Dome mids vs. cones--meaningful difference?

[Steve F]

Where has the dome midrange gone?

 

 

When AR introduced the domed drivers in the AR-3, things changed forever in the speaker world. The dome tweeter is still with us, some 59 years later, going reasonably strong. (Although, in all honesty, it does seem like the Heil/ribbon tweeter is gaining fast on the dome tweeter in recent new designs.)

 

 

But the dome midrange enjoyed a far briefer time in the spotlight. There are still a few dome mids here and there from random companies, but the mainstream, right-down-the-middle-of-the-road dome mid is pretty much gone.

 

 

The AR 1 ½-inch dome mid had a pretty long run, from around 1967 in the 3a until the mid-80’s with the 9/98/78 LS/LSi’s.

 

 

ADS utilized their own excellent dome mid throughout this approximate time period as well, and it was a suburb driver in its various guises.

 

 

The Allison 3-ways could be considered as having a “dome” mid. Yes, we all know it was not peripherally-driven like a conventional dome, but it looked dome-ish, so we’ll cut it some slack and include it.

 

 

The Avid 330 12-inch 3-way of the mid-1970’s used a nice dome mid, 2-inch IIRC. That was a nice-sounding speaker.

 

 

But AR’s big competitors at that time—Advent and EPI—were mostly 2-ways, and JBL used cone mids. Likewise for Boston Acoustics—mostly 2-ways and cones when they were 3-ways.

 

 

I worked at BA for 11 years. Andy [Kotsatos] knew I liked AR speakers, but he never knew to what extent, nor did he know that I was a member of this Forum during that time. (And the Forum didn’t know I worked at BA then, either—I was very careful to keep my posts totally free from conflict-of-interest suspicion and complication.)

 

 

I did sneak in a few questions to Andy during my BA tenure about his recollections of AR, since he was the main product manager at Advent before co-founding BA, so he knew first-hand about AR Classics and ADDs from a competitor’s standpoint. (BTW, he generally hated them, thought the 3a was ‘lousy,’ the 4x ‘totally muffled,’ but said the 2ax could be a ‘sneaky’ speaker if you weren’t careful.)

 

 

Anyway, he always said that the concept of a dome midrange was a ‘fraud’ (his exact word—I’ll never forget it). He felt that they were inefficient, yet also had a shorter excursion than a good 5-inch mid, so the dome couldn’t play as loud. “It can’t handle the same power, it can’t go as loud and yet it needs more power than a cone. Why bother? The “dispersion” advantage is simply untrue, because in the frequency regions where you’re going to use a mid in a well-designed 3-way, neither a cone nor a dome midrange is directional in any meaningful way over its operating band.  Do the math. The 3a’s mid ‘clips’—you can hear it running out of excursion at high power levels. It’s a joke.

 

 

But it looks cool with that screen and AR played the ‘technical dispersion angle’ to the hilt with lots of graphs and curves and the reviewers bought into it. All fine on paper, except that they didn’t sound that good.”

 

 

Was he right? Is the superior dispersion of a dome mid vs. a 4- or 5-inch cone not particularly important from 300-4000Hz? AR did blow a lot of mids (both my cousins blew their 3a mids playing jazz, not hard rock, in the late-60’s/early-70’s). BA never replaced blown mids under warranty, and I mean never.

 

 

I’m not talking about voicing. I always fought w Andy when I voiced a BA speaker, because I like the AR ADD balance and he liked a more forward mid-rangy sound. We’d go back and forth, mostly on woofer choke values.

 

 

I want to keep this discussion to the importance of mid-dispersion for forward-facing single-driver-per-band designs, and whether there is truly an audible/meaningful difference between a 1 ½-inch dome and a 4- or 5-inch cone when the listener is sitting in a normal position (less than 10 or 15 degrees off axis from the left and right speakers of a stereo pair).

 

 

Thoughts?

 

 

Steve F.

[Carlspeak]

I sold a number of AR3/AR3a Super-mod kits some years back. The kit basically had new drivers to replace original dome tweeter and dome midrange. The kit's midrange driver was a Seas H1262. It was a perfect drop in fit and came with a pre-fab'd chamber that was installed on the backside of the baffle board because the Seas mid did not have a closed back.  

All of the kits had positive feedback. The owners were satisfied with the sound. I did some dedicated measurements comparing the original dome mid with the Seas Mid. along with the tweeter and woofer; each done separately. They can be found on pg. 6 of my Super-Mod development document. The midrange curves show the Seas driver with a flatter response thru its pass band than the original dome mid. That's one test of one vintage dome mid and doesn't indicate a trend. 

 

 

AR3a_Super_Mod_development.pdf

[genek]

I have 3a's and 2ax's. Can't say that I notice any difference in soundstage between them.

