Musings on the AR-2ax crossover

[Mach3]

So recently was able to secure a very nice set of AR-2ax off the 'bay. Local guy and he delivered to my door in their original boxes. The grills had never been removed. Very nice cabinets with hardly a flaw. These were made in June, '73.

Of course the woofer foam surrounds were gone with the wind - replaced them easy enough and while in there I decided to "rebuild" the crossover which consisted of only two capacitors (4 uF and a 6 uF). Used Solen precision caps (+/- 1% tolerance) as I believe that a capacitor with a tighter specified tolerance is in fact a "better" capacitor - such means that the production processes being employed are "in control" and hence the product is more regular. I don't listen to various capacitors - but I do believe that AR speakers in general respond with considerable sonic improvement when the old electrolytics are replaced with modern film types (polystyrene, polypropylene, metallized films et cetera).

While doing this work I began to contemplate how this crossover worked - and in fact it is surprisingly crude (or sophisticated depending on your perspective). Both the mid and the tweeter are simply hi-passed - which means they receive ALL frequencies above their crossover point (1400 Hz for the mid and 5k Hz for the tweeter). The woofer is low-passed with a coil. Simplicity - not even a first order crossover - just two high-passes and a low pass in parallel.

This means the designer was relying on the natural electrical and mechanical roll-off of the mid range unit to cut off its response above 5 kHz. - I also suspect that this is the reason the mid-range on the AR-2ax has a thick section of felt on top of it (or is it fiberglass padding?). This tends to suppress break up sounds when the mid is being sent frequencies that are beyond its ability to play.

How does it sound? Exactly as I remember it - good clarity, detailed, very rich sounding, with a warm presence and with excellent bass per the AR 10" woofer. I am playing it with a Marantz 2252 receiver - and the combination makes one wonder if we in fact have made ANY progress in regards stereo gear in the intervening 39 years.

I have heard many $50k to $100k systems - and while they are possessed of supernatural clarity - almost create an x-ray image of the music - they are far less enjoyable than this $700 rig.

What gives with this?

BTW: Many of the "uber" speakers out there right now - the ones with the incredible price tags - employ a similar crossover/driver technique - such a minimalist approach to crossover design is considered very "moderne". Of course AR was doing this in what? 66? or so?

[genek]

Your assessment of the AR design is correct. The tonal balance is indeed based almost entirely on the characteristics of the drivers. The crossover's function is to protect the drivers from frequencies outside their working ranges that would result in undesirable distortions or driver damage, and to do a relatively minor amount of curve shaping,

Also, the history of the AR-2ax is that the AR-2 started out as a 2-way design, and got a dome tweeter added to it to become the 2a before it evolved into the 2ax.

[Steve F]

As proof of assertion that the 3 1/2" mid ran full range, see the attached 2ax FR curve of then individual drivers, from AR lit in 1974. The mid runs into the mid-teens kHz.

The tweeter can be seen to be well down in level, typical of pre-ferrofluid AR tweeters: well dispersed but low in FR level, hence that "classic" AR sound.

Steve F.

[Carlspeak]

DIY speaker builders, who can afford them, look for modern drivers with inherently flat response and intentionally design minimalist xo's.

http://techtalk.parts-express.com/showthread.php?t=230688&highlight=intimates

[tysontom]

So recently was able to secure a very nice set of AR-2ax off the 'bay. Local guy and he delivered to my door in their original boxes. The grills had never been removed. Very nice cabinets with hardly a flaw. These were made in June, '73.

While doing this work I began to contemplate how this crossover worked - and in fact it is surprisingly crude (or sophisticated depending on your perspective). Both the mid and the tweeter are simply hi-passed - which means they receive ALL frequencies above their crossover point (1400 Hz for the mid and 5k Hz for the tweeter). The woofer is low-passed with a coil. Simplicity - not even a first order crossover - just two high-passes and a low pass in parallel.

This means the designer was relying on the natural electrical and mechanical roll-off of the mid range unit to cut off its response above 5 kHz. - I also suspect that this is the reason the mid-range on the AR-2ax has a thick section of felt on top of it (or is it fiberglass padding?). This tends to suppress break up sounds when the mid is being sent frequencies that are beyond its ability to play.

