Jump to content

AR 9 and the anechoic chamber: An Out-of-Room Experienc...

Guest Goyescas

Recommended Posts

Guest Goyescas

Hello forum readers. I am new here and this is my first posting. On Saturday, Feb 21, 2004, a friend and I installed a pair of AR-9's in my back yard "listening room." I hope Barrydor posts a response, as he is the friend I speak of, and the progenitor of this project.

My intentions here are to simply post the results of our preliminary evaluation in order to further the discussion on these speakers and their (in)ability to perform well in modern home environments.

As a thesis of sorts, barrydor suggested we should hear these speakers in the closest approximation to an anechoic chamber. The other objective, from my own view, was to hear what a speaker which was designed to sound flat in an anechoic chamber sounds like.

I must start by stating by biases, which are formidable. I own these speakers, but do not really like them. Difficult to explain, but they sound like 9's. Blindfold me in any room and I would guess they are the 9's, so obvious is their sonic signature.

I don't know why anybody would want a speaker designed to measure flat in an environment (anechoic chamber) which fails to even approximate real world conditions; it makes no sense to me to build such a thing - instead I think speaker makers should concern themselves with the rooms of the buyer and not the anechoic chambers of scientists. So my bias is strongly against these speakers, although I remain open to hearing them at their best, which is the point of this post.

Now I realize this is a lofty subject, with room design, construction materials, and speaker location driving a very strong train that is the reproduction of music in ones home via a hi-fi system. But Saturday last we heard these speakers, I think, at their very best, adjusting of course for aging drivers, caps, and the like.


Musical Fidelity A-300 Integrated 150wpc dual mono amplifier

Magnavox CD-67SE

1M IXOS Gamma Encore IC (CDP-Amp)

10' bi-wire 1.4's w/Mlock/Mstr Pins (speaker wire)


Car Wheels On A Gravel Road, Lucinda Williams

Walter Schumann's "Voices" Lp recording on CD

Mahler Symphony #2, EMI, Otto Klemperer Cond.

Kind of Blue, Miles Davis

Prokofieff's Cinderalla (full score!), RSO

Time of Day: 10am-2pm

Temperature: 70-76 degress F

Humidity: app 70%

Area: Southwest Florida, app 10 feet above sea level

Conditions: little wind, but some, partly cloudy

Date: February 21, 2004

What follows then is a transcript of my written notes on the experiment (phase I).

BASS: Flawless; tight but pitch so well defined it boggles the mind. Listening to "So What" from Miles Davis' "Kind of Blue" revealed more HIGHLY RELEVENT musical information to have transformed the piece. As a player of double bass, I was A-S-T-O-U-N-D-E-D by what I heard. The music is not really centered around a specific harmonic progression, being instead modal, which makes the bass line far more important in that its work is not implied by the ear when absent. I will never be able to listen to this record (well, in this case, CD) the same way. It was a draw-dropping experience.

The dispersion characteristics of the 9's were as claimed; nothing beaming and everything well integrated. As difficult as this is to explain, the 9's did not emphasize one frequency over another. The soundstage was hardly neither deep nor wide, but the instruments sounded right.

We found a distance of 10' - 16' feet to be optimum; and closer or further and things sounded queer.

The clarity/intelligibility was stunning.

Bass Lines NEVER so articulate in my experience of 20+ years as an audiophile.

Timbre: Delineation does occur; one knows what one is listening to. But the micro-timbre stuff does not get through, no surprise considering the birds, lawn mowers, and other temporary but nevertheless spatially confusing cues. The noise floor was not measured (again, this is our preliminary evaluation!), but it was significant. Next time we will wait until 3am.

I have written next the following:

"Basically, a religious experience, will hear the room when it is obvious; meaningful b/c musically relevant and musically critical information us actually available!

Re Mahler 2: "What is on the CD, you hear it all.

Timing: questionable; need more data/better sources to evaluate. Drivers sound a bit slow. Some break-up/distortion issues.

Soundstage on classical music recorded in a hall: NOT GOOD; ill defined front and back, not a rear-hall sound."

This I think gets to the illusion of sound staging, and more to the point into the room reflections that help create this "soundstage" illusion.

There were other spatial cues that could or did very well throw off this soundstage. As barrydor points out, the "birds sound like taradactals."

Well, here it is: If you have a pair of these speakers drag them outside and hear what an anechoically flat loudspeaker sounds like anechoically (or close to it). It will BLOW YOUR MIND.

Next time SPL meter and some other measurements.

Thanks for listening. I am leaving town until March 1, 2004, but wanted to get these notes down and out here while they were fresh in my feeble mind. I imagine barrydor may participate in any discussion, which may ensue.

Kindest regards,


Link to comment
Share on other sites

  • Replies 60
  • Created
  • Last Reply


>As a thesis of sorts, barrydor suggested we should hear these

>speakers in the closest approximation to an anechoic chamber.