 

[michiganpat]

if there is a knock on midrange domes, it's that they can't go as low as a 3 or 4" cone mid-- the xover point of the AR  mid is somewhere in the 800hz range, IIRC.  I wouldn't be surprised if there is an audible improvement with a dome mid (A/B an AR2ax and an AR5, or an AR58S and a 58B to compare?) especially off axis....

 

[genek]

The 3a crosses from woofer to mid at 525 Hz.

[AR surround]

The dome UMR in the AR9 and AR90 operates in the 1200Hz to 7000Hz range.   However, something like the Sonus Second Mortgage Faber Olympica III has a 2500Hz crossover point and uses the ubiquitous cone midrange found on today's speakers.   I would imagine that dispersion is important in the upper operating frequency of the AR9 series UMR and thus benefits from utilizing a dome.  The big 8" cone LMR also takes a large part of the midrange load off of the UMR.

In my opinion, the UMR dome definitely plays LOUD, so much so that I have always attenuated it by 3dB.  Perhaps AR over-compensated for it a bit after moving beyond the Classic series. 

[ligs]

Just for comparison, here are some frequency response curves for a modern 2-way speaker system using 5" woofer and 1" dome tweeter. The thing caught my eye was the even dispersion of the 1" wave guided tweeter,

http://averagejoeaudiophile.blogspot.com/2015/12/infinity-reference-r152-infinitys-high.html

 

How would AR dome compare to a cone in the midrange?

[Carlspeak]

2 hours ago, ligs said:

 

 

How would AR dome compare to a cone in the midrange?

Take a look at the link noted in my first post in this thread. It's not a true apples to apples comparison because the Seas mid was new and the AR dome mid I tested was old.

However, there are AR generated on-axis curves for individual drivers when they were new that can be viewed in this forum with a little searching. IIRC, the midrange curve is as flat as the Seas curve.

A better telling set of measurements to satisfy the OP's question would be polars for the dome mid vs a 5 inch cone mid. I don't have those and don't think anybody else does either. 

 

[ligs]

http://www.classicspeakerpages.net/library/acoustic_research/special_sections/additional_ar_documents/the_sound_field_in_home_lis/

 

AR3a on-axis vs off axis responses are in the attachment. 

[Aadams]

I want to keep this discussion to the importance of mid-dispersion for forward-facing single-driver-per-band designs, and whether there is truly an audible/meaningful difference between a 1 ½-inch dome and a 4- or 5-inch cone when the listener is sitting in a normal position (less than 10 or 15 degrees off axis from the left and right speakers of a stereo pair).

I am just a long time hobbyist with subjective opinions keeping in mind the original question and constraints.  In my experience, for sweet spot listening there is probably no practical difference.   The 58s has a dome midrange vs 58b cone midrange with a similar low crossover point, maybe there is comparative data on those.   AR 5 vs AR2ax fits the midrange physical requirement but the 2ax woofer carries all of the LMR load and for that reason IMO the AR5 always sounds better at my house.   

Using an equalizer I have made full range Advents and Full range EPIs sound very close to AR 3a in direct AB test but I prefer the sound of the AR.  I have made Time Windows sound shockingly similar to 3as with all of their spaciousness but I used a subwoofer and the DCM has two cone midranges in the Allison 1 style.

Bottom line for me is it makes no difference.  I love AR 12 inchers. They are relatively inexpensive, they are easy to dial into my listening preferences and these days I listen to multiple AR 12 inch systems simultaneously in the same room so the difference between mid dome and mid cone is now lost for me.    

[tysontom]

On 2/15/2017 at 2:31 PM, Steve F said:

Where has the dome midrange gone?

I worked at BA for 11 years. Andy [Kotsatos] knew I liked AR speakers, but he never knew to what extent, nor did he know that I was a member of this Forum during that time. (And the Forum didn’t know I worked at BA then, either—I was very careful to keep my posts totally free from conflict-of-interest suspicion and complication.)

I did sneak in a few questions to Andy during my BA tenure about his recollections of AR, since he was the main product manager at Advent before co-founding BA, so he knew first-hand about AR Classics and ADDs from a competitor’s standpoint. (BTW, he generally hated them, thought the 3a was ‘lousy,’ the 4x ‘totally muffled,’ but said the 2ax could be a ‘sneaky’ speaker if you weren’t careful.)

Anyway, he always said that the concept of a dome midrange was a ‘fraud’ (his exact word—I’ll never forget it). He felt that they were inefficient, yet also had a shorter excursion than a good 5-inch mid, so the dome couldn’t play as loud. “It can’t handle the same power, it can’t go as loud and yet it needs more power than a cone. Why bother? The “dispersion” advantage is simply untrue, because in the frequency regions where you’re going to use a mid in a well-designed 3-way, neither a cone nor a dome midrange is directional in any meaningful way over its operating band.  Do the math. The 3a’s mid ‘clips’—you can hear it running out of excursion at high power levels. It’s a joke.