I have heard many $50k to $100k systems - and while they are possessed of supernatural clarity - almost create an x-ray image of the music - they are far less enjoyable than this $700 rig.

What gives with this?

What's wrong with this picture? Mach3 perhaps means well but is stating the obvious, that the AR-2ax crossover is simple—intentionally so—but this should not imply that it is "surprisingly crude." Damned with faint praise? In my view, Mach3's notion is oversimplified and naïve. The 2ax crossover is simple because the drivers perform well within their passband and don't require crossover equalization, and the speaker has proven itself time and again—probably far beyond any known $50k or $100k loudspeaker—as being an established "standard" of accurate-sound reproduction in the home. Name just one $100k loudspeaker that is considered to be an "accuracy standard" in the industry; in magazines yes, in the industry, no. There aren't any. By the way, most mega-buck loudspeakers are rarely designed in the laboratory but usually in someone's garage using a dose of black magic here and there. That "supernatural clarity" is something that you would never hear in live music. Perhaps Mach3 has been reading too many Stereophile or TAS magazines.

Incidentally, the fiberglass on top of the midrange is for damping purposes, not "…to suppress break up sounds when the mid is being sent frequencies that are beyond its ability to play." The driver doesn't need the fiberglass except to smooth the output on- and off-axis. Each driver in the AR-2ax was designed to perform accurately in the anechoic chamber to achieve flat on- and off-axis output, and the fiberglass pad was part of that design: to take an inexpensive, production CTS 3½-inch cone driver and make it perform up to standards expected of the speaker.

This is "what gives" with this.

--Tom Tyson

[ra.ra]

This topic is timely, for me at least - - - I am just about to re-assemble my recently re-capped and re-foamed AR-2ax's. On this pair of speakers, I decided to simple remove the original wax block capacitor which was only held in place with a few staples, and added a single #5 insulated staple to keep the blue wire firmly in place (see attached pic).

I, too, used Solen capacitors, but I am only familiar with the more common "fast-caps" with a 5% tolerance rated at 400V - - - I am unfamilar with the 1% tolerance variety mentioned by Mach3. Even though I am quite confident these will perform admirably, I harbor a minor twinge of regret that instead of employing the original spec 4.0 and 6.0 mf caps (e.g. from Madisound), I used the 3.9 and 6.2 mf values available from P/E. Alas, one more bit of web research may have pre-empted my impulsive purchase.

I removed the pots to clean them (very little corrosion this time) and I noticed one thing that surprised me a bit - - - it appeared to me that the wires connecting the pots to the capacitor(s) were of a sligthly heavier gage than the wires connecting the pots to the drivers. In the pic, this can best be seen in the green wires for the mid-range driver.

(to Steve F: not sure if your attachment was included with your post.)

[Carlspeak]

This topic is timely, for me at least - - - I am just about to re-assemble my recently re-capped and re-foamed AR-2ax's. On this pair of speakers, I decided to simple remove the original wax block capacitor which was only held in place with a few staples, and added a single #5 insulated staple to keep the blue wire firmly in place (see attached pic).

I, too, used Solen capacitors, .......

Let us know how you think they sound after you get them back together. Solens get beat up on sometimes in audio forums.

[Steve F]

2ax FR attached.

Steve F.2ax FR.pdf

[dynaco_dan]

Hi Steve

I know the 3 1/2" mid/tweeter was sourced from an outside supplier for all of it's life.

Early on there was just the screen cover and later the fiberglass cover was added.

There may have been an even earliest time period without the screen cover.

The AR-4 used this as their mid/tweeter driver with many notible reviews.

The earliest AR-2X speaker also used this driver.

To this day I never read a review of the AR-1X or AR-2X speaker system.

It was my intention to add one AR-1X center fill speaker to a pair of AR-4X's, it never happened though.

I mistakenly wrote elsewhere here that the earliest AR-1X used this same driver.

I was corrected by Tom Tyson, thank you.

