Jump to content

Inconsistencies in AR's 12" woofer curves


Steve F

Recommended Posts

This has been something I've noticed for 45 years or so.

The late AR-3/early AR-3a woofers were identical, right? Or at least, very nearly so. Certainly from a functional standpoint, if not absolutely identical, realistically identical, to the point that they could be drop-in replacements for each other.

Therefore, their published FR curves should be identical also. But the curves published by AR are certainly not identical. Unless AR held back a little info that would explain the differences.

See the attached graphs. In the (upper) AR-3 woofer FR, it goes all the way out to 1000 Hz. But the AR-3a woofer FR graph rolls off well before that, and very smoothly both on-and off-axis.

I get the feeling that these woofer curves are curves that AR ran with the woofers operating into their low-pass filters. The bass end through around 400-500 Hz of the two graphs is absolutely identical, which one would expect, seeing as they're indentical units.

But at the upper end of the woofers' range, the two deviate considerably.

There is nothing wrong or nefarious about this, of course, and the somewhat ragged nature of the AR-3 woofer being taken up to 1000 Hz is one reason why the 3a crossed over so much lower.

The part that makes me raise my eyebrows just a little is that AR doesn't explicitly state in this data that the woofers are operating into their low-pass filters, which they obviously are. Actually, AR kind of implies the exact opposite, that these are wide-open, driver-only curves by saying, "On-axis output, unretouched, automatically-recorded trace." Yes it is, perfectly unretouched--but operating into its low-pass filter.

The thing that makes this just ever-so-slightly misleading and/or confusing to the reader is that the midrange curves for both the AR-3 and 3a are obviously not operating into their low-pass filters, because both midrange units' FR curves go well up past 10kHz, far past their x-o points. See the attached curves. Again, the 3 is on top; the 3a is the lower.

So the 12" woofers are shown operating into their filters, but the midranges are shown wide-open. Just a little inconsistent. I have a feeling that AR wanted to show a really "neat and tidy" 12" woofer curve for the 3a, which is why its curve shows the perfectly well-controlled roll-off introduced by its low-pass filter.

I've been on the inside of several very "major" speaker companies for a long time and I know the thinking (Heck, I was the cause of some of the thinking!). These woofer curves have been published for 45 years and no one has caught this or commented on this before. But it struck me right away.

Steve F.

post-100522-0-94121100-1405460203_thumb.

post-100522-0-58916900-1405460227_thumb.

Link to comment
Share on other sites

Steve, You 're right. Woofer curves are obviously with the woofers operating throw their low-pass filters. It could be interesting to confront also AR 3 / 3a midrange and tweeter curves , operating without and into their filters: as You know, I think AR 3's tweeter has a more extended and linear response than Ar 3a's, which has an apparent roll-off on extreme highs. And no, David, crossovers aren't identical , first because AR 3a has lower crossover point, then AR 3 has a simpler circuitatation. It doesn't mean that one is better than other, in fact I think the better is the loudspeaker, the simpler will be the crossover.

Link to comment
Share on other sites

My point was that if the woofers were identical then the xovers might explain the difference in the curve.

However, I'm no expert on this subject.

For a long time now I've wanted to start a thread and gather opinions as to which speaker people think is the better: The AR3 or the AR3a.

Link to comment
Share on other sites

The thing that makes this just ever-so-slightly misleading and/or confusing to the reader is that the midrange curves for both the AR-3 and 3a are obviously not operating into their low-pass filters, because both midrange units' FR curves go well up past 10kHz, far past their x-o points. See the attached curves. Again, the 3 is on top; the 3a is the lower.

Steve F.

Hi Steve,

I believe it is entirely possible the mids are operating into their filters, primarily because they really don't have much (in some cases, nothing) in the way of electrical filters limiting their output at their highest frequencies. In the 3a there is only a series .044 mh coil. In the AR-3 there is absolutely nothing or a .06mh coil, depending on the version.

There are folks out there listening to these old beasts who are not aware their tweeters stopped functioning years ago. :)

Compared to most modern designs, the crossover points and slopes were much more dependent on AR's proprietary drivers' mechanical characteristics rather than electrical filters.

