Concert Halls and Speaker Cones
Name that Distortion
Toasting Tweeters - Part 2
-by Ed Poindexter
I would like to extend the discussion on speaker cones introduced at the September meeting, and I hope for some strong opinions by readers. I don't recall the exact context of remarks, but somehow the notion of live-vs- recorded comparisons came up, and I was admonished that paper is (still) a very fine material for direct-radiator speaker cones. The specific issue I had in mind was, however, how the best paper cones of more than 50 years ago would stack up against the best cones-paper, plastic, or composite-of today.
On 29 July 1947 in the Tanglewood Music Shed, RCA conducted a comparison of the live Boston Symphony Orchestra vs. music reproduced via the loudspeakers used in its own new Festival Series of radios and phonographs. The audience reaction was surprise and pleasure at the quality of reproduction. The demonstration was described in Harry Olson's book.1 Recently, I succeeded in getting a copy of Olson's internal RCA report, which includes some important additional info.2
Twelve speakers were distributed along the entire front edge of the stage, only a few feet from the first row of musicians. Perhaps surprisingly, the speakers were not the fabled LC1A,3 but rather, an older, simpler version lacking the 7 elliptical diffuser cones affixed to the main cone. They were of 15-inch diameter, with a 2-inch tweeter cone mounted in the center of the woofer cone, without any offset between woofer cone inner edge and tweeter cone outer edge. Both cones were paper; the woofer cone was 2.5 times as thick as a "normal" cone, making it 15X stiffer. The cones were both straight-sided; and though not described, the outer surrounds appear to be simply a molded extension of the cone paper. There is no discussion of cone paper in the report, or the referenced article4-perhaps to protect proprietary aspects.
It was found that the maximum acoustic power of the Orchestra could be reproduced with three 40-watt amplifiers. Presumably, this power was sufficient for the instantaneous wave-crest peaks-typically 10 to 11 db higher than the maximum observed with a standard "fast" sound level meter. In the Tanglewood Shed, which seats about 6000, internal volume 1.5M cu. ft., midrange TR 1.25 sec, the maximum observed SLM reading was 95 db at 85 ft from the stage. This figures out to an efficiency rating of over 100 db/one watt/one meter for the speakers. Stereo reproduction was adjudged to be unnecessary, due to the very reverberant stage shell and rather dead audience area.
The system included a crude threshold-cut diode circuit between two one-octave filters which reduced playback noise by 15 db, and somehow did not butcher the music.
The last 4 minutes of Beethoven's "Egmont" overture was played by the Orchestra, and recorded on disc-presumably 78 rpm in 1947. For the demo, the switch from live to recorded music was carefully rehearsed to allow continuous music without interruption or audible glitch.
This was not a blind test; the orchestra and the speakers were in full view. The demo got very nice comments from capacity audience members, well-known music critics, and the conductor-Serge Koussevitsky. Many persons, if eyes were closed, could not tell where live left off and reproduced began.
Several questions arise:
(1) Does the concert-hall acoustic ambience so dominate the quality of sound that other factors-primarily the electrical and acoustical inadequacies of the sound system-do not matter?
(2) Would the demo have been even more impressive with today's sound systems-better speakers, CD's instead of records-ie, able to withstand a blind test?
(3) In a home environment, how would the RCA 1947 sound system compare to an analogous one today? I note that the successor speaker, the LC1A, still has many fans on the web more than 50 years later.
(4) Would the cone paper in RCA's speakers be competitive today?
1. Harry F. Olson, Acoustical Engineering (Van Nostrand, Princeton, 1957), pp 606-611.
2. Harry F. Olson, RCA Engineering Memorandum PEM-79, 18 Aug. 1947. Kindly made available by the David Sarnoff Library, Princeton, NJ.
