AR-3a 4pi Anechoic Measurements from 1970 AES

[Pete B]

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.

http://members.aol.com/basconsultants/AR3a4pi.jpg

[Pete B]

I'm not going to take snap shots from my simulations of the 1.88 mH and the 2.85 mH inductors because it would be time consuming and we already have similar results here on the forum. I've not run my own in many months, I don't remember the exact results, but they were similar to what's shown below. The AR-11 crossover response (red), which is very similar to the AR-3a, has peaking around the upper low pass crossover that reduces the amount of compensation for baffle loss. It is rare that the full amount is used however the figure below is for the ideal case of zero ESR in the 120 uF cap, the peaking would be less with a real cap moving it closer to the AR-303 response shown in green.

It is interesting to look at the response of the AR-303 where it can be seen that the crossover, provides ~7 dB of boost, at 80 Hz as compared to 400 Hz. This approximately compensates for the droop seen in the above 4pi diagram where 400 Hz is ~7 dB higher than 100 Hz. Obviously, I'm comparing the AR-303 to the AR-3a (probably with the 1.88 mH coil) but these are the concepts that are used in crossover design. This boost is not needed if the 303 were measured into half space and this is why the peaking is seen at 60 Hz in the Stereophile review IMO. It seems based on this simulation that the AR-303 has full BSC compensation. It also seems that the AR-11 family, has about 4 db or about 2 dB less than full compensation. This should all be verified through measurements of a real system:

http://www.classicspeakerpages.net/dc/user_files/836.jpg

[Steve F]

What is also very interesting is Roy Allison’s description of the reason for the shape of the 3a’s response in Figure 8. It’s quite clear that Allison fully understood the 2 Pi/4 Pi situation:

“The most obvious feature of Fig. 8 is the woofer response. There is a continuous downhill slide from about 400 Hz, at which frequency the cabinet is a reasonably effective 2 Pi baffle, to about 170 Hz, then a flat response below that frequency. At 170 Hz and below the radiation angle is 4 Pi steradians and the output, quite predictably, is lower than it was into a 2 Pi angle.”

If one were to raise the 3a’s response in this graph in the frequency region in question by the appropriate 6dB below about 170Hz (progressively less above that frequency as the baffle transitions from 4 Pi to 2 Pi), then you’ll see a very smooth, even response from the low bass right through the midrange.

Except---the region around 400-500Hz would still be about 2dB higher than the rest of the woofer region. Which is why, not coincidentally, Roy Allison just told Tom a few days ago that AR changed the 3a’s woofer coil to reduce the very top end of the woofer’s response by a couple of dB in that region.

Nowhere in this article by Allison and Bob Berkovitz do they say (or even imply) that as a result of their real-world measurements of 16 pairs of AR-3a’s in 22 locations in 8 actual listening rooms they (AR) were going to change the 3a’s crossover to optimize the 3a for a 4 Pi environment. Forum members, if they don’t own the JAES Compendium Volume 1, can go the JAES website and download the article for $20.00. Enter the article’s title “The Sound Field in Home Listening Rooms” by Roy Allison in the search dialog box and follow the instructions from there. Read it yourself. The thought of “redesigning” the 3a is never mentioned by either of them.

The real points of this article were to ascertain how much of domestic listening environments are reverberant vs. direct, whether first arrival or energy response matters most, whether first-arrival determines directionality AND spectral balance or just directionality, etc. Fascinating topics, but this article by Allison and Berkovitz does NOT concern itself with whether the 3a should be or ever was redesigned for a 2 Pi vs. 4 Pi environment.

Yes, AR made both 2 pi and 4 Pi measurements. But it’s apparent that they made the 4 Pi and 2 Pi measurements in this study to get a better understanding of the various real-world acoustic settings in which loudspeakers were used—-not as a basis on which to redesign the 3a. The 3a—-like all AR speakers of that time period—-was designed for 2 Pi environments. Their literature said so. Their owner’s manuals said so. And after reading this article again and reading Roy Allison’s comments to Tom Tyson a few days ago, there doesn’t seem to be any question. Roy said that the coil change was to correct a response aberration at the top end of the 3a’s woofer’s range and I have no reason to disbelieve him.

Steve F.

[Guest Bret]

>If one were to raise the 3a’s response in this graph in the frequency region in question by the appropriate 6dB below about 170Hz (progressively less above that frequency as the baffle transitions from 4 Pi to 2 Pi), then you’ll see a very smooth, even response from the low bass right through the midrange<

Not-coincidently, the tap on the autotransformer in a 10pi which puts the woofer in 4pi mode, raises the signal to the woofer crossover by 6db in excess of the 2pi setting (2pi is 0db gain or loss).