I was the one that had previously posted that hand wrtten conversion note from the 8" to 3 1/2" instead of the inhouse made 2 1/2" driver..

[Carlspeak]

This thread I created last fall is an interesting sidelight to the AR2ax story

http://www.classicspeakerpages.net/IP.Board/index.php?showtopic=6705&hl=ar2ax&fromsearch=1

[genek]

What's wrong with this picture? Mach3 perhaps means well but is stating the obvious, that the AR-2ax crossover is simple—intentionally so—but this should not imply that it is "surprisingly crude." Damned with faint praise? In my view, Mach3's notion is oversimplified and naïve.

I think this just reflects the technical misconception that more complex circuits with more components must be better and more sophisticated. Remember when they used to actually put the number of transistors on the case of a radio or amplifier?

[RoyC]

I, too, used Solen capacitors, but I am only familiar with the more common "fast-caps" with a 5% tolerance rated at 400V - - - I am unfamilar with the 1% tolerance variety mentioned by Mach3. Even though I am quite confident these will perform admirably, I harbor a minor twinge of regret that instead of employing the original spec 4.0 and 6.0 mf caps (e.g. from Madisound), I used the 3.9 and 6.2 mf values available from P/E. Alas, one more bit of web research may have pre-empted my impulsive purchase.

I removed the pots to clean them (very little corrosion this time) and I noticed one thing that surprised me a bit - - - it appeared to me that the wires connecting the pots to the capacitor(s) were of a sligthly heavier gage than the wires connecting the pots to the drivers. In the pic, this can best be seen in the green wires for the mid-range driver.

I wouldn't give 1% caps a second thought, or worry about the performance of your new 3.9uf and 6.2uf replacements. I've measured many caps, and modern film caps seldom vary by more than 1% to 3%, if at all, from stated value. Even if they do, remember the original caps were +/-10%. The 1% caps are not made any better or differently. It just means some effort was made to measure, and find those that vary by no more than 1% from the intended value.

Another thing to remember is that the performance of those old drivers and pots are likely to vary by at least that much.

The heavier wires were original to the capacitors, which were manufactured that way and delivered to AR. The thinner ones were AR's crossover wires, which was typically tinned, 20ga.

Roy

[soundminded]

When used within the tweeter's power handling capabilities AR2ax can be equalized to exceptionally flat FR over the entire audible range. Also it's high frequency dispersion is excellent, better than AR9. Unfortunately you can't replace the stamped frame woofers with the earlier cast frame woofers directly, the cabinet opening isn't routed to accept them. For some reason the early cloth surround cast frame 10" and 12" alnico magnet AR woofers sound better to me than the later versions, their bass quality is different. I'm not sure why. Perhaps their distortion is lower?

The use of very complex crossover networks in modern speaker systems to passively equalize them, their narrow dispersion tweeters that beam and are shrill, and the ported designs to me add up to a product that is inferior in performance and far more expensive than they are worth. For example, among other characteristics, substantial resonant peaks in kevlar and aluminum cone drivers must be filtered out. The superiority of active equalization at the preamplifier signal level in controlling system FR is striking. It is inexpensive and very effective which probably esplains why the industry doesn't like it, it cuts into profits. This was not available to consumers at the time these classic speakers were manufactured and so their full potential for accurate tonal balance playing commercially made recordigs was never fully realized. When restored and used in systems that take advantage of modern electronics, these speakers never sounded better than they can today.

[Steve F]

As I’ve opined before on these pages, we—as a hobbyist community—have a tendency to compress the historical timeline into the same presence, such that we think that designers’ knowledge of the state of the art is a constant throughout the decades.

It is not, of course.

The awareness and knowledge of crossover design in the 60’s was different and more limited than that knowledge and awareness is today, some 50 years later. To think that the audio industry—or indeed, any industry—hasn’t progressed in its collective understanding of its art is incorrect.