3. H. F. Olson and J. Preston, RCA Review, Vol. VII, No. 2, p. 155 (Jun. 1946).
4. Harry F. Olson, John Preston, and Everett May, J. Audio Eng. Soc., Vol.2, p. 219 (Oct. 1954).
-by David Carlstrom
Some time ago the SMWTMS LC Silly Survey reported on the number of speakers that had expired while in the care of the members. Far earlier I reported in LC criticizing the oft told tale that small amplifiers destroy tweeters and powerful amplifiers protect tweeters. A more detailed view at the life and death of tweeters at the output terminals of amplifiers came up at the October 12, 2002 PSCAS meeting at amplifier maker Crown in Elkhart, Indiana. Gerald Stanley, Vice President of Crown and designer of all the Crown amplifiers explained his view. He said that early transistor amplifiers made with output devices of limited safe operating area and active protection circuits would allow tweeters to expire. But the amplifier did not supply the power that heated the tweeter! When active, the protection would turn off the output devices in the amplifier, momentarily letting the amplifier output terminals become an open circuit. This would happen when the woofer's current and voltage were out of phase. Just before the protection switched the amplifier off, there would be a high inductive current at low voltage. When the amplifier outputs went off, the current in the woofer refused to stop instantly, so the current went to the only remaining path, the tweeter. So it wasn't the amplifier that toasted the tweeter, it was the woofer! Higher power amplifiers usually activated their protection circuits less often and thus appeared to save tweeters. This further detail confirms that the notion that a high power amplifier protects a tweeter is a myth.
by David Clark
When I was a teenager I read Audio and other magazines from cover to cover. I especially liked a battle that took place over several months between Edgar Vilcher (founder of Acoustic Research) and Paul Klipsch (Klipschorn). They took opposite approaches to reproducing clean bass in the living room.
Klipsch maintained that efficiency was the key and horn loading produced the maximum efficiency. Horns seemed to not be applicable to home speakers because the size of a bass horn was too big. A hornís minimum dimensions are dictated by the wavelength of the lowest frequency that they intend to reproduce efficiently. A 40 Hz horn would need to have a mouth seven feet square and be about 14 feet long. Klipsch made bass horns practical by folding the air path and having it open into a corner of the room. The walls served to continue the flair and easily met the size requirements for 40 Hz.
Vilcherís approach was to clean up the distortion of speakers used in direct radiators. His AR-1 used a 12" woofer with (for the time) a heavy cone, long voice coil and loose suspension. The woofer went into a 1.5 cubic foot sealed box. The very linear air spring supplied by the box raised the resonance to 43 Hz for a -3 dB point below 40 Hz. The problem was that the system was inefficient. Fortunately, cheap, powerful amplifiers like the Dynaco Mk II came along at the same time (a build it yourself 50 watt "monoblock" for $69.95.)
I was fascinated by both approaches, but I could afford neither. A friend who made more money with his paper route than I made by mowing lawns let me talk him into building a folded horn. I designed it to be built into the house in the triangular area under the basement stairway. It opened into a corner of the basement "rec. room." I selected two 15" woofers of new design from a public address speaker company named Racon. The woofers had massive magnets, light cones with stiffening ribs, a green foam surround and a fair amount of stroke. Cool, or as we used to say, "neat."
The exponential horn design used well-known guidelines: throat area of Ĺ the diaphragm area; Cross-section doubled in area every 36" and a sealed back chamber of about 5 cubic feet. This should be good for a 32 Hz cutoff, according to the books. We wired the two 16 ohm woofers in parallel and hooked it up to my friendís 70 watt Eico amplifier.
We were dumbfounded at the results. The whole house seemed to shake and rattle. I was hooked forever and so was my friend. He still has the Racons, now in big vented boxes.
My next affair was with the Electro-Voice 30W. Soon after I had a house with enough room, I planned for one. Basically, one end of the room was to have a 32" high counter on which the amplifiers and upper range speakers were placed. The counter was, in fact, a 40 cubic foot sealed enclosure for the 30" diameter woofer. The woofer faced the wall behind with the cabinet pulled out about 3" for semi-horn loading. I drove it with a 40 watt amplifier and had all the deep bass necessary to keep up with my Acoustic Research AR-5 satellites.
