AR's honesty and the 3a's excellence

[Steve F]

This is an in-room curve of the 3a from the Berkovitz-Allison AES paper on Soundfields in domestic rooms. It’s a fascinating curve, because it demonstrates AR’s honesty and technical proficiency, attributes that put AR well ahead of other speaker companies of that time.

First of all, it’s doubtful if any other company would let anything other than some phony-baloney hand-drawn ruler-flat response curve even see the light of day. Sure, this was an AES paper (not an ad in Stereo Review), but it was out there in the public domain. Nonetheless, AR had such high principles of broadening the general understanding of speaker behavior/performance and had such (justifiably) high confidence in the correctness of their design ideals and approach that they willingly let an imperfect-looking FR curve of their best speaker—warts and all—out of the secret confines of their marketing department.

Let’s look at the curve a bit more closely. It’s the FR of an AR-3a in a home living room, with the Mid-Hi controls all the way up. The speakers were on stands, somewhat away from the room’s walls. This was not an uncommon usage scenario, although certainly not an optimum one.

Using the speakers on stands, away from the walls, is a classic “4π” placement. This placement provides the least amount of bass reinforcement. The 3a (like all AR speakers at that time) was designed for 2π bookshelf placement (as in Tom’s bookshelf).

You can see the woofer’s response falling off from the midrange level below 500Hz as the solid angle that the speaker “sees” transitions from 2π to 4π. (The 25” front baffle of the 3a serves as the 2π baffle for frequencies above around 5-600Hz. A 565Hz wavelength is 2 feet long, and 25 inches is…..it’s the math, the math…..). BTW, any box-type bookshelf speaker on a stand away from the walls will show the same mid-to-bass falloff as the 3a, when placed similarly.

If you raise the under 200Hz woofer level up to the mid’s level (as the woofer level would be in a 2π setting), you can see that the speaker’s response from just under 40 Hz to a bit over 2,000 Hz is remarkably flat and accurate. This is the bass-midrange region, where virtually all the music lies.

Above than, you see the falloff—but very smooth—in the upper-end response, partly due to AR’s willingness to sacrifice flat on axis upper treble for wide dispersion (in the days before ferrofluid-cooled tweeters) and party because RA’s somewhat misguided (IMO) view that the loudspeaker should intentionally introduce the particular coloration of the performance venue (in this case, the HF rolloff of major orchestral halls) into the domestic playback chain. In my opinion, the speaker should simply reproduce the electrical signal fed into it as accurately as possible. Nothing more, nothing less. It should be up to the recording engineer/record producer to impart any sense of the performance space into the recording, if desired. Not the loudspeaker.

In any event, with the woofer level raised to the mid’s level as it would be in a 2π space, the 3a’s real-world, in-room performance is quite excellent. If you were to tweak the treble just a hair, it goes from quite excellent to unconditionally superb.

This was quite a loudspeaker.

Steve F.

[Pete B]

When I listened to the AR-11s, I commented that -6 dB on the mid was a bit too much,

and -3 dB was not enough.  If you look at that curve dropping the mid by about 4-5 dB will

flatten it quite a bit.

Do you know if the measured 3a had the 1.9 mH or 2.85 mH inductor?

http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1270-anyone-have-ar-3a-measurements-system-drivers/

Did we ever figure out what year the 3a inductor went from 1.9 to 2.85 mH.

If I remember correctly there was not a connection to the change to the ceramic magnet woofer.

[frankmarsi]

Hi SteveF,   excellent and enjoyable write-up, as usual.  I hope some of the newer members take heed to the points you've made, especially about cabinet and correct room placement.

There seems to be a trend for some here of rewriting the rules about positioning and proper placement. I sometimes attempt to inform/correct them but, I'm usually stonewalled and besmirched for it.

FM

[Pete B]

Actually, where is the floor bounce dip in that chart if it is in room, it seems that claim is incorrect.

Here is the figure from one of the AES articles that I posted many years ago (Feb 1, 2006) where it is

clearly stated as 4pi anechoic:

I wrote this about it:  in this thread:  http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1571-ar-3a-4pi-anechoic-measurements-from-1970-aes/

AR-3a 4pi anechoic measurments from Allison and Berkovitz's "The Sound Field in Home Listening Rooms", JAES July/August 1972, Figure 8.