>The other objective, from my own view, was to hear what a

>speaker which was designed to sound flat in an anechoic

>chamber sounds like.

Listening to speakers outdoors -- unless you hoist the speaker up on a platform 30 feet in the air, or bury them flush in the ground, and you get up in the air on-axis to listen -- is not truly an anechoic experience. You are still going to have ground reflections. You remove part of the double-reverberation effect, but only a portion of it. Besides, an anechoic-listening experience, if it is truly echo-free, is a dreadful experience. On the other hand, recording out doors, with fairly close miking, can be sort of an anechoic thing, and has been successfully done before, even by AR.


>I must start by stating by biases, which are formidable. I own

>these speakers, but do not really like them. Difficult to

>explain, but they sound like 9's. Blindfold me in any room and

>I would guess they are the 9's, so obvious is their sonic



All speakers, regardless of design, exhibit some amount of coloration, and all speakers will sound like speakers in any normal listening room due to double-reverberation effects of recordings. You are getting the effects of a recording's reverberation and the room reflections, plus the standing waves, etc. So what you are describing above is more an effect of the room, or rooms, than the speaker itself. The AR-9 has relatively low coloration, even when compared with some very expensive, so-called "reference-standard" loudspeakers. They really don't call attention to themselves with blatant coloration or a sonic signature, which is to say they are actually very accurate reproducers. Nothing is perfect, but the AR-9 is high up on the list of accurate loudspeakers.

>I don't know why anybody would want a speaker designed to

>measure flat in an environment (anechoic chamber) which fails

>to even approximate real world conditions; it makes no sense

>to me to build such a thing - instead I think speaker makers

>should concern themselves with the rooms of the buyer and not

>the anechoic chambers of scientists. So my bias is strongly

>against these speakers, although I remain open to hearing them

>at their best, which is the point of this post.

Well, here is the problem with your reasoning: the definition of high-fidelity sound is "the accurate or faithful recreation of the original sound source." If the signal source for measuring frequency response (both on- and off-axis) exhibits flat, or uniform, response accross the continuum of frequencies, then the reproducer should also replicate same thing, measured in an echo-free environment or without the effects of reflections. Since no two rooms are alike, for the reproducer to be accurate one must use either room treatment or equilization or whatever to ameliorate the shortcomings of your listening environment.

--Tom Tyson

Link to comment
Share on other sites

I did love the title of the thread.

Very interesting "test." I may have to try that myself after it warms-up a bit and dries-out a bit here.

But it will be strictly a "for fun" thing since I can't listen that way most of the time.

Several comments, most of which are obvious and trite, but at least one of which might be worth your consideration -

My memory says that AR did not use an anechoic chamber to design many of their speakers, but rather used a "semi-reverberant" chamber in an attempt to more closely design with a real room in mind; yet not reverberating to the point that they were measuring the room more than the speaker. I'm not an engineer and so didn't memorize the characteristics of the chamber because it hardly matters - my room is my room.

Second, and probably most importantly, one of the real problems with reproducing any sound in any room is the recording. You just never know what the engineers did to the music when they recorded and equalized and compensated and mixed and mastered it.

Third, unless the music is miked close and binaurally in an anechoic environment, you just are not going to get a real reproduction of what those instruments would sound like in your room, period. So, any "realism" one speaker exhibits over another is as much a figment of your (my, our) imagination as any other. Obviously, it is really impossible to get an accurate reproduction of a symphony in your living room mostly because unless you are a Vanderbilt you can't get a full symphony into your living room. So trying to get an accurate reproduction of what the symphony sounded like in the symphony hall in your room would require a binaural recording and that your room be anechoic.

Fourth, I realize that you listed very fine equipment with expensive speaker wire, etc, however, I'm lucky enough to have a local friend as you do and we have run many, many experiments using many different pieces of audio gear with speakers being the only constant. I started line- listing what equipment we've goofed-around with and realized that it really doesn't much matter. Suffice it to say we've used a lot of equipment with a lot of different interconnects (and even same interconnects of varying lengths) and the *only* constant we've found is that there is no constant.

Okay, that isn't entirely true, either. Sony 400 disc CD or CD/DVD changers don't sound very good no matter what interconnects you use with them or what you are connecting them to.

But there have been times that one setup sounded very listenable on the AR-90s that then sounded absolutely anemic on the 10pi's and vice-versa.

In all our listening sessions where we are A/Bing or swapping equipment or cords or something my cliche line has become, "It's _______ as an equalizer." We have yet to find a perfect piece of gear or a perfect combination of gear. It seems to both of us that by emphasizing or de- emphasizing any part of the music spectrum that we hear more or less of some things while hearing less or more of something else. We're listening, of course, in an attempt to read the manufacturer's name off of the piano and tell what brand of sneaker the bass player is wearing, or take the percussionists fingerprints.

Just as importantly we have yet to find the "perfect recording" actually preferring that which we both suspect is inferior (equipment-wise) if the recording is so bad that it *needs* coloring or filtering or something in order to make it enjoyable.