When I moved, the big enclosure would no longer fit, but I had other plans anyway. I sawed off the ends, leaving a compact 48" wide, 32" tall 16" deep box turned around with the 30 incher facing forward. This gave a system resonance of 50 Hz, far too high. My plan, however, was a servo system to flatten the response and reduce distortion. I machined a magnesium plate to adapt a Bruel and Kjaer accelerometer to the voicecoil former. A "Tigersaurus" 200-watt amplifier was selected to drive it. I designed a charge amp for the accelerometer and selectable high-pass filters so the system would not go nuts on warps in the black vinyl playback medium. I measured gain and phase and designed a third-order compensator to close the loop of the feedback system.
Technically, it was an incredible success, but less of a success on music. At low levels, frequency response was flat from 6 Hz to 130 Hz. Distortion at maximum power was no more than two or three percent from 20 Hz to 100 Hz.
The problem on music was that it didnít seem to play very loud. As you turned it up it would gobble huge amounts of power and as soon as the Ďsaurus clipped, huge ugly popping and humming sounds were produced. This was 1976. Lasting lessons were learned: if you have really clean bass, you had better have a lot, and if you donít have enough, you must go into distortion gradually.
By this time, I was designing custom systems for recording studios and concert sound. For concert sound, I specialized in fully horn-loaded systems because of their high efficiency. Amplifier power was costly at $3.00 per watt, speakers had low power handling and plywood was cheap. I designed wedge-shaped mid and high frequency modules that hooked together and were raised above the stage in an array. The modules were designed to "truck pack" dimensions for carrying to the next gig.
My biggest client wanted new bass horns to replace the huge Western-Electric "tub" designs from the Ď30s. These held two 18s and sounded great but were too big and didnít "truck pack." They didnít have to twist my arm very hard to get me to say, "Iíll do it."
I decided on a horn module size half the width of the truck box and full height. This was 90" high and 45" wide. Depth was also 45." Each module contained four 18" Electro-Voice EVM-18B speakers. Flare rate cutoff was 28 Hz. The module was built with rigging hardware, chain sacks for hoisting it, 6" casters and fork lift pockets for moving it around. Modules were made in mirror-image pairs so that a single, symmetrical mouth almost 8 feet square was formed. The idea was that as many pairs as necessary were used at a concert. The module was named "Bigfoot." Two or more were called "Bigfoots."
At a Bob Seger concert in Cobo Arena that I attended, a solid block of 4 pairs of Bigfoots was flown about 30 feet above the stage. The array totaled 32, 18" horn-loaded woofers with a combined output of about 1200 acoustic watts. Best of all, the flare rate, total mouth size (16í X 16í) and path length are good for true horn loading down to 28 Hz, independent of room boundaries. The keyboardist performed a special deep bass synthesizer solo because the band knew I was in the audience.
In the 20 years since that time, I have designed many home woofer systems, many around the JBL 2245 18" and the Audio Concepts AC-12. Mostly, however, I have designed woofers for cars and used Velodyne servo woofers for myself. As I contemplated a new woofer for home, I realized that the tradeoffs have changed again. David Griesinger talks of "lateral bass" and "envelopment." Amplifiers get down to $0.40 per watt. Speakers that will stroke an inch peak to peak are common.
My listening room is built on a slab, so I cannot use the basement as an infinite baffle. I mainly listen to home theater DVDs, but I want to experiment with Griesingerís multi-woofer concepts. I wanted small subwoofers that would serve as lamp tables. Most of all, I wanted plenty of clean bass down to 16 Hz.
My choice of a woofer is from the aftermarket car audio world, the JL Audio 18W6. 18" diameter, strokes +/-16 mm, cone mass is 1.1 pounds, rated at 500 watts continuous. I need four to keep up with the Klaczas and other SMWTMS extremists. By putting two on opposite faces of a box, the mass reactions cancel so the lamp wonít fall off. The box would be as small as possible, limited only by the power available to compress and expand the air.