.

The 4pi woofer setting does other things too, which I suspect is correcting a "sag" caused by the 4pi setting. It appears to alter the woofer's crossover. Nobody quote me just yet, but at this stage of study it appears to me that the infamous #9, 2.85mH inductor stays in the circuit no matter what "pi" position is selected by the "environmental control."

Bret

[Guest richd]

Attached is a circuit simulation showing the two inductor values. Again I point out this is the voltage at the input of the speaker (Kantor supplied model in library) in free air not the acoustic pressure at the output of the speaker.

Blue 1.85uH Red 2.85uH Cap ESR 100mohm

David Rich

[Guest Bret]

>Not-coincidently, the tap on the autotransformer in a 10pi which puts the woofer in 4pi mode, raises the signal to the woofer crossover by 6db in excess of the 2pi setting (2pi is 0db gain or loss).<

Correction: 5db

Bret

[Pete B]

First, I notice that you seem alarmed by the use of the phrase "redesign", I have never intended for my use of this phrase to refer to the #7 to #9 inductor change alone, or it being a major redesign. I was referring to a redesign going from the ALNICO woofer to ceramic, combined with the inductor change and perhaps others. Taken as a whole it was a redesign or modernization in my view, it is just semantics in any case, no need to become alarmed over it.

You make many obvious statements here, many based on owners manuals and marketing literature. Yes I know that AR produced primarily bookshelf speakers, intended for placement against the wall. And yes the marketing literature might lead one to believe that all the 2pi testing and measurements might make AR speakers ideal for such placement. However, the fact is that a real room is a highly complex environment. It is a cavity resonator - room modes, it is certainly not exactly 2pi since at low frequencies the other boundaries come into play, and it is certainly not 4pi, it is complex - let's keep this in mind.

Again, let me remind you that in the past you suggested that there was no inductor change and that the different woofer measurements were minor differences - you were wrong. Now you come back again with your spin on the article claiming:

>Nowhere in this article by Allison and Bob Berkovitz do they

>say (or even imply) that as a result of their real-world

>measurements of 16 pairs of AR-3a’s in 22 locations in 8

>actual listening rooms they (AR) were going to change the 3a’s

>crossover to optimize the 3a for a 4 Pi environment.

Then you suggest your view of the intent of the article as if to say that there is no other point being made, or perhaps none to support my perspective:

>The real points of this article were to ascertain how much of

>domestic listening environments are reverberant vs. direct,

>whether first arrival or energy response matters most, whether

>first-arrival determines directionality AND spectral balance

>or just directionality, etc. Fascinating topics, but this

>article by Allison and Berkovitz does NOT concern itself with

>whether the 3a should be or ever was redesigned for a 2 Pi vs.

>4 Pi environment.

Here's the point in support of my view, they write with regard to the 16 pairs of AR-3a's in 22 locations in 8 actual living rooms:

"The general trend in the range below 250 Hz demonstrates that the average real room does not give the low-frequency support that is commonly assumed. Since this falling response is not a property of the AR-3a when radiating into a hemisphere, and cannot be attributed to a large radiating angle (the average angle in these rooms is visually smaller, not larger, than 2pi), it must clearly be the result of inreasing energy absorption at low frequencies. Room furnishings do not absorb much at low frequencies (certainly not more than at middle frequencies). Evidently most of the loss is the result of inadequate boundary stiffness, flapping walls, floors, ceilings, and windows."

The key point above is "the average real room does not give the low-frequency support that is commonly assumed."

Commonly assumed of course for 2pi radiation, my conclusion is that real rooms are somewhere between 2pi and 4pi. This is why I have stated that full BSC compensation is not typically used.

You are stuck on semantics, one man's power response change is another's BSC change, it is just a matter of perspective. As I have also said before change the top end of the woofer response, is relative to something, that something is the low end. You and Allison say that the top end was lowered, I say that the low end was raised in a relative sense, once again it is a matter of perspective.

You should have gone into politics Steve the way that you spin things.

I do not have time to put into these long discussions, I'm going to take a break.

Pete B.

>What is also very interesting is Roy Allison’s description of

>the reason for the shape of the 3a’s response in Figure 8.

>It’s quite clear that Allison fully understood the 2 Pi/4 Pi

>situation:

>

>“The most obvious feature of Fig. 8 is the woofer response.