What I am suggesting, to the suggestible reader, is that you might try listening to your AR-9s using different interconnects (even cheap ones) or a different amplifier or preamplifier or CD deck and see if you like them better. They will certainly be different. Better? Maybe / maybe not. But don't assume that because so-and-so magazine says this CD player is wonderful that it will be in your room. (I know you know that already, I just like hearing myself talk.)

I would seriously suggest that try some different equipment regardless of how good yours is. I'm sure it will surprise you even if you don't much care for the results.



Link to comment
Share on other sites

The irony is that the 9 was one of the first systems actually *designed* with practical room considerations in mind. From the Acoustic Blanket to the upper driver arrangement to the location of the side-firing woofers, this speaker was meant to successfully integrate into a real-world listening room.

From years of comparing the 9 to other speakers, I would characterize it's "sonic signature" as that of uniform frequency response, in that it closely mirrors the recording. And my guess is that the break-up and distortion issues are a function of either out-of-spec crossover capacitors (possible, but not probable), or a misbehaving amplifier (as has been well-documented in this forum, some otherwise fine amplifiers do not suit the AR-9's prodigious current requirements).

Link to comment
Share on other sites

Reading the rataionale behind the side firing woofers and the front firing lower midrange in the design paper on this site demonstrates that far more thought went into the integration of this speaker into real rooms at least as far as bass reproduction is concerned than any other speaker I can think of. And it works well in that respect. My problem with this speaker is that it does seem to me to have a sonic signiture which I've figured out how to overcome with great difficutly and patience. Its main problem as I see it is its restricted high frequency dispersion, which although certainly no worse than any other front firing speaker is inadequate to compensate for the early reflections at mid and low frequencies it produces. This problem cannot show up in an anechoic chamber and has been given virtually no attention by audio equipment designers. However, audiophiles know the problem all too well even if they don't understand the cause. They have tried many different schemes to get around it including dead end/live end acoustic treatments, pulling the speakers away from the back wall (killing the deep bass at the same time), bipolar radiators, and indirect firing tweeters. IMO, using supplementary indirect firing tweeters (I use three per channel) works best and allows the speakers to remain against the wall. Very careful equalization can compensate for the spectral imbalance this creates and the program level controls help a lot in this regard. With the speaker augmented this way and properly equalized, it is hard to find fault or room for improvement in the sound of the Teledyne AR9 IMO. It is not merely spectrally accurate, it is the only speaker I ever heard that can play with great power and extreme delicacy of detail at the same time.

Link to comment
Share on other sites

Guest rickcee

Well, just to throw in a wrench . . . Bose 901 series - designed totally for 'room reflections'. . . .? As well as the LST and various original Allisons .

AR itself sold models with designed in limited dispersion (well, European 'connisour' series, I believe. mid '80's)

Seems to me it just comes down to what you like, maybe a tightly controlled dispersion.(probably 'brighter' but not required.) Maybe a wide dispersion sortof surround sound.- probably for most, somewhere in between. As stated, there's all that production / recording engineering in there, as well as listening area, etc. Rick

Link to comment
Share on other sites

I don't agree. I still own a pair of original Bose 901s in perfect condition. This loudspeaker had at least one serious problem and that is that it could not produce the highest octave of sound very well because the 4 inch drivers had so much inertia that their response fell off at too low a frequency. As a result, the overtones which are so important to accurtely reproducing the timbre of musical instruments are lacking. IMO, this is the major reason most audiophiles find them objectionable. Also, their placement and finding a suitable room for them seems far more critical to getting the optimum performance form than than most other lodspeakers.

The question is not one of preference. It is one of accuracy. As I said, perhaps not explicitly enough, I have come to the conclusion after much consideration that one of the major failings of most loudspeaker systems is that they have a relatively limited high frequency energy dispersion characteristic compared to their radiation of middle and low frequencies. The proof of this is very simple, just look at the polar response graph of loudspeakers or even just the tweeter response as a function of angle off axis. The result is that in a real room, not an anechoic chamber, you must hear early reflections caused by surfaces near the loudspeaker but these reflections contain no energy in the highest octaves, only in the middle and lower octaves. This distortion leads manufacturers to make their speakers overly bright to compensate but the results are unsatisfactory at least to me. An additional indirect firing array of tweeters can be used to create early reflections at high frequencies bringing the speaker back into proper spectral balance over the time spanning the first few milliseconds of each note. I find this far more satisfactory.

Link to comment
Share on other sites

>The irony is that the 9 was one of the first systems actually

>*designed* with practical room considerations in mind. From

>the Acoustic Blanket to the upper driver arrangement to the

>location of the side-firing woofers, this speaker was meant to

>successfully integrate into a real-world listening room.

This is a GREAT post and very accurate. Thank you for summing things up SO well in so few words.

>From years of comparing the 9 to other speakers, I would

>characterize it's "sonic signature" as that of uniform

>frequency response, in that it closely mirrors the recording.