A Crown CE-4000 puts out 1800 watts per channel into 2 ohms and can be obtained for less than $1400.00. Parallel connection of two 18W6s results in about 2 ohms, so lets see how small we can make the box and still compress the air enough to stroke the full +/-16 mm. The answer turns out to be about 3.6 cubic feet for each two-woofer box. This translates to outer dimensions of a pair of 21-inch cubes. Perfect for the lamps.
You might ask if this design approach is leaving something out: frequency response, for instance. The system wonít be flat. It will roll off slowly below about 50 Hz. The 20 dB of eq needed to bring 16 Hz up to flat is a simple design that can be made from an old preamp. For the time being, mine is an external equalizer ahead of the power amp.
What about distortion caused by the small box? The air spring itself is very linear, but not perfect. At peak excursion, the air is only being compressed or expanded by about 3.6%. The non-linearity of compression is only a small percent of the 3.6 percent. The suspension of the 18W6 does not contribute to non-linearity over this range,
The only significant non-linearity is the Bl fall off away from rest position. This is 70.7% of the rest position value at +/-16 mm excursion. From past experience, this results in about 10% THD at peak excursion. This is quite low for a woofer, and it builds slowly from very low values at smaller excursions. Still, harmonics are "amplified" by the rising response. Having just installed the system, I can testify that my houseís windows rattle so as to mask any distortion. And, my windows donít rattle much.
That was the thinking that led to the dual boxter sub system. Now it is a reality thanks to the work of Justin Putty who built the box and Eric Busch who built the electronics interface in the DLC shop.
The first test was a single box driven by sine waves. Not enough power was available from the 120 volt wall outlet to drive both boxes to full output. Over an inch of peak to peak excursion is available at any frequency below 40 Hz. At maximum output, there is no extraneous noise audible. It is quite possible that listeners were so overcome by the pressure sensation that distortion was not audible. At lesser levels, some "crackling" noise coming from the plastic sack holding the damping material is audible through the cone. This was not considered to be worth fixing.
On music, there is never a condition in a normal listening room that requires approaching full output of the subs. Using bass effects discs and turning up the subs in the small conference room, it was possible to reach significant excursions. The effect is numbing. I was aware that the ceiling tiles danced in the grid
Finally finished and installed at home, the little boxes are dwarfed by the AR-91 towers on either side of them and the rear projection TV in the middle. The Crown amp joins two Haffler XL-600s and a DH-220 in a six foot rack in the basement. The system is pure 5.1 with full-range five channels and LFE going to the two subs via the interface box. The interface box has a switchable all-pass filter that rotates phase of one of the subs from in-phase at 16 Hz to 180 out at about 150 Hz to give the Griesinger "envelopment" (hopefully).
All the bass you ever wanted makes watching movies interesting. I first noticed that the bass is much faster than the French doors. They continue rattling long after the depth charge has exploded. Next, both front drapery valences slipped off and crashed to the floor. (I didnít think it would happen twice). The dog and cats stand paralyzed. They are not afraid, they just donít know what to do. A lot of bass causes your hearing to compress and the rest of the movie sounds softer. (I explained to Mary that this is why you have to turn it up.) Last Friday the video started losing sync with bass hits and finally went blank--just a wiggled out connector. Mary thought a helicopter was near as she drove up to the house one night.
All this and the enclosures are absolutely free of vibration. I guess its okay, then, that they are on wheels rather than spikes. (Can we get a golden-ear read on that?)
I donít have any plans for a new woofer at this time. This system satisfies my need for bass and for perverse creativity. (Car woofers, PA amplifier running on 220 volts, tiny pink boxes and intentional phase shift. To top it off, I bought new mid-line Radio Shack cables to hook it up.) Iíll be happy to show any one else how to build a pair.
by David Clark
Mark Ziemba presented a version of his paper on POSition Dependent Amplitude Modulation Distortion at the November meeting. I think this investigation into speaker distortion is great, but I donít think Markís acronym of ďPOSDAMĒ is so great. There is enough confusion in the audio business intentionally introduced by marketing spin artists. Serious researchers should use accurate acronyms.