Was he right? Is the superior dispersion of a dome mid vs. a 4- or 5-inch cone not particularly important from 300-4000Hz? AR did blow a lot of mids (both my cousins blew their 3a mids playing jazz, not hard rock, in the late-60’s/early-70’s). BA never replaced blown mids under warranty, and I mean never.

I’m not talking about voicing. I always fought w Andy when I voiced a BA speaker, because I like the AR ADD balance and he liked a more forward mid-rangy sound. We’d go back and forth, mostly on woofer choke values.

I want to keep this discussion to the importance of mid-dispersion for forward-facing single-driver-per-band designs, and whether there is truly an audible/meaningful difference between a 1 ½-inch dome and a 4- or 5-inch cone when the listener is sitting in a normal position (less than 10 or 15 degrees off axis from the left and right speakers of a stereo pair).

Thoughts?

Steve F.

Steve, what a provocative topic!  Where has the dome midrange gone?  How about, where has the high-fidelity loudspeaker gone?  It's all a moot point today.

This is a very interesting subject and, for sure, very controversial.  In my opinion, I think it comes down to the philosophical differences between speaker designers Henry Kloss/Andy Petite (Kotsatos) and those of Ed Villchur/Roy Allison, and how each perceived speaker quality and a speaker's sound in a listening room.  The answer lies somewhere in the midst of what is important to the listener, acoustic power or near-field frequency response.  Is flat, forward-sounding speaker output with average acoustic-power output better than rolled-off near-field response with better acoustic-power response?  A louder, more forward-sounding speaker sound usually trumps the more laid-back, "reticent" sound when listened to in the near field.

For many years, Acoustic Research studied the sound of concert halls and what the listener hears in the reverberant listening environment.  Much of this research by AR was done in conjunction with Leo Beranek of Bolt, Beranek and Newman (now Raytheon BBN Technologies).  BBN studied the acoustics of concert halls all over the world, and their research offered a great body of data in helping AR determine the sound reaching the listener's ears in concert halls, etc.  In essence, flat acoustic-power output in a listening room, though slightly downward-tilting to a degree, was closest to the experience of live music in a concert hall.  But not everyone listens to classical music; nevertheless, any type of music from jazz to rock will have a widely dispersed spectral balance, especially the farther back in the room one listens, so acoustic power applies to all types of music in one form or another.

Also, a speaker with very wide dispersion will have a relatively flat acoustic-power (or integrated-power) response when measured in a typical listening room.  A speaker with less midrange and treble dispersion will have a rolled-off power response in that same environment, a function of less sound reflection off the boundaries of the room.  Important: a "bright," forward-sounding speaker with relatively even on-axis frequency response across the spectrum, but embodying average dispersion and average acoustic-power response, will invariably sound duller and less three-dimensional back in the reverberant field; alternatively, the rolled-off, reticent-sounding speaker with excellent dispersion and power response will sound brighter and more "spacious" back in the listening environment.  Therefore, I would add, a 3-way BA speaker with a 1-inch dome tweeter and a 5-inch midrange could not approach the acoustic-power response of an AR-3a in the reverberant field, but the BA will surpass the AR-3a up close.

Dispersion is a function of the frequency wave length and the radiating-diaphragm diameter, and the smaller the radiator diameter, generally, the wider the dispersion down to the frequency at which the driver is omnidirectional.  AR designed dome tweeters—not so much because of their small diameter—but because they were driven at the periphery of the diaphragm (the edge of the moving system), which gave significantly better control over the moving system with fewer break-up modes, lower distortion, better transient behavior due to the lighter mass of the moving systems and so forth.  This translated into lower distortion, greater clarity and greater smoothness.  The small size does give better dispersion, naturally, but a 1-inch dome will not give better dispersion than a 1-inch cone driver!  So the dome design is more than just dispersion. 

An AR-3a 1½-inch dome midrange dome will produce wider dispersion throughout its higher-frequency operating range than any 5-inch cone midrange, despite the superior sensitivity and on-axis output advantage of the cone driver.  In the near field, particularly, the cone driver will typically be more efficient, brighter and more forward-sounding than the AR-3a midrange dome.  The AR-3a's small midrange dome does has limited excursion, as the voice coil has very little overhang in the gap and simply cannot produce very high output.  It is sufficient for nearly any type of music for home use, but not by any means sufficient for high-output sound-reinforcement levels.  Typically, a 5-inch midrange driver can belt out much great output, simply because it has a significantly larger moving system to move more air, greater overhand with more excursion and a voice coil wound with heavier wire, allowing higher electrical and acoustic power levels.  The AR-3a dome uses a very large magnet with a lot of gauss, but the small dome simply can't move much air.   Under duress, therefore, the AR-3a midrange can burn out, whereas the 5-inch midrange will distort badly before it gives up, and this is likely at a high output level.