My experience is that the stock 9's have a somewhat sloped frequency response i.e. tilted up in the warmth region with a very slow but steady roll-off as the frequency rises. There is a slight "peakiness" or "ragged" characteristic in the upper mids. Much of this is due to the 1.5" dome breaking up at higher spl's. The problem with the domes breaking up can be further compounded / more apparent if the part values in the crossovers have shifted with age.

>And my guess is that the break-up and distortion issues are a

>function of either out-of-spec crossover capacitors (possible,

>but not probable),

See my notes above about out of spec caps.

As a side note, these speakers can be DRASTICALLY improved using the following methods. These are not listed in any specific order as they are all important.

1) Upgrade the internal caps. How expensive you want to go is up to you. The use of low grade "budget" caps may be a step backwards from the stock caps. Solen's and / or Axon's should be considered minimum. Having said that, they are a VERY acceptable minimum. Keep all the leads on the caps as short as possible. Otherwise, they tend to ring and smear the sonics. If you doubt this, hold the body of the cap and "flick" the leg with your finger. Damping the lead where it enters the body of the cap is not a bad idea.

2) Bypass the attenuation network. All of the drivers running wide open are still below the output level of the 12's. As such, there is no need to attenuate drivers that are already quieter to begin with. Not only does this reduce loss in the circuitry, it keeps the signal path shorter with quite a few less connections. You might not think that a "few connections" are audible, but you might be surprised. You can leave them in place for cosmetic purposes, but get them out of the circuit path.

3) Replace the internal wiring. Low inductance cable should be used. I used Kimber 8TC for the woofers and 4TC for everything else. If you want to do this inexpensively and get good results, i can suggest other wire to use.

4) Replace the factory binding posts. These things are a horrid joke and were installed simply to torture those of us trying to run real wire.

5) Add a few more internal braces to the cabinet. The cabinet is "relatively sturdy", but it still resonates. I added one from the back of the mid-woofer chamber to the rear panel. I then crossed right over the top of that brace and installed another from side to side. These were so close that i glued the two braces together, adding further rigidity. I then installed another brace higher up in the cabinet running from side to side.

6) Line the internal cabinet walls as thoroughly as possible with acoustic foam. This can be pretty tough to do due to the very tall and narrow passageway inside the cabinet. If you can't work with the foam, use thin self-adhesive damping sheets. Something is better than nothing.

7) Line the walls of the chamber for the 8" mid-woofer with a think layer of acoustic foam.

8) The baskets of the woofers and ESPECIALLY those of the mid-woofers should be heavily damped. The use of rope caulk or "liquid nails" type material applied in layers works wonders. If you would like to test this out, pull one of the 8's from the box and "flick" the basket of the driver with your finger. It will ring like a bell. While this IS reduced via mounting the driver, the sonic colouration of the basket being excited IS both measurable and audible. The 12's also ring when excited, but not nearly as much as the 8's do.

9) If you're really interested in doing things properly, you CAN fit all of the parts from the two crossover boards onto one board AND maintain proper spacing without having the fields of the coils interact with each other. This makes it easier overall and reduces cabling from one board to the secondary board, making the signal path shorter with fewer connections.

10) Add at least one solid layer of damping material ( i prefer fiberglass, your option to use foam or poly fiber-fil ) over the top of the crossover board(s). Take the necessary steps to keep this in place.

11) Install some type of cone or spiked footers to the bottom of the cabinet.

12) Make sure that the cabinet and all of the drivers are properly sealed. Use of "smashed down" rope caulk as a gasket works great.

13) You can install aftermarket inductors if you would like. If you do this, i would recommend using something along the lines of Goertz flat wire inductors. The stock inductors work pretty well, but if you've got the cash.... If you re-use the OEM inductors, try to orient them for the same polarity. That is, use the wire coming out of the center of the toroid as positive and the wire at the outer edge as negative or vice-versa. It really doesn't matter so long as you are consistent on input & output from coil to coil. Don't forget to position the inductors so that they aren't all laying in the same vertical or horizontal plane. This minimizes crosstalk between the various drivers.

14) While you've got your speakers apart, look the 12's over VERY carefully. I've found that many of these drivers are suffering from damage that many are unaware of or simply over-look. Pay close attention to where the voice coil former is attached to the spider. Due to heat and / or long excursions, i've seen several of these drivers come apart here. If you find one driver like this, i would recommend re-gluing all of them. Take a look at the 8's while you're at it. Due to the higher cut-off frequency though, these should be fine.

All of this should be done in addition to retaining the factory installed damping materials and bracing. In other words, adding the foam to all of the internal walls, mid-woofer chamber and over the crossover boards does not replace the the factory installed poly damping material. All of this is doneand added IN ADDITION TO what was already there.

>or a misbehaving amplifier (as has been well-documented in this >forum, some otherwise fine amplifiers do not suit the AR-9's >prodigious current requirements).