Amplitude modulation (AM) distortion, or intermodulation (IM) are essentially the same thing and have been observed in speakers for years. The basic cause is well known: a low frequency causes the cone to stroke over a significant range and any simultaneous treble sound is affected by the different sensitivity of the speaker at different points in its stroke range. Generally, the motor strength (and therefore sensitivity) falls off as the cone moves away from center. Yes, this causes some harmonic distortion of the low frequency itself, but the warbling effect on simultaneous highs can be more annoying. Markís work has been on a different cause of this distortion: a change in the frequency response at different points in the stroke range. This is more than a sensitivity reduction as the cone leaves the rest position, this is a change in the shape of the response curve. The modulation distortion effect is the same. As the cone strokes, a high frequency tone is reproduced with cyclic amplitude variations. The interesting thing is that this change is not predicted by the motor strength falloff away from rest position.
How can this happen? Iíd rather not get into it. For now, please believe that it does. Mark covered some of it and gave some examples.
About the acronym. First either conventional IM or the new cause depends on excursion from one position to another, not on a specific cone position. This makes it EXcursion Dependent Amplitude Modulation Distortion. I think I like EXDAM even better than POSDAM.
EXDAM would include the old and known cause as well as the new one. To differentiate, we could separate the newly defined cause and call it eXcess eXDAM, or 2XDAM. This isnít working so well.
Why do we need a new name anyway? It still good old amplitude modulation distortion. We donít invent a new name every time we discover a new cause of distortion. Sometimes we do. Take transient intermodulation distortion (TIM). It and its refinement, slew induced distortion (SID), named measurement methods to uncover this illusive distortion. (It turned out to be so illusive, or imaginary, that we donít hear of it any more.)
So what is the best measurement method for Markís distortion? Mark finds likely candidates by measuring frequency response difference at excursion, which he calls FRAX. When he finds a suspect, he measures modulation products at suspect frequencies. The amplitude of the products is described and the order is noted. This does not yield a simple number or plot for comparing speakers.
What we need is a method for revealing the presence of this cause of modulation. Out of this procedure will probably fall a good name. Letís try.
I would apply a low-frequency signal around the resonance or below that strokes the speaker through a portion of its full excursion. Added to this would be a small swept signal that covers the operating range of the speaker. A mike would pick up the acoustic output, a filter would remove the LF signal, and the HF would be analyzed for peak to peak amplitude variation. This would be expressed as a percent of the average amplitude. This would result in a plot of AM distortion vs. frequency for that stroke. Other stroke amplitudes could be plotted on the same graph. Alternatively, worst-case distortion could be plotted as a function of stroke. I would call the measurement swept amplitude modulation distortion. SWAM, if we must have an acronym.
A bad speaker would be referred to as SWAM I am. Testing could be called taking it for a swim. A rejected speaker would get the SWAM jam or a SWAM dunk. Having to listen to one would be SWAM spam.
by David Carlstrom
Fifteen years ago consumers and audiophiles had the opportunity to choose compact discs or reject them in the marketplace. The fact CD's adoption was not a sure thing may not stick in our minds, but no designer knows for sure if a product will be accepted. Today we are offered DVD and DTV each with advanced picture and audio resolution and are told these will replace CD ad TV as we know them today. I am not so sure.
My reminder is a technological advance of the late 1940's. For twenty-five dollars in a rural antique shop my sister picked up an RCA Victor 45 RPM record player and promptly sent it to me with a plea to make it work. It immediately struck me that this player represented both real technological advances and a marketplace fizzle. The machine only plays 7 inch 45 RPM records. Since it has no backwards compatibility with the prior 78 RPM art, clearly RCA meant it be a revolutionary product to completely replace 78's. The record is smaller, lighter,easier to handle and store due to the use of vinyl and a large center hole. The little player is far cheaper and simpler in operation than any 78 RPM phono with a record changer. The arm is lightweight plastic and tracks at remarkably low 6 grams. The fidelity of the microgrove 45 surpasses the 78 in signal to noise and bandwidth. Clearly RCA brought it to market with the sincere expectation the innovations would get the public to dump both their big heavy Victrolas and brittle easy to break 78's.