Note that while the date of the AES article is 1972, the paper was presented on October 12, 1970 at the 39th Convention of the Audio Engineering Society, New York. Obviously the research was done prior to presenting the paper around the time of the change from the #7 to #9 inductor. Clearly, AR was doing 2pi and 4pi measurements given that they're shown in the publication. I'm not going to show the 2pi measurements since it is not disputed that these were made.

 

 

[Aadams]

23 hours ago, Steve F said:

Above than, you see the falloff—but very smooth—in the upper-end response, partly due to AR’s willingness to sacrifice flat on axis upper treble for wide dispersion (in the days before ferrofluid-cooled tweeters) and party because RA’s somewhat misguided (IMO) view that the loudspeaker should intentionally introduce the particular coloration of the performance venue (in this case, the HF rolloff of major orchestral halls) into the domestic playback chain.

 IMO, the 3a is unlike any other 12” AR speaker with 3 drivers.  It took me a while to realize this but when folks refer to the 91 as the ultimate expression of the 3a it just isn’t true.  The 3a, has a presentation that cannot be duplicated by any AR made after the LST.  The difference only becomes very obvious with recordings that are not close miked, almost all of which are classical and even then not numerous.  If you don’t listen to classical music, you will most probably never hear the big advantage of the 3a. Instead the difference in presentation between it and the later models will be heard as an improvement in imaging because most recordings since the 1960s come from a close miked mix.

Adams

 

[tysontom]

I think this is a great message about the AR-3a, and Steve has described the AR-3a well!  Figure 8 of Roy Allison’s paper was indeed a 4π anechoic rather than “room” measurement, but Steve’s description is nevertheless completely accurate. 

As he states, the curve show’s Acoustic Research’s basic honesty in presenting the performance curves, and it shows AR’s proficiency in quantifying the performance of their speakers.  To repeat what we’ve said before, not all speaker companies adhered to such a strict “standard,” but from the beginning this attitude of honesty prevailed at AR.  The important thing about this AES paper was to show the AR-3a’s true acoustic-power response, the “integrated-output” response of the system, which was shown in later figures within Allison’s paper.   It is noteworthy that the grill molding did cause some disturbances in the midrange, particularly, that could be seen in the pressure response, but those irregularities do not affect the total integrated output, so that in effect one does not hear interference differences in amplitude that are picked up by a microphone placed 1 meter in front of a loudspeaker.  It has also been shown that by merely moving the microphone a few inches one way or the other changes the frequencies in which those irregularities occur, further reinforcing that fact.

The hard thing to understand is that, despite the lower relative energy level of the AR-3a tweeter, the total energy put into a listening room by the AR-3a is strong all the way out to the limits of audibility.  By comparison, if you had another speaker with completely flat on-axis response out to 20kHz, but poor off-axis response, its total energy into the room would be lower at these highest frequencies than the AR-3a; that is, the AR-3a would sound brighter in the reverberant listening field.  A highly directional speaker in the upper midrange and treble might measure "flat" on axis, but it will invariably sound "dull" well back into a normally damped, reverberant listening environment.

The drop-off in the AR-3a’s bass response below about 400 Hz—the difference between the 4π and 2π mounting configuration in the anechoic chamber—is quite dramatic, but predictable, and shows how bass response responds to different mounting positions; this bass difference shows up when mounting a speaker in a room, such as in a corner, along a wall or out into the room. 

In the mid-1960s, a couple years before the introduction of the AR-3a, Edgar Villchur received a letter from a new AR-3 owner who had taken his speaker to the Harvard University anechoic chamber and measured the woofer’s response.  An engineer friend had prompted the test, saying that the AR woofer wasn’t flat into the deep bass and was defective and perhaps overrated.  The engineer then told the owner that he could prove it by testing it in an anechoic chamber.  Sure enough, the measured response did show a severe drop-off in response, and the owner was distraught.  What the engineer-friend hadn’t considered was the measurement solid angle.  He had measured it into a 4π angle, and the big Harvard chamber was anechoic down to fairly low frequencies.  Had the speaker been measured into a 180° sold angle, the output would have been flat down to resonance.