So what does all this mean?  To Andy Kotsatos, the AR-3a midrange was a "joke," so he says, but philosophically, Andy was from the Henry Kloss school-of-thought to build a speaker that sounds good both in the near field as well as "ok" in the far field, despite the small differences that  might be measurable.  Near-field clarity and brightness was important in the high-fidelity showroom—thus to Andy and Henry—not  the reverberant field.  Kloss found this out immediately after leaving AR in 1957, and he sold his KLH speakers directly against AR speakers (some of which he helped design!) by demonstrating that the AR speakers were low in treble output and low in sensitivity.  In most hifi showrooms, people listened a few feet away to speakers mounted on shelves or on speaker stands,  and many decisions on "speaker quality" were made based on this method of audition. 

The end game in the design of all of the BA and Advent loudspeakers (not to mention KLH and others earlier) was subjective "voicing" in a listening room by Andy (and Henry Kloss before that), deciding how loud or how bright, etc.  Then, for the fun of it, they place an AR-3a up on the shelf and sit back a few feet and listen to that speaker: reticent and hard to take at 5 feet away, correct.  All sorts of interference effects, lobing and interaction, coupled with an output that is 3-7 dB down from the woofer level at close range.  Therefore, the speaker is a "joke" in their minds.  To them it is overrated, "lousy" and probably "stinks" and is "no good."  People in a showroom will be able to hear this.  KLH and Advent dealerships across the globe were "coached" on methods of dealing with the "AR" sound, and since there were many more mail-order outfits selling AR with its "laissez-faire" pricing policy (other manufacturers insisted on fixed suggested retail prices), there were relatively few dedicated AR dealers to defend the AR speakers.  Still, AR outsold everyone in the 60s.     

Neither Advent nor Boston Acoustics did much, if any, anechoic-chamber or free-field measurements, and certainly there were no power-response measurements.  Not even driver response curves or system anechoic curves from Advent, KLH or BA were ever published in any literature to my knowledge, so we have no way of knowing how good these speakers measured at the time.  They sounded good, of course, and these speakers sold extremely well in the late-60s and into the 70s, eventually outpacing AR sales.  So Advent, KLH, Advent (and many other) product-design decisions were driven heavily on market demands by dealers; AR, on the other hand, with its mail-order sales through the mid-60s didn't care about hifi showroom effects, and relied on reviews and critical acclaim and so forth.  AR won this battle in the 1954 through 1967 era, but after Ed Villchur sold AR and after Roy Allison and others left AR in the early 1970s, things changed a bit at the company.

--Tom Tyson

 

[onplane]

Steve thanks for starting such an interesting topic and Tom thanks for your terrific historical perspective.

I happen to own two AR three-ways, my 3a with the dome mid and my TSW-610 with a 6.5 inch cone mid. Prior to bi-amping the TSW sounded orders of magnitude brighter than the more "muffled sounding" 3a. With bi-amping and tone controls, I can make them sound pretty much alike except for dispersion!  Here the 3a dominates with distinction.

My TSW's are in a large room (40 x 20) and I listen in a normal position. Result is excellent sound and good separation. 

My AR-3a's are in my den, a much smaller room. Here the speaker's are NOT normally positioned. Because of the room and the "stuff" in it, the speaker's are way, way too far apart and I sit too close. When I tried the TSW's in this position, the separation was so much that you got the impression of a "hole" in the middle. The 3a's with their superior dispersion masked this problem such that separation sounds both normal and pleasing.

Seriously doubt AR designed these speakers with this problem in mind, but the dome mid just seems to fill with sound space that the cone mid cannot equal.

Regards,

Jerry

[valkyrie]

Wow!!  Thanks Tom Tyson for a great commentary on this issue. 

[Aadams]

...............But the dome midrange enjoyed a far briefer time in the spotlight. There are still a few dome mids here and there from random companies, but the mainstream, right-down-the-middle-of-the-road dome mid is pretty much gone.

The reason is: today most speakers for home use appear to be two way designs or variations using identical drivers in ported enclosures.

A 5 or 6 inch cone driver with a tweeter will suit the needs of the average listener who doesn't care about sounds below 85z.  Apparently a midrange dome can't cover the same spectrum in a two way configuration. As to which is better, dome or cone, for my type of listening, domes are better.