Most people don't have anywhere near enough sustained or peak power capacity to obtain optimum results from these speakers. I've mentioned this before, but you really need BEHEMOTH sized amplification to make this things sing. Those that think 350 - 400 wpc @ 4 ohms is "behemoth" sized will be sadly disappointed with the results. That is, if they are trying to obtain high peak or sustained spl's with these speakers and do so cleanly. Your amp will give out before the speakers will.

If you perform all of these tasks, you'll be pretty shocked at how open, transparent, fast and cohesive these speakers sound. The difference is VERY noticeable and beneficial. If you've got a REAL amplifier, you'll have retained the speed and tunefullness of the bass and should now have even greater impact i.e. "slam" on the bottom end. I have performed very similar tasks to these and other similarly constructed speakers and the results are quite staggering and HIGHLY worthwhile. I just got done doing a set of Legacy's for my Father using this approach and the results were well worth the effort invested. As a side note, applying felt to the baffle on his Legacy's ala the 9's "acoustic blanket" worked wonders in terms of reducing treble smear while improving imaging. If you would like to read a bit more about the Legacy mod's, try the web address listed below. If you have further questions about these or other modifications, let me know. Sean



Link to comment
Share on other sites

While this response isn't taken directly in the context posted, i "slightly" re-arranged things to prove a point.

>I have come to the

>conclusion after much consideration that one of the major

>failings of most loudspeaker systems is that they have a

>relatively limited high frequency energy dispersion

>characteristic. / >As a result, the overtones which are so important to accurtely reproducing the timbre of musical instruments are lacking.

This is VERY true and why most "boxes" sound like boxes. Thinking outside of the box will get you better radiation characteristics with a more natural presentation. Not only are the harmonic overtones "squished", the notes lack the dimensionality that they would normally have when being played live in an acoustic venue. This is one aspect where the 9's, even in highly modified form, can't compete with some of my other speakers.

>An additional indirect firing array of tweeters can be used to

>create early reflections at high frequencies bringing the

>speaker back into proper spectral balance over the time

>spanning the first few milliseconds of each note. I find this

>far more satisfactory.

While you might prefer this, and i can understand why in terms of a more even tonal balance, the idea of a full range omni-directional driver is that it is time and phase coherent across the audible band. While Bose chose to use multiple drivers, which creates its' own problems, the idea that you mentioned above ( wide high frequency dispersion ) combined with a single driver ( to maintain time and phase accuracy of both the primary notes and harmonics ) is an incredible thing. As far as i know of, there were only two drivers made that did this and they are no longer available. Care to take a guess at what they were? Sean


Link to comment
Share on other sites

When I referred to bringing the speaker back into spectral balance over time, I am not referring to the usual notion of time coherent drivers where the time delay of one driver is comparable to the time delay of another. Some people refer to this as phase coherence. As far as I know, there is not proof or evidence that so called time aligned loudspeakers are in any way audibly different than identical non time aligned speakers. Early time aligned models like Rectilinear Five which merely tilted the speaker back a few degrees placing the tweeter slightly further away from the listener than the midrange also put the listener off the tweeter axis changing the frequency response at the listener. Moving the tweeter back from the woofer makes its relative loudness slightly less. When any two sources radiate at the same time, there will inevitably be cancellations and reinforcements because they are not at the same location. Just throw two rocks in a still pond at the same time and that will become obvious. What I am referring to is the total energy of each note as a function of frequency over time, something AR aluded to but never convinced me that they got even close to right. The problem isn't just the total energy radiated as a function of frequency, it's the energy as a function of frequency distributed over time that reaches the listener. The typical loudspeaker radiates energy fairly uniformly in all directions except at high frequencies. Most manufacturers don't see this as a problem. The directly radiated component has a flat frequency response. But the listener immediately hears a large number of echoes from nearby objects which have no high frequency components. This is unavoidable unless your room is 100 percent dead. IMO, the difference in the radiating patterns as a function of frequency is the main reason two speakers with identical on axis frequency responses sound different.

I have half heartedly played with the spectral radiating pattern of Bose 901 with limited success. Because of the active equalizer, you have basically two choices. First, you can choose tweeters with relatively limited hf response and accpet the treble boost provided by the equalizer, or you can bi-amplify and use the same kind of wide range tweeters I used successfully with AR9 and KLH6. So far I have done the former and even worse, I've used only one indirect firing tweeter per channel. The results are encouraging but not entirely satisfactory. One additional challenge Bose 901 presents is that unlike multiway forward firing speaker systems, it's forward high frequency response is also unsatisfactory. Therefore more drastic measures are called for. My next serious step would probably be a four tweeter array per channel with one soft dome firing forward and three polys firing upward using the unequalized signal of course. I would probably have to also cross over the front and indirect tweeters at different frequencies and tweak the relative levels of each.