Today’s best speakers can and often do accomplish more with their crossovers than did the rudimentary crossovers of a half century ago. That’s not to say that everything today is automatically better, because as we all can attest, that is certainly not the case. But greater understanding exists today, which makes better results possible. Case in point: Speaker Dave’s re-do of the crude 4x original crossover, using the 4x’s original drivers. Result: A far superior response and better overall performance, without a tremendous increase in crossover cost or component count.

This subject does engage a lot of emotional-defensive reaction, as many of us believe that the old AR of 1954-1974 “walked on water,” and could do no wrong. Near perfection in design, near perfection in the inevitable compromises chosen, and near perfection in the results for a given price/materials cost. While the Classic ARs have certainly stood the test of time better than perhaps any speakers from that time period, they weren’t perfect, and neither was Villchur’s or Allison’s understanding of every single possible technical/engineering design point. Excellent? Yes. Far ahead of their time? Yes. Perfect for the ages, never to be improved upon, ever?

You answer that.

As an aside to Soundminded’s excellent point about active vs. passive EQ: This is quite true, of course. The better way to do it is actively. Fortunately, something like 90% of the speakers sold today are part of a ‘closed’ system, where the speakers (or headphones/buds) are fed an EQ’d signal from the captive source unit/amplifier. If—and it’s one whale of a huge if—the source unit/amplifier is EQ’d properly to take the companion speaker’s characteristics into account, then these systems can achieve an excellent final result.

Whether they do or not in reality is a function of the designer’s skill and the marketing department’s understanding and direction as to whether that matters to the end customer or not.

Steve F.

[Gerry S]

I've done LOT's crossover circuits in the 12 year period when voicing "bread and butter" designs. What I can say without question is this. The crossover designer's job is made MUCH easier IF the drivers for a particular design is really thoroughly developed to the "nth degree" before it's given to the crossover designer.

Even then, it's no easy task for the the crossover designer because even the best drivers will occasionally have relatively large variations from the reference driver. This can be minimized by doing a trial production run to determine the "average" frequency of what's possible for a given driver. This is true for ALL drivers (woofers, mid-range, tweeters).

When matching two or more drivers at the "pass-band" or "linear piston range" , one must take into account what happens OUTSIDE the driver's ideal operating range. If the drivers are "sloppy' or unpredictable at their extremes, then a more complicated network must be employed to insure best possible unit-to-unit uniformity when the finished product comes off the manufacturing line.

Sometimes a "simple" 1st order isn't enough for the system to pass because of what happens at the extremes where the drivers "blend". That's why the final crossover slope and frequency may be higher or lower than what's actually "stated". Thus the +/- "window" you see in specification sheets.

When speaking of "crossover frequency", there is the " electrical crossover point " of the circuit, as well as the natural "acoustical" roll-off of the drivers themselves. There is also the tolerance of the crossover components themselves. Close tolerance parts must cost more than high tolerance ones. But close-tolerance crossover components by themselves will not automatically yield a superior product if the drivers themselves are not made to tight tolerances.

The speaker system that is commercially successful (profitable) is that which can be made consistently in numbers, and has the sound and reliability expected by both the consumer and the designer. Given the inherently high potential of "built -up of tolerances" in manufacturing transducers, it's remarkable that the very early AR's were remarkably similar from one system to another.

"Sound quality" issues aside, simple "first order" (or no crossover) will in most cases provide a speaker that's either very acoustically different from unit to unit, and/or prone to early failure.

[genek]

Even then, it's no easy task for the the crossover designer because even the best drivers will occasionally have relatively large variations from the reference driver. This can be minimized by doing a trial production run to determine the "average" frequency of what's possible for a given driver. This is true for ALL drivers (woofers, mid-range, tweeters).

AR's approach to this in the classic era was a bit different. They scrapped about a third of all the drivers they tested because they varied excesively from the reference driver.

[Gerry S]

AR's approach to this in the classic era was a bit different. They scrapped about a third of all the drivers they tested because they varied excesively from the reference driver.

Maybe that's why "true classics" of that era generally sound "better" to readers of this forum (experienced listeners) when A/B 'ing a "similar" product made today by the same "brand". I'm guessing virtually no "quality control" unless a particular driver is obviously distorting or dead.