 

[tysontom]

____________________________________________________

I took the measurements (from that link) of the Infinity Reference r152, with its 5-inch woofer/mid and 1-inch semi-horn tweeter, and I compared the off-axis measurements to those of an AR-3a midrange measured in AR's anechoic chamber in 1967 made around the time the AR-3a was commercially introduced.  Some observations:

•  The graph (chart) used for the Infinity measurement is "spread" a bit in the horizontal plane, giving a picture of a smoother-appearing output across the band;

•  The Infinity was only measured out to 30˚ off axis and not beyond (the AR-3a was measured to 60˚ off axis, a much more difficult measurement;

•  The Infinity measurement was not a true anechoic measurement, but rather a "gaited" measurement, and although extremely close, it is not a true "apples-to-apples"    comparison, but definitely close enough to make a quality judgment.  The AR-3a measurement was taken from 575Hz (lower crossover) to 5 kHz crossover to the  tweeter.

Taking these characteristics into account, the AR-3a midrange is flat, ±3 dB, 575 to 5 kHz at 30˚ off axis and ± 4.5 dB, 575 to 5 kHz, 60˚ off axis!  The 5-inch Infinity woofer-midrange, assuming ≈3 kHz crossover, is ±6 dB, 600 – 3 kHz at 30˚ off axis.  This comparison demonstrably shows the superior on-and off-axis performance of the AR-3a 1½-inch dome midrange, and that it is flatter, off-axis, even at 60˚ off-axis, up to a higher frequency than the 5-inch cone woofer-midrange driver in the Infinity.  There is really no comparison.

That said, it is very important to understand that acoustic-power output is not nearly important today with speaker designers as in the past.  It is more academic today, but in terms of true acoustic performance, the small dome midrange will easily out-perform a 5-inch cone midrange in dispersion, and that was the purpose of the original message!  It might be possible to say that the AR-3a dome is a more accurate reproducer, but that might open another can of worms!

—Tom Tyson

[ligs]

Thanks for the analysis. 

Here is an additional graph  of a speaker with a 4 inch midrange. Indeed  it is is now closer to AR3-a in terms of midrange performance. 

http://hometheaterreview.com/measurements-i-also-did-full/

 

Now if we can somehow combine the mid/high performance of this speaker with the 12" AR woofer. bass from AR!

[tysontom]

Thanks Ligs.  That certainly is a pretty curve, but....  

The problem is not so much the 30-degree off-axis performance; the real issue is how it performs 60- and 75-degrees off axis.  This is where the small dome has a huge advantage.  The problem with any cone driver is that you cannot make a 1.5-inch cone driver that can handle sufficient power that would otherwise be able to compete with the 1.5-inch AR dome midrange.  It's down to the physics of the moving system size, and a dome driver has it handled.  In addition, these should be anechoic-chamber measurements of both types of drivers to get true comparisons, avoiding the various smoothing and gait settings that are necessary when measuring in free field or rooms.  There is no real standard here.  In essence, room measurements are helpful but do not necessarily tell the true picture of what the driver is doing -- warts and all.  Thoughts?

--Tom Tyson

[ligs]

Thanks Tom. You have insights that can not be found elsewhere!

Nowadays Harman is probably the only company that has the resources to do in-depth speaker testing. It is interesting they emphasize both listening and lab testing. 

I don't think neither Revel, Infinity nor JBL uses dome midrange in their current speaker lineups. 

 

[Steve F]

Ok, playing devil’s advocate here.  First, let me state for the record that I love the sound of the AR 1 ½-inch dome mid, from about 800Hz on up. Below that it gets in a bit of trouble. Good loudspeaker design dictates that you cross a mid or tweeter over at least an octave above its resonance. The 1 ½-inch dome (essentially the same in all AR 3-ways, whether they had the metal screen or not) had a FAR of 400Hz, so 800Hz would be the “minimum” x-over point. (The 10π and 11 ADD spec sheets in the Library show the drivers’ FARs.)

 

 

But as Tom correctly points out, it's the AR MR dome’s off-axis performance—especially greater than 30 degrees—that sets it apart from a 5-inch cone MR and gives the 3a/11 etc. their spaciousness and flexibility of both speaker and listener placement. Not the dome’s shape, per se. The 1 ½-inch diameter.

 

 

It’s a matter of math. Divide 13560 (the speed of sound at sea level in inches per second) by the diameter of the driver to get the frequency at which that driver becomes noticeably directional.

 

 

So, a 5-inch driver becomes directional at 2712 Hz. A 1 ½-inch dome driver becomes directional at 9040 Hz. You can see that a 3-way speaker with a 5-inch mid crossing over to the tweeter at 4 or 5000 Hz is going to be pretty beamy at the top end of its operating band.

 

 

But when I was at BA, we developed this really great 3 ½” cone midrange for our best speakers. A 3 ½-inch driver is good to 3900 Hz. (Actually, you only count ½ of the surround as the driver’s diameter, so it was really a 3-inch piston, good to 4520 Hz. But let’s not get that niggling about all this, ok?)

