreciate good music

you don't have a proper listening room

you sit too close

you like the wrong sort of imaging

you like that new-fangled bright sound

your hearing was ruined by rock and roll

No hope for me.

[genek]

Those who hear current AR-3a systems are probably experiencing "updated" or "modified" versions that are the result of restoration projects, and so we have no idea of just how good or bad they might be, compared to today's best systems.

No hope for me.

If you're ever in my neighborhood, look me up. The only change from 3a original is the grill fabric. Tweeters haven't been measured, but their sound seems to correspond to the listening impressions of reviews published when new. Can't make any comparisons to "todays best systems," whatever they may be because I haven't got any of those here, and I don't have any measuring instruments. You'll have to bring them with you.