As seems obvious to us now, the 78 was not killed off by the 45 RPM system alone. It took the LP and the 45 together to do that and the changeover process was slow. The 45 was never anyone's complete record player. The 45 found only an unplanned niche as a single song media, while the LP opened new territory with its longer uninterrupted playing time. The RCA 45 missed the mark because it couldn't play any longer than a 12 inch 78. The 45 system clearly is chocked full of advantages over the 78 and equals everything else about the prior state of recording art, but that alone was not enough. How could RCA have foreseen that longer playing time meant more to the music fan than the convenience of a smaller record and a large center hole?
Who is to say the DVD and DTV won't suffer the same fate by being good but not good enough. The RCA team thought they had a winner, but while the 45 became the dominant single song media, they failed at their goal to completely replace the 78. The advocates of the digital media schemes think they have clear winners, but are they good enough to sway an entire market. While we each have an idea about the fate of DVD and DTV, I think the promoters would be wise not to be so sure. I don't think anyone knows the fate of new digital systems any more than RCA knew what would happen when they introduced the 45 RPM system. Looking at the clear advantages of the 45 I am sure I don't know if the DVD and DTV are convincing enough to sway an entire market.
Toasting Tweeters- A Silly Survey
by David Carlstrom
Twenty-one of the thirty-two (65.625%) members and guests at the SMWTMS meeting in January 1992 responded to the survey question on the number of speakers they had damaged.
The largest number of speaker kills was for tweeters. The modest power handling ability of the bird-like drivers may explain the high mortality. But would a better explanation be an abusive tweeter stalking trait in SMWTMS members revealed by these statistics? Twenty-seven of the little speaker-babes were destroyed in the hands of the respondents, or 1.286 per person. A middling one per person, or twenty-one woofers were met their paper's end at the hands of the survey group. Midranges fared better with thirteen expirations, only 0.619 per person.
Overall, sixty-one speakers were snuffed, or 0.968 per person. Lest this seem a low number, David Clark had to be excluded from the scores. He has no doubt exterminated more speakers than the rest of the group combined and replied, with humor or modesty: Woofers - 4; Midranges - 8; and Tweeters - 16. Runners up in speaker murder-death-kills were John Budrys reporting 10-0-6, and Arny Krueger at 3-2-2.
by John Tyrer
For the second time in as many months, SMWTMS members were treated to a performance of live music. Unplugged, as they say.
SMWTMS founder, violinist, violist, and man-about-town, Bern Muller, sartorially resplendent in white tie, tails, and jeans, played for our enjoyment and edification, the Bach Suite #1 for unaccompanied cello. The viola is not an instrument one often hears played alone, and Bern's presentation was quite a treat. Complete I should add with printed concert notes.
I have heard a few musical instruments live in my living room such as the acoustic guitar, banjo, and string bass, but because I'm usually quite close to the instruments - especially the bass - I never really heard them. This time I was well to the back of the room, just inside the door by the kitchen. Not what I would call an ideal listening position.
But much to my surprise, this did not detract from the listening experience one whit. The angular walls and vaulted ceiling, all hard surfaced, that combine to murder the sound of my home audio system only seemed to enrich the sound of the viola. From where I stood, the instrumental sound was brilliant, clear and resonant. And where were all those room peaks and valleys?!
Bern's playing was most ingratiating. Those few notes that may not have been up the Bern's standard shrank to insignificance compared with the enjoyment of hearing such music live. For one accustomed almost exclusively to listing to recorded music, I found the experience both illuminating and enormously satisfying. I thank Bern for having the courage to stand up in front of what can be, at times, a hostile crowd, and giving us a taste of the real.
To those in SMWTMS even moderately musically inclined, I urge you to share with the rest of us your personal musical point of view. We'd like to hear you play.