 

 

This 3 ½-inch driver had a FAR of 175 Hz and we crossed it over at 400 Hz and 2500 Hz—well above the 2x FAR point and well below the directional point.

 

 

Plus, it had a 1.5-inch voice coil, which was really ridiculously large for a 3 ½-inch cone. Plus the entire cast aluminum basket functioned as a heatsink (it was ribbed like an amp’s heatsink where it encased the neo slug magnet), so between the oversized voice coil (the same diameter as the AR 10-inch woofer’s) and the heatsinking, it could handle virtually unlimited midrange power. The large voice coil exerted great control over the cone’s motion, so the distortion was non-existent and the FR was perfectly flat. This was the way to do a midrange driver. It was twice the driver that the AR 1 ½-ich dome was—far more power-handling, not used below its FAR like AR foolishly did. The BA mid crossed over to the tweeter way before it became directional, so there was no advantage of the AR mid dispersion vs. this BA mid.

 

 

I spent decades in the speaker biz. This is probably the single best driver I’ve ever encountered.

 

 

Everyone pretty much agrees that the 9 and 90 sound so good partly because they relieve the 1 ½-inch dome from having to “stretch down” to a point where they’d rather not go.

 

 

The LST is so good because it divides up the midrange stress among 4 drivers, so there’s not too much distortion or thermal stress on any single midrange driver, like there is in a 3a.

 

 

So....yes, it’s the diameter that counts for dispersion, but you’ve got to use the driver intelligently, within its usable band. Total system design counts more than just one parameter (like dispersion) and when I was at BA, I had the advantage of hindsight, so I didn’t repeat AR’s mistakes with the 1 ½-inch dome.

 

 

Steve F.

[fedeleluigi]

Hello Steve,

Which BA loudspeakers used this 3-1/2" cone midrange and how did they altogether sound in your opinion?

Regards

Luigi

[tysontom]

On 2/24/2017 at 4:43 PM, Steve F said:

Ok, playing devil’s advocate here.  First, let me state for the record that I love the sound of the AR 1 ½-inch dome mid, from about 800Hz on up. Below that it gets in a bit of trouble. Good loudspeaker design dictates that you cross a mid or tweeter over at least an octave above its resonance. The 1 ½-inch dome (essentially the same in all AR 3-ways, whether they had the metal screen or not) had a FAR of 400Hz, so 800Hz would be the “minimum” x-over point. (The 10π and 11 ADD spec sheets in the Library show the drivers’ FARs.)

 

 

But as Tom correctly points out, it's the AR MR dome’s off-axis performance—especially greater than 30 degrees—that sets it apart from a 5-inch cone MR and gives the 3a/11 etc. their spaciousness and flexibility of both speaker and listener placement. Not the dome’s shape, per se. The 1 ½-inch diameter.

 

 

It’s a matter of math. Divide 13560 (the speed of sound at sea level in inches per second) by the diameter of the driver to get the frequency at which that driver becomes noticeably directional.

 

 

So, a 5-inch driver becomes directional at 2712 Hz. A 1 ½-inch dome driver becomes directional at 9040 Hz. You can see that a 3-way speaker with a 5-inch mid crossing over to the tweeter at 4 or 5000 Hz is going to be pretty beamy at the top end of its operating band.

 

 

But when I was at BA, we developed this really great 3 ½” cone midrange for our best speakers. A 3 ½-inch driver is good to 3900 Hz. (Actually, you only count ½ of the surround as the driver’s diameter, so it was really a 3-inch piston, good to 4520 Hz. But let’s not get that niggling about all this, ok?)

 

 

This 3 ½-inch driver had a FAR of 175 Hz and we crossed it over at 400 Hz and 2500 Hz—well above the 2x FAR point and well below the directional point.

 

 

Plus, it had a 1.5-inch voice coil, which was really ridiculously large for a 3 ½-inch cone. Plus the entire cast aluminum basket functioned as a heatsink (it was ribbed like an amp’s heatsink where it encased the neo slug magnet), so between the oversized voice coil (the same diameter as the AR 10-inch woofer’s) and the heatsinking, it could handle virtually unlimited midrange power. The large voice coil exerted great control over the cone’s motion, so the distortion was non-existent and the FR was perfectly flat. This was the way to do a midrange driver. It was twice the driver that the AR 1 ½-ich dome was—far more power-handling, not used below its FAR like AR foolishly did. The BA mid crossed over to the tweeter way before it became directional, so there was no advantage of the AR mid dispersion vs. this BA mid.

 

 

I spent decades in the speaker biz. This is probably the single best driver I’ve ever encountered.

 

 

Everyone pretty much agrees that the 9 and 90 sound so good partly because they relieve the 1 ½-inch dome from having to “stretch down” to a point where they’d rather not go.