Roy

Link to comment
Share on other sites

I' ve ever had the idea that AR 3a's tweeter made a gentle roll-off since I was a boy of fourteen ( 1974 ) and the tweeters are new ando so my ears.And the same idea I have now. The tweeters of my two AR 3 pairs, surely older, are working correctly, when i recapped crossovers I tested every loudspeaker with a battery just to make sure they works .But probably my impression is dued to the different tonal balance midrange/tweeter between AR 3 / 3a.

Link to comment
Share on other sites

My point was that if the woofers were identical then the xovers might explain the difference in the curve.

However, I'm no expert on this subject.

For a long time now I've wanted to start a thread and gather opinions as to which speaker people think is the better: The AR3 or the AR3a.

In my opinion, AR 3a is a more " dynamic " speaker, while AR 3 has a sort of electrostatic ethereal mid-high range, precise, selective, a bit chilly but very fascinating . Human voices are incredibly marvelous in AR 3a, and so a great classical orchestra. Strings are unbelievably real and " woody " . In this moment I' m listening " The Alchymist " by G.F. Haendel directed by Cristopher Hogwood, and believe me, it's almost identical to the real I listened some years ago in the Auditorium of Rome. In that occasion also my wife , which is unable to discern radios from hi-fi loudspeakers, told me " it sounds just like your AR ".
Link to comment
Share on other sites

Roy,

Perhaps the mids are operating into their networks, but that was not really my point.

My concern here are the woofer curves:

Two identical woofers (late 3 and early 3a), two "unretouched automatically recorded traces," yet two completely different FRs above 600 Hz.

AR implies that these are wide-open, driver-only curves, but it is very obvious that the 3a woofer curve is taken with the LP section in place. Perhaps the 3's is also, but certainly, the 3a's woofer curve is with the LP network in place. That's why the 3a curve shows such a marked "neat, tidy" rolloff above 600 Hz and the 3's curve doesn't.

I just thought it was amusing that AR kind of implies that there is no network in place for any of these curves, but it's obvious that there is for the 3a's woofer curve.

Did AR think no one would notice the difference between the 3 and 3a's woofer FR above 600 Hz even though the two woofers are identical?

Apparently, they did. But as I said, I understand the thinking.

Steve F.

Link to comment
Share on other sites

I completely agree, Steve. There is no way the woofer curves make sense unless the crossovers are included. With crossovers included the curves make alot of sense, as the AR-3 has a very simple, much less restrictive filter on the woofer compared to the 3a.

Roy

Link to comment
Share on other sites

The treble of all of the classic AR speakers IMO was dreadful. Rolled off, muffled, not clear or distinct. In experiments directly comparing AR3 with KLH Model 17 AR3 had far superior deep bass but above that KLH Model 17 blew it away. In showroom after showroom, in experiments at school, it was always the same. How? Why? And then a strange thing happened. I heard two LvR deomos of AR3 and remarkably they sounded a lot like actual musical instruments. How was it done? Well I never found out until a few years ago on this site. And then experimenting with AR2ax. The tonal balance of these speakers as manufactured is awful. The bass overpowers everything. Villchur tried to justify the high end rolloff using the same faulty logic BBC used 20 years later. Restoring AR2ax, amazingly with properly functioning tweeters revealed that when carefully equalized these speakers sound very accurate. They beat KLH Model 6 every way there is. Dispersion of the AR 3/4" tweeter is superb. There is no audible difference on axis and 45 degrees off axis. Pushed hard the bass of AR2ax is also amazing. Personally I like the sound of the original cloth surrounds with ferrite magnets and the cast frame better than the newer versions. But the square magnet AR9 woofers in my version show this beast is a killer. No other speaker I ever heard seems able to dominate a room like AR9 if you let it. It is a very strange effect. Cutting the bass back makes it sound more like the original ferrite version. Unfortunately IMO the treble is still poor. It had to be substantially "augmented" with additional tweeters and the entire system rebalanced. When properly tuned, it's hard to beat for accuracy. Unfortunately I cannot convert it to a direct/reflecting system. Too much time and effort. But it can be made to have the timbre much like the Steinway piano at the other end of the room...but like other speakers, not the size or power of what is far closer to an omnidirectional radiator which is different from loudness. It can also emulate the sound of outstanding famous violins made in Cremona Italy very well, sounds I'm also familiar with. The AR 12" woofer works best as a subwoofer cut off at 200 hz. In AR9 Acoustic Research finally gave up and conceded that they cannot get 3 drivers having typical usable ranges of 2 1/2 to 3 octaves cover the full 10 octaves of sound. No one else can either at least without extraordinary effort.