Today's buyers (an ultra-small number compaired to the 70's -80's ) would probably plop them where convienient and play mp3 stuff. If one channel sounded significantly different than the other, they probably wouldn't notice. And if they did notice, they probably wouldn't care. For these consumers, if they LOOK identical, then the speakers are a "matched stereo pair".

[speaker dave]

2ax FR attached.

Steve F.2ax FR.pdf

That is pretty much what I saw when I measured my pair a couple of years ago. The 2ax is essentially a 2 way design with an added dome tweeter. The level of the dome is low enough that it adds vey little on axis but perhaps adds some extra to the power response and some off axis sparkle (even there you can see that the power response drops at the tweeters frequency range from the low sensitivity.

I would comment on the evolution of crossover design since that era but perhaps I shouldn't!

David

[genek]

I would comment on the evolution of crossover design since that era but perhaps I shouldn't!

Come on, Dave, there must be a Mods and Tweaks 3-way crossover project in there somewhere. There's at least as many 2ax's still out in the world as 4x's.

[dynaco_dan]

That is pretty much what I saw when I measured my pair a couple of years ago. The 2ax is essentially a 2 way design with an added dome tweeter. The level of the dome is low enough that it adds vey little on axis but perhaps adds some extra to the power response and some off axis sparkle (even there you can see that the power response drops at the tweeters frequency range from the low sensitivity.

I would comment on the evolution of crossover design since that era but perhaps I shouldn't!

David

Hi David

By all means comment on evolution of crossover design, that is how we all learn, my friend.

With all of the different speakers I own, either buying new or rescuing I never owned a pair, until a few years ago at least, of AR-2AX's that I listened to in my home, yet.

Now that I looked at the posted 2AX speaker response I noticed something that I had overlooked before, it wasn't important to me, then.

The on axis photo only shows a relatively flat response out to about 10,000 hz from 500 hz with the mid driver actually rising in it's overall output, relative to the woofers output.

I feel that the woofer and mid driver would have a very close output, up to where they actually crossover and the midrange would be very slightly brighter or have more output at least.

The dome tweeter, 3/4" as listed, has a very substantially flat output, but a very much reduced output level, of which does not seem logical to me at least.

My permanent impaired hearing is generally from gut shaking AR-3A / AR-LST bass, out to maybe 2,000 - 2,500 hz unaided.

After the 2,000 - 2,500 hz crossover my hearing generally drops 6 1/2 dB out to maybe ? perhaps 7 - 8,000 hz., not precisely evaluated, as they cannot improve much higher.

It would appear that I will definitely not hear any or much of this 3/4" tweeter output any time soon.

I don't have any AR-2AX reviews with and without the 1 3/8" tweeter.

Can anyone help me out here, please?

I seem to remember reading, somewhere, long long ago, that there was an audible and improved difference in the different 2AX's, particularly the later version.

There must be reviews somewhere of even the different woofers they used, alnico, transition adapted steel frame and steel frame.

There is a proven history of the consumer and reviewers tremendous appreciation of the AR-2AX in both it's earliest and especially the later form.

I am curious now as to what the on axis photo would look like using the 1 3/8" tweeter instead.

This is all about the on axis photo only, not the combined power output, which combines from all angles, with cabinet reflection included, not just on axis.

Dispersion is where the AR speakers generally excel.

[tysontom]

As I’ve opined before on these pages, we—as a hobbyist community—have a tendency to compress the historical timeline into the same presence, such that we think that designers’ knowledge of the state of the art is a constant throughout the decades.

It is not, of course.

The awareness and knowledge of crossover design in the 60’s was different and more limited than that knowledge and awareness is today, some 50 years later. To think that the audio industry—or indeed, any industry—hasn’t progressed in its collective understanding of its art is incorrect.

Today’s best speakers can and often do accomplish more with their crossovers than did the rudimentary crossovers of a half century ago. That’s not to say that everything today is automatically better, because as we all can attest, that is certainly not the case. But greater understanding exists today, which makes better results possible. Case in point: Speaker Dave’s re-do of the crude 4x original crossover, using the 4x’s original drivers. Result: A far superior response and better overall performance, without a tremendous increase in crossover cost or component count.