Link to comment
Share on other sites

I have been following the thread on “AR 9 and the anechoic chamber”

and the suggested improvements by Sean. I own a set of AR90s. While I like the idea of having the “perfect AR90 speaker” I am reluctant to spend heaps of money and fiddle (sometimes it backfires with undesired results) but this thread inspired me look for a quantitative way to identifying if I really need to fiddle with the crossover networks. Also I wanted to identify if the crossovers are working as listed in the AR specifications, namely 20, 1200, 7000.

I have tried the following but would like some comments on how to interpret the results.

I recorded four tracks from a master CD which has a sweep frequency track running from 20Hz to 20kHz. ( I think my sound card starts attenutating after 16KHz).

Track 1 - 20Hz to 200Hz, 5 seconds 1kHz tone L&R, repeated 3 more times

Track 2 - 200Hz to 1200Hz, 5 seconds 1kHz tone L&R, repeated 3 more times

Track 3 - 1200Hz to 7000Hz 5 seconds 1kHz tone L&R, repeated 3 more times

Track 4 - 7000 to ??. 5 seconds 1kHz tone L&R, repeated 3 more times

As there is no analogue filtering required, I assume the cutover points should be very accurate.

I listened at low levels with one ear to the sound from the speakers above and below (where appropriate) about 6 inches away from the speaker and very lightly touched the speaker cone. I found that the frequencies appeared to run into the next speaker. I assume crossover rolloff was coming into affect.

Before I decide whether I will “fiddle” or not, I three questions I would appreciate comments on;

---- Will this test (or a modified one ) quantitatively identify if I have to upgrade my crossover components, replacement or enhancement if the frequencies move into the next speaker?

----Are the nominated crossover frequencies the best frequencies to use or would the design of the frequency handover and rolloff influence the decision?

----How long should I expect to hear the frequency in the adjacent speaker due to the rolloff of the crossover?


Link to comment
Share on other sites

The crossover points are gradual, Richard, so you're definitely going to hear spillover from one driver to the next - that's absolutely normal. Without a real signal generator, and some way of quantifying the output (and duplicating it for the second AR-90), it's going to be difficult to reveal anything other than gross anomalies with the CD test tones.

Link to comment
Share on other sites

I/ve a pair of AR90s, and did upgrade the crossover; well, rebuilt them from scratch truth be known. The change in clarity, definition, and overall openness was profound. Each note of each instrument took on a clarity that I wasn't aware was missing beforehand. Yet, they still sound like an AR90, just one heck of a lot better.

Link to comment
Share on other sites

Sounds like a lot of work trying to turn a sow's ear into a silk purse. Having said that, i used to own a set of the original 901's and still have all of the drivers from them. Harry Pearson of TAS stated that, as far as he knew, the Bose 901 was the only speaker that has gotten worse with every new incarnation.

Other than that, why not look for a Walsh tweeter that Infinity used to use many moons ago? That will give you both the direct and reflected sound that you are after. Mounting one on top of or below the 901 ( depending on if the 901's are on stands or hanging ) will approximate the center of the front and rear drivers. This should provide you with a simple but elegant solution to your problem. Sean


Link to comment
Share on other sites

Timbertop: AR Pro is right on the money. To add to his response, i would also mention that not all of the crossover slopes used in the 9's and 90's are at the same rate. That is, one driver might come in / fall off faster than another. This is purposely done so as to sum the total response of all of the drivers for a relatively flat response. As such, you have to be able to take complex measurements of each driver and them sum them together as a whole to actually see what the big picture looks like in terms of frequency response. After doing so, one might be able to "wiggle" certain aspects of the crossover around, but i would be VERY careful.

While i am generally loaded with "contempt" for most loudspeaker designs / designers, the 9 / 90 is pretty well researched and put together. If i was to make one change in specific, it might be to bump the crossover frequency between the 8" and the larger dome from 1200 Hz up to appr 1500 Hz or so and use a sharper slope on at least the dome. This would increase power handling and reduce break-up at high spl's just a bit. One would assume that going just a bit higher than this would be even better in this regards, but the 8" starts getting nasty around 1800 Hz and above, so crossing over below this point helps avoid adding just one more variable to what could be a mess : )

Other than that, you can check the actual slopes of the crossover individually with the use of a signal generator / amplifier and misc test equipment. Most signal generators will not directly drive a loudspeaker with any type of linearity and checking the crossovers without the actual impedance of the driver being used in that circuit will alter the actual measured response.

As a side note, the improvements made to these speakers after performing such mods is like night and day in terms of overall transparency and musicality. What was a good design but limited by manufacturing production costs is now infinitely more refined and far more dynamic sounding.

RRCrain: I agree with your basic assesment of the results, but wonder what exact steps you took with the rebuild / modifications of these speakers? Would you mind sharing notes with us? Sean


Link to comment
Share on other sites

"Sounds like a lot of work trying to turn a sow's ear into a silk purse."