Link to comment
Share on other sites

Soundminded,

This is an excellent summation of your observations of the older AR speakers, coupled with your impressions of the AR9 and your modifications of it. I also remember reading and enjoying your original write-up on the restored 2ax's and how they sounded to you when EQ'd to your liking.

But your response has nothing to do with my original topic, which was the seeming inconsistency of AR's published 12" woofer FR curves and the fact that apparently both the AR-3 and -3a woofer curves (certainly the 3a's woofer curve at least) are shown operating into their LP filters even though AR doesn't explicitly say so.

My feeling is that AR wanted to show a very "good-looking" woofer FR curve for the 3a, which is why it's shown with the LP filter (even though AR doesn't say so).

But the late 3 and early 3a woofers are identical, so if these were really driver-only curves in a cabinet, they'd be identical. Identical. AR's intentional use of the word "unretouched" kind of vaguely implies that there is no network filter in play, even though there obviously is with the 3a. I know how speaker companies think, and how their Eng and Mktg depts interact.

"Publish the 3a woofer curve with the LP filter engaged."

"No. That's deceptive, and besides, we didn't do that for the 3."

"It's not deceptive. That's how the 3a's woofer actually responds in real life, because in real life, it goes through its LP filer. Besides, it'll look so much better than the 3's curve, and we want to promote our latest speaker as being the best."

"Ok, ok."

Since you are a well-informed and veteran contributor to these pages, we'd be interested in your reactions to this specific issue.

Steve F.

Link to comment
Share on other sites

I have a question: AR 3a ferrite woofer without filter has the same response of AR 3 Alnico ? In other words, the 500 Hz cutting is only the effect of the crossover network ?

Link to comment
Share on other sites

This has been something I've noticed for 45 years or so.

The late AR-3/early AR-3a woofers were identical, right? Or at least, very nearly so. Certainly from a functional standpoint, if not absolutely identical, realistically identical, to the point that they could be drop-in replacements for each other.

Therefore, their published FR curves should be identical also. But the curves published by AR are certainly not identical. Unless AR held back a little info that would explain the differences.

See the attached graphs. In the (upper) AR-3 woofer FR, it goes all the way out to 1000 Hz. But the AR-3a woofer FR graph rolls off well before that, and very smoothly both on-and off-axis.

I get the feeling that these woofer curves are curves that AR ran with the woofers operating into their low-pass filters. The bass end through around 400-500 Hz of the two graphs is absolutely identical, which one would expect, seeing as they're indentical units.

But at the upper end of the woofers' range, the two deviate considerably.

There is nothing wrong or nefarious about this, of course, and the somewhat ragged nature of the AR-3 woofer being taken up to 1000 Hz is one reason why the 3a crossed over so much lower.

The part that makes me raise my eyebrows just a little is that AR doesn't explicitly state in this data that the woofers are operating into their low-pass filters, which they obviously are. Actually, AR kind of implies the exact opposite, that these are wide-open, driver-only curves by saying, "On-axis output, unretouched, automatically-recorded trace." Yes it is, perfectly unretouched--but operating into its low-pass filter.

The thing that makes this just ever-so-slightly misleading and/or confusing to the reader is that the midrange curves for both the AR-3 and 3a are obviously not operating into their low-pass filters, because both midrange units' FR curves go well up past 10kHz, far past their x-o points. See the attached curves. Again, the 3 is on top; the 3a is the lower.

So the 12" woofers are shown operating into their filters, but the midranges are shown wide-open. Just a little inconsistent. I have a feeling that AR wanted to show a really "neat and tidy" 12" woofer curve for the 3a, which is why its curve shows the perfectly well-controlled roll-off introduced by its low-pass filter.

I've been on the inside of several very "major" speaker companies for a long time and I know the thinking (Heck, I was the cause of some of the thinking!). These woofer curves have been published for 45 years and no one has caught this or commented on this before. But it struck me right away.

Steve F.