This subject does engage a lot of emotional-defensive reaction, as many of us believe that the old AR of 1954-1974 “walked on water,” and could do no wrong. Near perfection in design, near perfection in the inevitable compromises chosen, and near perfection in the results for a given price/materials cost. While the Classic ARs have certainly stood the test of time better than perhaps any speakers from that time period, they weren’t perfect, and neither was Villchur’s or Allison’s understanding of every single possible technical/engineering design point. Excellent? Yes. Far ahead of their time? Yes. Perfect for the ages, never to be improved upon, ever?

You answer that.

As an aside to Soundminded’s excellent point about active vs. passive EQ: This is quite true, of course. The better way to do it is actively. Fortunately, something like 90% of the speakers sold today are part of a ‘closed’ system, where the speakers (or headphones/buds) are fed an EQ’d signal from the captive source unit/amplifier. If—and it’s one whale of a huge if—the source unit/amplifier is EQ’d properly to take the companion speaker’s characteristics into account, then these systems can achieve an excellent final result.

Whether they do or not in reality is a function of the designer’s skill and the marketing department’s understanding and direction as to whether that matters to the end customer or not.

Steve F.

This message was somewhat of a lecture, but it also misses the point I was trying to make.

My original message regarding Mach3's comments about the "surprisingly crude" AR-2ax crossover was not meant to defend AR speakers because we are simply emotionally attached to them or that the old Acoustic Research "walked on water." It was neither an emotional-defense of AR, nor was it meant to imply that progress in loudspeaker development stopped with the Ed Villchur/Roy Allison designs of the 1950s and 1960s. Improvements and enhancements over time are completely self-evident! And insofar as we are attempting to recreate—not create—sound, the improvements should be in science, not art.

My point was that blatant criticism of a design—without really understanding what the design was intended to accomplish—is naïve and pointless—even if no harm was meant. Saying the 2ax crossover was "surprisingly crude" or that midrange had fiberglass "to suppress break up sounds when the mid is being sent frequencies that are beyond its ability to play, are flip comments, pure and simple. Mach3 was implying that the crossover was under-designed, or purpose of the fiberglass was to keep the cone from perhaps self-destructing or distorting when fed signals out of its passband. This was criticism for the sake of example, a hindsight perspective—and yes—I will defend AR's old design in this regard. Of course the AR-2ax and the AR-4x designs were simple in concept and intentionally so. They were designed for a certain cost point to perform well in an average listening room. It wasn't that crossover design wasn't understood—there were many well-understood, complex crossover designs dating back many years—but the purpose of these AR designs was smooth acoustical response in the average listening room. Nothing more, and it worked well.

--Tom Tyson

[soundminded]

"The 2ax is essentially a 2 way design with an added dome tweeter."

You could make the same derogatory comment about any 3 way design, it's a two way design with a tweeter added. The inefficiency of the tweeter even unattenuated compared to the other drivers made the speaker sound muffled, dull. Fortunately today this is easily overcome in any of several ways. The stuffing in front of the midrange (seemed more likely lamb's wool to me than fiberglass) mechanically filters its hf output. I noticed a small rise in the midrange possibly due to the crossover filter F3s being slightly too close together and this was easily filtered out too. The deep bass can also be improved with equalization although my experience with it is that the 10" woofer isn't nearly so robust as the 12" model. When optimized AR2ax IMO has a wide flat FR, outstanding dispersion, no harshness or shrillness, much less of a "sweet spot" than modern speakers. and can play at reasonably loud level without distorting although it's not usable in single pairs for very loud sound or large room. It's also easy to drive presenting no difficulty to almost any amplifier. Widely available, inexpensive, and easily restored if the tweeters are in good condition AR2ax is a winner. A much better choice than far more expensive audiophile speakers today that cost in the thousands.