It's more like trying to turn a leather purse into a silk purse. It is foolish to ignore the special qualities of Bose 901 by dismissing the entire concept. While audiophiles find the shortcomings of this speaker unacceptable, Other listeners including non audiophiles and audiophiles of yesteryear find that the special qualities of the direct/reflecting principle have much of value that cannot be duplicated by other designs including omnidirectional designs like Ohm's. The Walsh driver is not ideal for my purpose. The reason is that the requriements for the front tweeter and indirect firing tweeter array are different if for no other reason than volume alone. The front firing tweeter must integrate its output both frequency wise and volume wise with the single 4" front firing driver. The indirect firing tweeters must compensate for the variables of both the absolute and relative frequency absorbancy of the room acoustics near the speaker. So while the front firing speaker has a fixed requirement, the indirect array must have variable adjustments for both frequency contour and volume. Is it worth it? I'm betting it is. To bad I have to do Bose Corporation's work for them. This is something they should have figured out and incorporated in their design a long time ago. If they had, they might have found much greater acceptance for their product among audiophiles today. But they probably don't care. They're laughing all the way to the bank. Come to think of it, I'm surprised Sidney Harman hasn't bought them out yet. Don't be surprised if you wake up one day and find out that he just did.

Link to comment
Share on other sites

>the front firing speaker has a fixed requirement, the indirect

>array must have variable adjustments for both frequency

>contour and volume. Is it worth it? I'm betting it is. To

>bad I have to do Bose Corporation's work for them. This is

>something they should have figured out and incorporated in

>their design a long time ago. If they had, they might have

>found much greater acceptance for their product among

>audiophiles today. But they probably don't care. They're

>laughing all the way to the bank. Come to think of it, I'm

>surprised Sidney Harman hasn't bought them out yet. Don't be

>surprised if you wake up one day and find out that he just


I believe that the Bose 901 is a much-maligned, over-critized design. Part of the problem comes from one's mental image of all nine 4-inch drivers trying to reproduce 20,000 Hz, while at the same time attempting to pour out 30 Hz tones. Furthermore, the necessity for equilization to make the speaker approximately flat, and the somewhat bloated image that can result on single instruments, have all worked to hurt the audiophile image of the 901.

The speaker is really not that bad. Properly installed, it can sound very realistic. It may lack certain qualities of audiophile-type speakers with multiple drivers, but overall it is capable of giving the illusion of a concert hall, and this is in fact the very goal of high fidelity.

One thing is for certain: it is an enduring design, has sold in the tens-of-thousands, and continues to be available today after nearly forty years of continuous production. A lot of that is due to Amar Bose's edict that it shall be made available, but it nevertheless continues to sell well in spite of its age. I don't know the figures on Harman Industries vs. Bose, because the latter is a privately held corporation, but I would think that Bose is at least as large, if not larger, company than Harman. The latter has a lot of commercial divisions, such as Crown, JBL and such, but Bose is very innovative, clever and the undisputed market leader in consumer-audio today, and I suspect that the company is very profitable.

--Tom Tyson

Link to comment
Share on other sites

Bose 901 does severeral things right that most other speakers including AR and KLH do very wrong. The spatial radiating pattern of Bose 901 is much closer to the radiating pattern of real instruments than front firing speakers. For recordings of instruments or small ensembles intended to perform in relatively small venues, this difference brings the sound to life.

Anyone who doubts that a single driver can reproduce the entire audio spectrum should consider that the performace of the best headphone drivers easily accomplishes this. But headphone drivers have a big advantage. They operate at much lower power levels, they operate into a known acoustical environment with a fixed relationship to the ear, and they operate directly on axis to the ear. The problem for creating a single loudspeaker driver that is of comparable quality is a materials problem and one of geometry and diameter. At the moment, it's beyond the state of the art.

In my experience, Bose 901 works best in an acoustically live room where there are side walls on both sides of the back wall. The back wall should be relatively empty with no absorbant materials. I got the best results with the speakers on their pedistals, the backs of the speakers 15 inches from the wall, and the speakers fairly far apart with the fronts towed out (backs pointing in about 30 degrees.) This is one of the few speaker set ups in my experience where a two channel arrangement completely eliminated the hole in the middle.

Some people complain that Bose 901 makes a piano sound 10 feet wide. These people have obviously never seen a concert grand like a Steinway A or B. I never experienced such a problem with a violin recording.

In experiments several years ago, with an outboard tweeter and careful equalization, it was possible to get Bose 901 to have a remarkably similar spectral balance to AR9. However, I can't say whether it is my lack of adequate available electrical power or inherent in the 901 or both but frankly in my experience, the original Bose 901 is no match for AR9 in deep bass. AR9 can be made to sound just like rolling thunder shaking the walls, rattling the windows, and vibrating everything in the room. It also easily excites every resonance node in the room, a drawback I accept as inevitable. Unfortunately when moved to another room, Bose 901 no longer sounded like AR9 any more. Back to the drawing board.

BTW, the amplifier I used to drive it for many years was the AR integrated.