I think this is a case of over-reaction to a standard measurement protocol used by AR since the AR-1 was first introduced and presented in its original and subsequent product disclosure articles, but I don't think it was intentionally misleading. For example, it is stated nowhere that AR measured raw drivers without the crossover in place, and AR did not imply, as Steve suggests, that the drivers were measured driver-only, without crossovers in place by the statement, "On-axis output, unretouched, automatically-recorded trace. I think this statement by AR is misinterpreted by Steve; it implies nothing other than to say that the traces were not modified or "hand-drawn" or altered in any way; on the contrary, the measurements were the direct input through the measurement microphone into the automatically recorded trace from a General Radio (and subsequently B&K) automatic level recorder. All of the measurements made by AR were in accordance with the original RETMA Standard SE-103, later updated to the ASA C16.4. AR went to great lengths to describe the careful—and completely repeatable—method of frequency-response and distortion measurement so that anyone could duplicate the results if the speakers were measured in accordance with known standards. Who else in the entire loudspeaker industry—before or since—did this? No one.

In the case of the AR-1 and AR-3, the Alnico woofer enters its natural roll-off around the 1kHz crossover—and that original curve basically shows the response of the AR-1W—but of course that same woofer is crossed over at approximately 575 Hz in the AR-3a, and AR shows the action of the filter on the top end of the response, thus not showing the extended response above it. While there is a lot of criticism about the AR 12" (10-inch radiating surface) woofer's "ragged" response up at the top end, a response of 38-1000 Hz, ±1.5 dB, on axis, is anything but that. Many an otherwise good speaker could not approach that flatness. There are off-axis losses, true, but -5dB down at 60˚ off axis at 1 kHz is not terrible; yet it could be improved, of course, and it was in the AR-3a. It is also true that the ferrite woofer, used in the AR-3a after 1969, has a different response above 500 Hz when compared with the Alnico woofer, and it actually rolls off sooner. Even the bottom-end of those two woofers is not identical, but it is close enough for all practical purposes, and all prototypes were measured flush in the ground radiating into 2π steradians. Splices of this response up to about 300 Hz were used in subsequent measurements, since AR's anechoic chamber was not echo-free below that frequency. The response curves used in the anechoic chamber were thus calibrated with the original outdoor response.

It is true that the AR-3 midrange was shown giving off full output beyond the gradual-slope 7500 Hz crossover, but originally this was described in detail by Edgar Villchur in the AR-3 disclosure article and subsequently in his paper, "Loudspeaker Testing and Measurement." In this paper, Villchur described the 2-inch midrange, "Although the on-axis curve by itself would seem to indicate response within ±1½ dB to about 15kHz, the actual radiated power is considered rolling off at 7500 Hz. Hence, this is the crossover frequency employed for a super-tweeter that is used with the midrange unit." So the idea was to show the reason for the chosen 7500 Hz crossover. Perhaps the midrange driver's output should have been shown exactly as through the crossover, but the crossover is explicitly shown and one would infer that the tweeter begins to take over at 7500 Hz. In the case of the AR-3a woofer, the midrange begins to take over at 575, etc.

So, overall, I think this is a case of stepping on gnats and letting the elephants get away. Yes, there is some inconsistency, but this is not a perfect world, and in no way was AR being intentionally deceptive or misleading.

—Tom Tyson

Link to comment
Share on other sites

Tom,

I went to great lengths to explicitly state that I did not think AR was being "intentionally deceptive or misleading," but you knock down that straw man anyway.

Here's what I said:

There is nothing wrong or nefarious about this, of course

It's not deceptive. That's how the 3a's woofer actually responds in real life, because in real life, it goes through its LP filer.

AR's intentional use of the word "unretouched" kind of vaguely implies that there is no network filter in play, even though there obviously is with the 3a.

"Vaguely Implies," is what I said (not "AR intentionally mislead the reader with bogus information"), and it does sort of imply that. That, and the fact that the mids do not seem to be running into a filter, when the 3a's woofer is obviously running into a filter.

My point was that AR should have simply added the qualifying phrase, "Woofers shown with crossover in place, on-axis output, unretouched, automatically recorded trace," and then there'd have been zero confusion or room for question.

But they didn't.

Intentionally misleading? Doubtful. That wasn't AR's style.

Lack of attention to detail? Far more likely. As I've pointed out on so many previous occasions, the Classic AR crowd weren't exactly great marketeers. After all, they managed to go from 32% US market share to left-in-the-dust in the US retail market in a span of only about 5 years.