 

 

The LST is so good because it divides up the midrange stress among 4 drivers, so there’s not too much distortion or thermal stress on any single midrange driver, like there is in a 3a.

 

 

So....yes, it’s the diameter that counts for dispersion, but you’ve got to use the driver intelligently, within its usable band. Total system design counts more than just one parameter (like dispersion) and when I was at BA, I had the advantage of hindsight, so I didn’t repeat AR’s mistakes with the 1 ½-inch dome.

 

 

Steve F.

Steve,

Great discussion!

That 3.5-inch midrange BA driver description sounds fascinating and quite impressive.  It was undoubtedly a stellar performer with little or no stress under normal circumstances.  It is very true that the AR-3a midrange can be stressed if pushed (but generally there is negligible distortion in the midrange at any reasonable output level).  The voice coil has marginal overhang in the gap, but there is considerable flux outside the immediate gap, too.  Nevertheless, the the dome is small and can move only so much air, especially as the lowest-operating frequencies required of it. 

Do you have any measurements on the BA 3.5-inch driver, on and off-axis?  It would be interesting to compare its anechoic-frequency curve, on- and off-axis up to 60 degrees vs, the AR 1.5-inch midrange to compare output linearity.  Insofar as the BA 3.5-inch MR used a neodymium slug-type magnet, it would mandate a large-diameter voice coil just to have room for the magnet assembly within the speaker driver itself.  Yoke and slug magnets are also somewhat susceptible to demagnetization from the voice coil if power or heat is sufficiently high enough, but this was more of a problem with woofers, and not so much of a problem with neo magnets, though high heat is a problem (thus the big heat sinks used by BA).  There were often reports of some yoke-magnet woofers and mids losing magnetization because of excessive current through the voice coils, as the coils were directly adjacent to the magnet itself. 

Also, the 1.5-inch MR diameter should not be conflated with an AR-2/5/6 woofer's 1.5-inch diameter voice coil; the windings on the AR 10-inch woofer were much larger bell wire with a much longer coil height, and the woofers were capable of several times as much actual input current.  Nevertheless, a 1.5-inch voice coil in a 3.5-inch driver is impressive. 

I wonder what constitutes "noticeably directional?"  Is it beyond 30 degree off axis, 60 degrees?  I think it is just the point at which it becomes noticeable, and that might be just a few degrees off axis at higher frequencies.  The beauty of the AR-3a 1.5-inch dome midrange is that it does well out to 60-75 degrees off axis -- well beyond its intended operating band.  When I look at AR's 3.5-inch midrange driver, as used in the AR-2ax, early 2x and AR-4, the 30-degree performance is pretty much equal to the 3a dome up fairly high, but the wider the measurement angle, the dome is much better; thus the superior acoustic-power response of the smaller-diameter driver.  It's all an academic argument, but fun to contemplate.

So, therefore, the 5-inch driver is clearly no match for the 1.5-inch driver in off-axis performance, but the 3.5-inch MR is a much closer match.  The answer as to linearity and smoothness when comparing those two would reside in the anechoic-chamber measurements themselves.  For sure, if power levels get too high with the AR-3a, the dome midrange can run out of excursion, yet it is amazing that the AR-10Pi could sustain 800+ watt peaks during the Neil Grover live-vs.-recorded drum demonstration!  You were there, too!  This is the same midrange, though these may have had Ferrofluid!  I do know that the tweeters had Ferrofluid.  Victor Campos told me that the bulk of the high-energy peaks went to the midrange/tweeter units (mostly drum "rim" shots), and that no drivers were lost due to thermal damage during the actual demonstrations.  Two woofers were damaged because of the Dunlap-Clarke Dreadnaught 1000 amplifier-bias failure. 

--Tom

 

 

[Steve F]

 

The BA mid only had to go to 2500Hz, so its off-axis performance was easily equal to the AR dome MR to that point. And it was a far superior performer at the lower end of its band, due to greater excursion, more piston area and a lower FAR. The BA engineer who originally used this driver for the VR-M90 was Mike Chamness, an Allison/power response devotee, who spec'd this driver for precisely the reasons I've laid out. As I said, it was unquestionably the single best driver I've ever worked with (although Atlantic Technology's 3 1/2-in mid came very close. Shhhh....I pretty much copied the BA driver once I got to AT).

Yes, the 10π did a fabulous job in the L v R demo. Remember, we're talking about absolutely superb, top-tier drivers here, with both the AR MR dome and the BA 3 1/2-incher. It's not that the AR unit was deficient. It was an A- or A by today's standards. The BA mid noses it out in overall performance by a tad, given the benefit of hindsight, and more modern materials, manufacturing, etc. AR's fanatical QA probably rejected 20-30% of their mids in order to hold their very tight QA window. AR used to show big trash bins of rejected drivers in one of their ads. BA's QA standard was +/- 1dB per driver, yet we rejected less than 1%. Trash bins of rejected drivers might have been a badge of honor in the 1960's, but they'd be a total humiliation in the late 1990's-2000's.