Dr. Bose made several errors or at least misleading statements in his original white paper. His rationale for the 11% direct/89% reflective design based on his measurements at Boston Symphony Hall is absurd. First of all, the ratio is different for every location closer or further back from the stage, even according to his own graphs. At 19 feet, this is not necessarily the best seat in the house. At Carnegie hall, the critics who get the best seats sit in row K, somewhat further back. And the percentage will depend on where on the stage the musician is located. The fact that 89% of the radiated sound from 901 is reflected doesn't mean that what you hear is 89% due to reflections. This depends on room acoustics and placement. But most important and most obvious error is that the nature of the relationship between the direct and reflected sound produced by 901 in a home environment has virtually nothing in common with the corresponding relationships of live musicians in concert halls.

Still, with its shortcomings, it was unusual for someone to dare to market a product which was so innovative that it flew in the face of what everybody else in the industry belived. Frankly I take it for what it is, neither totally condemning it nor completely ecstatic about it. I think I will move mine soon into another room where it may perform much better than the room it's in now. I'll probably play around with it till the end of my days.

Link to comment
Share on other sites

There is a separate thread on the AR90s and what I did to the crossovers, but I'm not against repeating myself.

In a nutshell, I built new crossovers (upper and lower boards) using Solen poly caps and 15 gauge perfect lay coils except for the inductor in the tweeter section where I used an Alpha Core 15 gauge foil inductor. All internal wire was replaced with Kimber 15 gauge high purity copper. I removed the attenuator switch assembly to shorten and simplify the circuit given I've never used it in 20 plus years. The new inductors were located in the same orientation as the old inductors.

Link to comment
Share on other sites

"Well, here is the problem with your reasoning: the definition of high-fidelity sound is "the accurate or faithful recreation of the original sound source."

This says an awful lot. And it gives the thoughtful audiophile who demands perfection good reason to dispair. There must have come a time soon after the novelty of being able to make recordings at all when the incredible shortcomings of the early technology making wax cylinders for Edison and his competitors gave the equivalent of today's audio engineers drive to do much better. These people had the opportunity to hear the greatest musical artists of the day perform the greatest music ever written in both recording studios and in live performances before audiences. And when they listened to the results of their efforts, any objective comparison must have lead them to the conclusion that their results were appalling. Over a hundred years later, for all of our advanced technology, modern engineers who perform the same function must come to exactly the same conclusion. For while it is possible to successfully set up carefully contrived live versus recorded experiments and demonstrations as AR did and as others have done in the past, with real recordings you can buy when played on ANY contemporary sound system, the auditory experience of a live concert cannot yet be successfully duplicated by a machine. On a scale of 0 to 100 where 0 represents the abilitiy of even the casual listener with normal hearing to instantily identify blindfolded that he is hearing a recording and not a live performance and 100 represents experienced concert goers with excellent hearing being fooled most of the time, with all of our progress the best our technology can do today is still at or pretty close to zero. That is not to say that someday when far more is known and the best of these current systems are antique curiousities, that new technology won't emerge that will be radically different in concept and results. However, the continued refinement and perfection of sound systems within the two channel paradyme we have lived with since the 1940s when Walt Disney Studios invented it for theaters and the 1950s when Wollensak and Westrex brought it into our homes is a dead end.

Link to comment
Share on other sites

Of course my main intention is to quantify if there are gross anomalies in the components of the crossover circuits and therefore a real need to change the capacitors. If I modify my approach, and play the frequencies from my computer through my amplifier to the "speakers" with a 4 ohm test load, would that be sufficient to identify any anomalies in the crossover components?

I assume I could prepare a theoretical frequency response curve, in Excel, for the crossovers feeding a 4 Ohm resistor for comparison. (If the original frequency response curves of the crossovers ONLY, were available, and the test load used stated, a direct comparison with the original could be made). The CD I'm using is the Denon Audio Technical CD which they state is produced from a "high precision digital 16bit signal generator controlled by a computer and recorded in digital form onto the compact disc". Of course I'm not trying to measure the sound pressure levels in a room, rather, if the crossovers are working close or near to specifications I'll leave the capacitors alone and consider the other improvements mentioned elsewhere in the forum.


Link to comment
Share on other sites

Unfortunately, your model using a 4 ohm resister won't really work. The resister can not replicate the capacitive and reactive loading that the voice coil inflicts into the circuit with AC applied compounded with the simple resistive loading it represents to a simple DC signal.

Link to comment
Share on other sites

I hear what you say about the speaker's reactive load on the crossover, but does that matter for the trial? If I measured the crossover points for the 4 Ohm load and they matched those calculated from the Excel sheet (Using forum cct schematics), wouldn't that confirm the crossover capacitors were working correctly for a 4 Ohm load? Then couldn't I assume that the addition of the speakers into the crossover network would revert back to the speaker crossover frequency specifications, No?

Link to comment
Share on other sites


This topic is now archived and is closed to further replies.

  • Create New...