For intelligent, informed consumers like me, who paid such close attention to everything that AR said and published back then in the late 60's/early 70's, the 12" AR-3/3a woofer FR curves delivered somewhat ambiguous information, making precise conclusions difficult. This very discussion is self-evident proof of that.

"Woofers shown with crossovers in place" would have cleared it all up.

Steve F.

Link to comment
Share on other sites

I know that you were not implying that AR was being nefarious or intentionally deceptive, but I wanted to emphasize that point. I think your statement below does misunderstand what AR was saying about their unretouched response curves, and I particularly wanted to stress what was meant by Villchur's statement on the "automatically recorded trace."

As far as AR's marketing was concerned, most of these measurements (AR-3) were made back in the Ed Villchur days prior to 1967, and understated honesty was considered to be a mainstay of the business model at AR during those years, exemplified by AR's many excellent magazine ads. But as I said, it's not a perfect world, and this is a small detail that would trip up only the most technically savvy or detail-oriented audiophile.

"The part that makes me raise my eyebrows just a little is that AR doesn't explicitly state in this data that the woofers are operating into their low-pass filters, which they obviously are. Actually, AR kind of implies the exact opposite, that these are wide-open, driver-only curves by saying, 'On-axis output, unretouched, automatically-recorded trace.' Yes it is, perfectly unretouched--but operating into its low-pass filter."

Link to comment
Share on other sites

Villchur’s polite, sophisticated ads from the late 50’s/mid 60’s are irrelevant to this discussion. This concerns technical data and its unambiguous presentation, not earlier magazine ads. The 3a was into’d in very late 1967. I first saw this tech data on my first trip to the AR Soundroom on Brattle Street in Cambridge MA in 1971. EV was long gone by then.

This statement (‘’ 'On-axis output, unretouched, automatically-recorded trace.' “) does imply that there is no extraneous circuitry or filtering applied. It doesn’t state it, it implies it. Unintentional? No doubt. Just plain sloppy? Yes, far more likely.

As I said, this statement, "Woofers shown with crossovers in place" would have cleared it all up.

This level of detail and consistency is very difficult for a company to maintain throughout all its published material, unless that company is particularly fastidious about it. AR was not, obviously.

This technical data wasn’t advertising per se; however, it can be fairly said that anything that is published by a consumer products company is “marketing material,” since it’s intended to influence a potential customer in some way on some level. AR’s lack of clarity and consistency with these FR curves was a detriment to their marketing efforts, even if it was unintentional and even if only a small percentage of the readership noticed it. I noticed; others must have as well.

AR made all kinds of mistakes and inaccuracies in their Classic lit.

The 3a and 5 w-m x-o changed from 575 to 525 and from 650 to 550 respectively, and it took the lit 2-3 years to catch up to production.

The 6 x-o started out at 1500 and went to 1800 and again, it took the lit a while to catch up.

These errors are acknowledged by AR to me in “Steve F’s Letters to AR” in the Library on this site.

AR made lots of other mistakes as well:

The LST-2 w-m x-o is somehow spec’d as “525” when that was never a Classic-era 10” w-m x-o. 525 was always a 12” w-m x-o. The 10” spec for the w-to-1.5” dome mid was 550, not 525.

The early LST-2 lit showed a 6-pos autotransformer, but in reality, it was a 3-pos switch.

And on and on and on.

The Classic AR was a great company whose products we love and respect even to this day, and whose no-nonsense, customer-friendly approach to business was unique.

But their marketing? Well….uh……

Steve F.

Link to comment
Share on other sites

"Villchur’s polite, sophisticated ads from the late 50’s/mid 60’s are irrelevant to this discussion. This concerns technical data and its unambiguous presentation, not earlier magazine ads. The 3a was into’d in very late 1967. I first saw this tech data on my first trip to the AR Soundroom on Brattle Street in Cambridge MA in 1971. EV was long gone by then."

We're beating this topic to death, but here goes one more time.

The ads were relevant, because it showed AR's—not necessarily just "polite and sophisticated"—but honest, straight-forward and unambiguous truth-telling, and I mentioned those ads as symbolic of the manner in which AR conducted business in general: honestly and truthfully. The AR-3a measurements were directly the result of the earlier AR-3 measurements, all done in accordance with the accepted standards. Those same measurements were carried over into later models up through the end of the AR Classic period.