[RoyC]

Obviously, the earliest AR philosphy was to physically/mechanically design, and consistently manufacture, speaker systems to meet design goals with minimal electrical ("passive crossover") modification. In an era of complex, computer driven theoretical crossover formulas mated to a multitude of assorted drivers, it is easy to overlook the effort and expertise it took to achieve the performance of systems such as the 2ax. It is also easy to underestimate the real-world performance of these old models in a typical residence. The degree of practical measurement, testing, and quality control on the part of AR was likely exhaustive by any standard of any era.

AR produced visually unrefined looking, finely tuned drivers. Today, these are often naively referred to as "crude" by audio hobbyists, who may otherwise believe in dubious refinements such as exotic cables, line voltage conditioners, heavy gauge speaker wire, overpriced capacitors, etc....not to mention flashy looking drivers constructed of all kinds of materials.

Fiberglass was the most durable of AR's modification materials, and foam was the least. Fiberglass is found under and over cones and domes of a number of drivers (btw, Soundminded, the material under the 2ax mid grille is simply compressed fiberglass). Early urethane foam is found in tweeter suspensions and under the 3/4" paper domes, and wooden plugs are installed under later 3/4" cloth domes. Some of AR's proprietary drivers had silicone grease in voice coil gaps, and some were among the first to contain ferro-fluid. The AR woofer cones were very thick, soft, and often modified with rubber compounds and foam damping rings. The point here is nothing about the quality of these old beasts can be determined by the appearance of the drivers, or complexity, or lack thereof, of the crossovers.

Roy

[speaker dave]

"The 2ax is essentially a 2 way design with an added dome tweeter."

You could make the same derogatory comment about any 3 way design, it's a two way design with a tweeter added. The inefficiency of the tweeter even unattenuated compared to the other drivers made the speaker sound muffled, dull. Fortunately today this is easily overcome in any of several ways. The stuffing in front of the midrange (seemed more likely lamb's wool to me than fiberglass) mechanically filters its hf output. I noticed a small rise in the midrange possibly due to the crossover filter F3s being slightly too close together and this was easily filtered out too. .....

http://www.classicsp...?showtopic=5282

Here is the thread where I took a number of measurements of my pair of later AR2ax.

In the first post there is a curve family showing the range of curves from the midrange control (tweeter at max). This duplicates what Steve's factory curves show, that the "midrange" is really quite wide range and is enough more sensitive than the dome tweeter to really be considered the primary tweeter itself.

I prefer listening at the plum middle curve setting and at that point the dome tweeter is, on average about 8dB quieter than the mid (over the range the dome operates in). I had to listen quite carefully initially to convince myself that I could hear the dome at all (dome set to max).

Since the "mid" operates from about 1500 to 15,000 Hz I would consider this a "2 way design with dome tweeter added". This is not meant as a slight but more as a description of the general design. No attempt is made to roll off above the low treble so that each driver works over its own range only, as with the typical 3-way. The turn on point of the tweeter and natural rolloff of the mid overlap by a considerable amount and I don't remember any electrical rolloff of the mid unit. I'm sure the fiberglass on the mid flattens the response by attenuating the upper range, acting more as a shelving device.

With the significant level difference between the mid and the tweeter, it isn't clear that the dome adds much to the top Octave power response. Again, this is seen in Steve's curves where the axial differences between mid and tweeter output are pretty well duplicated in the power response curve.

Regards,

David S.

[Carlspeak]

http://www.classicsp...?showtopic=5282

Here is the thread where I took a number of measurements of my pair of later AR2ax.

In the first post there is a curve family showing the range of curves from the midrange control (tweeter at max). This duplicates what Steve's factory curves show, that the "midrange" is really quite wide range and is enough more sensitive than the dome tweeter to really be considered the primary tweeter itself.

I prefer listening at the plum middle curve setting and at that point the dome tweeter is, on average about 8dB quieter than the mid (over the range the dome operates in). I had to listen quite carefully initially to convince myself that I could hear the dome at all (dome set to max).

Since the "mid" operates from about 1500 to 15,000 Hz I would consider this a "2 way design with dome tweeter added". .........

Regards,

David S.

I came to essentially the same conclusion at the link back at post #10 of this thread.