>There is a continuous downhill slide from about 400 Hz, at

>which frequency the cabinet is a reasonably effective 2 Pi

>baffle, to about 170 Hz, then a flat response below that

>frequency. At 170 Hz and below the radiation angle is 4 Pi

>steradians and the output, quite predictably, is lower than it

>was into a 2 Pi angle.”

>

>If one were to raise the 3a’s response in this graph in the

>frequency region in question by the appropriate 6dB below

>about 170Hz (progressively less above that frequency as the

>baffle transitions from 4 Pi to 2 Pi), then you’ll see a very

>smooth, even response from the low bass right through the

>midrange.

>

>Except---the region around 400-500Hz would still be about 2dB

>higher than the rest of the woofer region. Which is why, not

>coincidentally, Roy Allison just told Tom a few days ago that

>AR changed the 3a’s woofer coil to reduce the very top end of

>the woofer’s response by a couple of dB in that region.

>

>Nowhere in this article by Allison and Bob Berkovitz do they

>say (or even imply) that as a result of their real-world

>measurements of 16 pairs of AR-3a’s in 22 locations in 8

>actual listening rooms they (AR) were going to change the 3a’s

>crossover to optimize the 3a for a 4 Pi environment. Forum

>members, if they don’t own the JAES Compendium Volume 1, can

>go the JAES website and download the article for $20.00. Enter

>the article’s title “The Sound Field in Home Listening Rooms”

>by Roy Allison in the search dialog box and follow the

>instructions from there. Read it yourself. The thought of

>“redesigning” the 3a is never mentioned by either of them.

>

>The real points of this article were to ascertain how much of

>domestic listening environments are reverberant vs. direct,

>whether first arrival or energy response matters most, whether

>first-arrival determines directionality AND spectral balance

>or just directionality, etc. Fascinating topics, but this

>article by Allison and Berkovitz does NOT concern itself with

>whether the 3a should be or ever was redesigned for a 2 Pi vs.

>4 Pi environment.

>

>Yes, AR made both 2 pi and 4 Pi measurements. But it’s

>apparent that they made the 4 Pi and 2 Pi measurements in this

>study to get a better understanding of the various real-world

>acoustic settings in which loudspeakers were used—-not as a

>basis on which to redesign the 3a. The 3a—-like all AR

>speakers of that time period—-was designed for 2 Pi

>environments. Their literature said so. Their owner’s manuals

>said so. And after reading this article again and reading Roy

>Allison’s comments to Tom Tyson a few days ago, there doesn’t

>seem to be any question. Roy said that the coil change was to

>correct a response aberration at the top end of the 3a’s

>woofer’s range and I have no reason to disbelieve him.

>

>Steve F.

>

[Steve F]

Pete,

We all admire your spirited posts, put forth with your customary heart-felt enthusiasm and passion. It makes for a lively Forum, worthy of the time spent here by all of us. The fact that you are extremely technically adept makes your input that much more interesting.

It would seem, after we discard all the semantics, that everyone here on the Forum is in agreement that:

1. The 3a underwent significant upgrades and improvements during its 8-year (1967-1975) market lifespan. These upgrades included a change from Alnico to ceramic woofer magnet, from cloth woofer surround to foam woofer surround, from front-wired MF/HF drivers to rear-wired drivers, from glued-in grille cloth to Velcro-removable grille cloth, from the #7 to the #9 coil change (which was probably the cause of the spec change from 575Hz to 525Hz). There were undoubtedly other changes and upgrades as well that I’m forgetting at the moment.

2. The change to the coil was, as Roy said (not what I said—I’m just repeating Mr. Allison), to reduce the upper end of the woofer’s response between 400-500Hz by about 2dB. Does a reduction in a woofer’s TOP end response elevate its BOTTOM end response by comparison? Yes, of course it does! You’re not wrong, and no one here is saying or implying that you are. Roy’s words were “reduce the top end of the woofer’s response.” He could have said it slightly differently and still imparted the same information.

3. The 3a was never consciously, explicitly redesigned for a specific 4Pi environment.

However we all say it, by whatever route we take to reach those conclusions, by whatever phrases we choose to employ, it appears that all of us agree on those points.

This has been a terrific discussion and your participation was instrumental. We all enjoyed it. Thanks again.

Steve F.

[dynaco_dan]

Hi there;

It sure is a pleasure, to read praise and compliments now.

Makes for a much better and productive atmosphere.

Thank you.