This statement (‘’ 'On-axis output, unretouched, automatically-recorded trace.' “) does imply that there is no extraneous circuitry or filtering applied. It doesn’t state it, it implies it. Unintentional? No doubt. Just plain sloppy? Yes, far more likely.

This is nonsense in my view. There is nothing whatsoever in this statement that implies anything other than the method of measuring being done by an automatic strip-chart recorder and not retouched or hand-drawn or the like. How do you read into it anything about the filtering or extraneous circuitry? That is immaterial. If the trace were made, using the same equipment, of someone pounding on a drum, it is simply states that the trace was recorded automatically and not altered or modified. It implies nothing else, and nothing else should be read into that statement.

As I said, this statement, "Woofers shown with crossovers in place" would have cleared it all up.

This is actually the first time I think this has ever come forward as an issue in anyone's mind. I do see the point you are trying to make, but it is insignificant and to my knowledge has never been mentioned in earlier critical reviews or correspondence of which I am aware. I can't name a single reviewer among the several dozen AR reviews during this era who questioned this, or asked "were the woofers measured with or without the crossover in place," can you? There were complaints about the fact that AR chose to publish individual-driver measurements and not publish full-system frequency response measurements, but there were valid reasons for this, of course.

This level of detail and consistency is very difficult for a company to maintain throughout all its published material, unless that company is particularly fastidious about it. AR was not, obviously.

Yeah, like I said, AR wasn't perfect, and I don't know of any other company without technical mistakes and misstatements. I mean, look at the original The Advent Loudspeaker. In the original literature, the crossover was clearly stated to be 1000 Hz; however, test samples measured by AR (with component values verified) and others, showed the crossover to be ~1500 Hz or higher, so misstatements and inaccuracies were not uncommon.

I would add, too, that much of the technical information that came forward after Teledyne acquired AR—particularly after 1973 when AR moved to Norwood—began to seem a bit sloppy and more error-prone than in the Cambridge years. The one-page product data sheets that were published in Norwood were full of mistakes, and the AR-3a crossover data is a good example of that, not to mention some of the crossover points of some of the 2-way speakers towards the end.

This technical data wasn’t advertising per se; however, it can be fairly said that anything that is published by a consumer products company is “marketing material,” since it’s intended to influence a potential customer in some way on some level. AR’s lack of clarity and consistency with these FR curves was a detriment to their marketing efforts, even if it was unintentional and even if only a small percentage of the readership noticed it. I noticed; others must have as well.

I strongly disagree with this assessment. None of the response curves shown for any of the drivers deviates from what could be considered the pass band for that driver within the crossover frequencies. In other words, what was shown in the response graphs was 100% indicative of that driver's behavior within the speaker's overall response parameters. It is not as though what you got was something different from what was measured.

If this were such a big deal, or if AR had fudged the response curves to make them look smoother or more extended than they were, there would indeed be cause for alarm, but this is not the case.

AR made all kinds of mistakes and inaccuracies in their Classic lit.

Mostly those mistakes were made after AR moved to Norwood. There were not that many mistakes or inaccuracies in the Cambridge-era literature to my knowledge, other than a few gaffs here and there. Everybody made some mistakes.

The 3a and 5 w-m x-o changed from 575 to 525 and from 650 to 550 respectively, and it took the lit 2-3 years to catch up to production.

This is another case of the Norwood Teledyne AR crowd republishing data. I don't think the crossover values changed—and Roy Allison said as such—but the Norwood crowd interpreted the crossover frequencies differently from the Cambridge crowd. It is true that the Norwood crossovers used entirely different components for the AR-3a and the AR-5, but with the same basic values.

The 6 x-o started out at 1500 and went to 1800 and again, it took the lit a while to catch up.

The AR-6 did make the change, of course, because of the Series-2 version of the the crossover (the second one used a simple 6 µfd capacitor, whereas the original version used a more complex LCR network).

These errors are acknowledged by AR to me in “Steve F’s Letters to AR” in the Library on this site.

AR made lots of other mistakes as well:

The LST-2 w-m x-o is somehow spec’d as “525” when that was never a Classic-era 10” w-m x-o. 525 was always a 12” w-m x-o. The 10” spec for the w-to-1.5” dome mid was 550, not 525.