—Tom

[RoyC]

18 hours ago, Pete B said:

Do you know if the measured 3a had the 1.9 mH or 2.85 mH inductor?

http://www.classicspeakerpages.net/IP.Board/index.php?/topic/1270-anyone-have-ar-3a-measurements-system-drivers/

Did we ever figure out what year the 3a inductor went from 1.9 to 2.85 mH.

If I remember correctly there was not a connection to the change to the ceramic magnet woofer.

Pete, The inductor change was definitely linked to the later woofer.

Roy

[tysontom]

2 hours ago, RoyC said:

Pete, The inductor change was definitely linked to the later woofer.

Roy

I think it was a matter of a slight difference in the BL of the two woofers.  The newer ferrite woofer definitely has a slightly larger magnet (≈11 lb. structure vs. 9.6 lbs), but the voice coil is almost identical and the moving system is nearly identical in weight.  The production details of the new woofer were done mainly by Chuck McShane under Roy Allison's direction.

—Tom

[lARrybody]

Exactly 24 hours ago I acquired my third set of AR3a's. What makes these different is that they were from the original owner who purchased them 48 years ago while in the Army (1970).

At the time he told me they cost him a months wages. Serial numbers 28801 and 27932. My first pair have numbers in the 33XXX range, but upon removing the grills found Tonegen replacement woofers with the number 7 inductor still in place. I replaced them with a 18ga 3.0mh Erse coil unwound to 2.85mh.when they were recapped, but have been on the lookout for a set of alnico woofers so I saved the original coils. So far that has not happened. I am 99% sure these new 3a's have the cloth surrounds.

He assured me that they had never been apart and upon inspecting them I tend to agree. Today I hooked them up to a Lepai Class D bluetooth amp which probably puts out 80 watts per channel at 4 ohms. Good little amp currently available  at Parts Express.  Its hard to believe that they sound so good. Pots work flawlessly, and all drivers have full output. It took about 30 minutes for the capacitors to re energise and smooth out. Still amazed at how these 48 year old speakers that had been in storage for years can perform at this level. Yes the AR3a is a special speaker. I have a brand new XLS 2502 that I will soon hook them up to. If they rock out on 80 watts I Can't wait to hear them with 800 watts per channel.   A few pictures and back to listening.

[ra.ra]

Those 3a's are gorgeous, Larry - - the cabs, the grilles, the badges - - they all look like they are still in the audio sales showroom, except for that milk crate. Now tell us about your second pair, too.

Despite having previously read dozens and dozens of threads here about the AR-3a, the pics of these speakers sparked an intense personal recollection of my first real exposure to this historic speaker model. The year was 1970 or 71 - - location: northern Indiana - - and I knew two guys who were into this audio gear much more than me. One was a year or so younger, and I think he had an AR-XA turntable and OLA speakers and Dynaco electronics, but what really peaked my interest was his Advent 201 cassette deck - - the first I'd ever heard about Dolby anything - - and the recordings made on his deck sounded so superior to anything from the recording heads of my simple Sony deck. The second guy was a year or so older and he had a Dual table, a big fat Marantz receiver, and a pair of walnut AR-3a's. We'd listen to things like Hendrix, Santana, Cream, Beatles and Albert King and probably even Cat Stevens, and despite my deep appreciation for my AR-4x's at home, it was so obviously apparent to me that his system displayed a much larger orchestra than mine did. This guy's name was Bo, and he rode a Harley-Davidson Sportster motor pickle at high speeds. Thinking back about him, he strikes me as probably 5% Bo Schembechler (minimally disciplined); 10% Bo Derek (kind of a pretty boy); 25% Bo Jackson (strong and multi-talented); and 60% Bo Diddley (a real blues guy). Whenever I see that unique wide face frame on the AR-3a's, I remember my time with Bo.     

Back to the topic at hand - - primarily due to this site and with hindsight, I really appreciate having become aware of the marketing honesty promulgated by AR with its advertisements and product literature. And that pic of Tom's 3a's in a true bookshelf configuration is the way I always like to see these speakers situated within a home environment.