Time moves on.

Steve F.

[iso]

There are some good mid range domes made even today. ATC and probably Volt unit too can produce something like continuous system SPL @1m: 117dB SPL  when crossed at 380 Hz/12dB. They use usually short horn to control dispersion. ATC SM75-150S has been in production since mid 70`s. Volt unit is conventional long coil in short gap design but ATC-dome does use more efficient short coil in long gap and both are of double suspension design to improve linearity.

http://atcloudspeakers.co.uk/technology/soft-dome-mid-range/

http://www.voltloudspeakers.co.uk/loudspeakers/vm752-3/

 

Kimmo

 

[ligs]

I just could not resist pulling out my AR 58 1.5” dome midrange/Ribbon tweeter combination from the basement and gave another audition. The bass is a pro version of JBL 14” Neo woofer (similar to that used JBL Synthesis 1400 Array), crossovered around 700 hz to AR midrange. Next to it I have JBL 104H-2 4” mid and 1” 035Tia Titanium tweeter used in JBL L100 T3. I have to say the AR combo has no lacking of mid or high and it has the same sparkling high as JBL when program calls for it.  I admit AR is winning me over again.

[tysontom]

6 hours ago, Steve F said:

 

The BA mid only had to go to 2500Hz, so its off-axis performance was easily equal to the AR dome MR to that point. And it was a far superior performer at the lower end of its band, due to greater excursion, more piston area and a lower FAR. The BA engineer who originally used this driver for the VR-M90 was Mike Chamness, an Allison/power response devotee, who spec'd this driver for precisely the reasons I've laid out. As I said, it was unquestionably the single best driver I've ever worked with (although Atlantic Technology's 3 1/2-in mid came very close. Shhhh....I pretty much copied the BA driver once I got to AT).

Yes, the 10π did a fabulous job in the L v R demo. Remember, we're talking about absolutely superb, top-tier drivers here, with both the AR MR dome and the BA 3 1/2-incher. It's not that the AR unit was deficient. It was an A- or A by today's standards. The BA mid noses it out in overall performance by a tad, given the benefit of hindsight, and more modern materials, manufacturing, etc. AR's fanatical QA probably rejected 20-30% of their mids in order to hold their very tight QA window. AR used to show big trash bins of rejected drivers in one of their ads. BA's QA standard was +/- 1dB per driver, yet we rejected less than 1%. Trash bins of rejected drivers might have been a badge of honor in the 1960's, but they'd be a total humiliation in the late 1990's-2000's.

Time moves on.

Steve F.

This is a interesting discussion, but it is more subjective than objective.  Without frequency-response and power data on the BA midrange unit, there is no way to know how well that unit would actually compare with the 1½-inch AR-3a dome mid.  It might actually outperform it, but we don't have hard data.  "And it was a far superior performer at the lower end of its band" might also be true, but that is a subjective description and not a completely valid argument. 

Technically, of course, a low-resonance BA 3½-inch cone driver with a larger surface area—and likely an underhung voice coil—should  perform much better at lower frequencies than the AR midrange.  Nevertheless, the AR-3a midrange is uniform and very flat within its operating region of 575-5kHz, and our discussion was originally about dispersion and ultimately, power response.  It's just that at the lower end of its operating range, the AR-3a midrange won't blow you out of the room.  At normal levels, there is never a problem.

It is also true that AR recognized the need for greater power-handling in the midrange, and raised the low limit on those drivers in the newer AR9, AR90 and so forth, to reduce the stress on the midrange. 

At 30˚ to perhaps 45˚ off axis, the 3½-inch BA cone midrange should do well within its range; however, beyond that angle, there is a question as to how it would perform compared with the AR dome midrange, and flat power response requires very wide dispersion at all frequencies.

You have hit on one of the big differences between the early manufacturing processes and later, more automated production, when you described the "rejected" drivers at AR during production.  The rejection rate of AR midrange drivers was not as high as you describe—the ¾-inch tweeters were—but it was obviously higher than speakers built with higher automation in later years.  Although the 1½ MR used in the AR-3a was less "hand-made" than earlier drivers, it still was tested for any quality issues and rejected if it didn't meet certain standards.  AR did have barrels of rejected drivers that were destined to be rebuilt and returned to inspection, but in later years, production methods greatly improved for the high-fidelity industry, for sure.   In the end, the performance and quality of the product are what counted, and this required more QC intervention with the more hand-assembled speakers of the 50s-80s than later automated production.  Time does "move on," but this is really not our discussion here.

—Tom