The early LST-2 lit showed a 6-pos autotransformer, but in reality, it was a 3-pos switch.

And on and on and on.

The Classic AR was a great company whose products we love and respect even to this day, and whose no-nonsense, customer-friendly approach to business was unique.

But their marketing? Well….uh……

Steve F.

Yeah, there are errors in the literature, but this is always the case (exceptions might be for companies like Hewlett-Packard with enormous marcom budgets to make sure everything was precisely correct). The errors got decidedly worse once AR transitioned from the original leaders to the Teledyne group, but there should be no cause for alarm. No one was intentionally lied to or mislead.

—Tom

Link to comment
Share on other sites

Tom,

I guess we are kind of beating things to death here, on a fanatical detailed level that either qualifies us as experts on old AR lit or totally nuts or both!

Published tech data is marketing material: it's designed to influence tech-minded consumers. That's why companies publish tech data.

As I've said now at least 6 times, AR was not being intentionally disingenuous; that wasn't their style. (I'm in complete agreement with you on this--please take 'yes' for an answer!) They were honest and straightforward. But I still maintain that the differences in the manner that the 3 and 3a woofer curves were taken (with their quite different LP filters in place) vs. how the mids were taken (where it doesn't appear that their LP filters were in place) should have been explicitly mentioned.

Stating the test conditions that were in effect when you derive your tech data is simply good practice. AR was sloppy here. Not dishonest (ok--that's 7 times!), but sloppy. This is my point: AR should have stated the test conditions ("Woofer LP filters in place") and they didn't. Sloppy. Not dishonest or deceptive. Just sloppy.

Not the way I would do it, that's for sure.

Steve F.

Link to comment
Share on other sites

But I still maintain that the differences in the manner that the 3 and 3a woofer curves were taken (with their quite different LP filters in place) vs. how the mids were taken (where it doesn't appear that their LP filters were in place) should have been explicitly mentioned.

Steve, If you are still using the mids' curves as part of the basis for your argument regardng the woofers, it is a weak one, imo. The mids' curves ARE shown into their filters, just as the woofers are! The mids simply do not have much of anything in the way of "filters". In that sense, I believe AR's curves are indeed "consistent" for ALL drivers in question. They are just model dependent.

I won't get into the debate as to whether AR's wording was sloppy or not, but I do believe nobody else on the planet ever had this discussion. :rolleyes:

Roy

Link to comment
Share on other sites

Isn't all this discussion about the curves 'much ado about' almost nothing? What really matters is how the woofer, mid and tweeter performed in their final resting places in the cabinet's BB. To my knowledge, 'in box' FR curves for those speakers were never published by AR as promotional advertising.

Today's simmers who build their own speakers take FR measurements with their drivers in their final locations in order to generate computer models with optimized xovers that turn out to be remarkably accurate. Understandably, those models didn't exist back in the day, so AR did the best they could with what they had to sell their product.

Link to comment
Share on other sites

Steve, If you are still using the mids' curves as part of the basis for your argument regardng the woofers, it is a weak one, imo. The mids' curves ARE shown into their filters, just as the woofers are! The mids simply do not have much of anything in the way of "filters". In that sense, I believe AR's curves are indeed "consistent" for ALL drivers in question. They are just model dependent.

I won't get into the debate as to whether AR's wording was sloppy or not, but I do believe nobody else on the planet ever had this discussion.

Yes, this is a pretty subtle discussion, but if it's bothered me for 45 years, then surely, some other detail-obsessed AR aficionados over the years must've felt the same way.

All AR had to do, in my view, was include the wrods, "All curves shown with filters in place."

The 3 and 3a curves are so different at their high end that an explanation would have been nice. As to the mids, way back in '67 I certainly didn't have access to AR's x-o schematics, so I didn't know that the curves showed the mids operating into what was in reality a very minimal LP filter. It looked to me, instead, that they were operating flat out, since they were responding so far past their x-o points.

The 3 and 3a woofer curves being so different, the 3a's curve obviously being filtered, and the mids appearing to be flat-out, it all combined to raise these questions in my mind, that's all.

Remember, there was no Forum in 1967-71 to give everyone all this insight (like how minimal the mids' LP filtering really was), so I just thought--at the time-- that AR's driver curves were a bit confusing and inconsistent.

Steve F.

Link to comment
Share on other sites

Remember, there was no Forum in 1967-71 to give everyone all this insight (like how minimal the mids' LP filtering really was), so I just thought--at the time-- that AR's driver curves were a bit confusing and inconsistent.

Steve F.

...and this forum is certainly the best place for such things to be discussed today.

Actually, to get somewhat more technical, I doubt the published curves were absolutely accurate for all versions of each model. For example, there were 3 different mids used in the AR-3 and, as I mentioned earlier, in some iterations of this model they are operating flat out, with no filter at all through the high frequencies. (...and the AR-3a mid, of course, was different from the AR-3 mid.)

I have been working with all of these mids quite a bit lately, and the differing response characteristics of the various midrange drivers can easily be heard without any measurements at all (out of cabinet)...so if AR had provided curves for each of the AR-3 mids without the crossover, it probably would have fallen into the category of "too much information"...certainly confusing at best.

Given driver changes along the way in all AR models of this era, as well as crossover modifications, I'm willing to bet a single published curve was not precisely accurate for ALL iterations of each model. Otoh, they were close enough to provide a general picture of what was going on.

Roy

Link to comment
Share on other sites

Isn't all this discussion about the curves 'much ado about' almost nothing? What really matters is how the woofer, mid and tweeter performed in their final resting places in the cabinet's BB. To my knowledge, 'in box' FR curves for those speakers were never published by AR as promotional advertising.

Today's simmers who build their own speakers take FR measurements with their drivers in their final locations in order to generate computer models with optimized xovers that turn out to be remarkably accurate. Understandably, those models didn't exist back in the day, so AR did the best they could with what they had to sell their product.

Carl,

You make a good point, but do you mean that AR might have better served the cause by publishing a full-system, 1-meter frequency response? Perhaps, but what exactly would that show to a listener? It will show interference effects, diffraction, lobing, driver interaction and so forth on axis, but these artifacts are very hard to detect back in a listening room and come and go as you move the microphone to different positions in front of the speaker. If you were seated at a distance of 39 inches in front of a speaker, this measurement would perfectly correlate with the sound of that speaker, but in reality it tells precious little about the total energy radiated by that speaker back into a listening roomthe age-old problem. While a 1-meter fr test of a speaker system is very important in the process of optimizing speaker parameters, crossover, etc., it will not give a true picture of what the speaker will be doing back in the reverberant field, where most people listen. It is very true that gated measurements and computer modeling today enable designers (and DIYers) to quickly optimize frequency response of a speaker, a process that was labor-intensive in the anechoic chamber years ago, but this still doesn't tell the designer how the speaker will sound with regard to its overall radiation pattern. The best way to do this is to have an integrated-frequency response of the total output of a speaker, and this is still difficult to do even today.

AR did, in fact, make many thousands of system frequency-response measurements, but none was published to demonstrate how the speaker would respond to an electrical signal simply because it will not give even close to a true picture of what one hears in a real room. AR's individual-driver curves were shown to demonstrate the performance on- and off-axis of each individual driver to show the overall performance characteristics of each driver. Roy Allison did splice some individual-driver curves into a full-system response, with disclaimers, to show how the speaker integrates each driver. This clearly demonstrated the downward tilt of the high frequencies, and it was described. More important, though, were the acoustic-power response curves done at the time in AR's reverberant test chamber. This gave a better idea of how the speaker would sound in a real room, and these power-response curves show a very smooth extension well into the highest frequencies, with the expected downward tilt at the highest frequencies.

AR speakers are probably among the most-measured speakers of all times, considering the early history of Ed Villchur's interest in objective-test measurement procedures. But AR felt that a speaker-system frequency-response measurementwhile telling the speaker engineer a great deal about consistency and so forthdid not tell much about how the speaker will sound back in a typical listening room. The acoustic-power test did effectively tell the story of the speaker's radiation pattern and its behavior in a room, and acoustic-power response curves were extensively published. In the late 1970s, AR did begin to publish anechoic rta measurements of speakers (one of the first was the AR9), and these measurements gave a better idea of how the speaker would behave back in the far field.

—Tom

Link to comment
Share on other sites

Archived

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

×
×
  • Create New...