The AR-3a Woofer-Coil Mystery -- A Solution

[tysontom]

There has been a great deal of discussion about the #7 1.88 mH choke vs. the #9 2.85 mH choke in the AR-3a crossover. Many observant contributors, notably John, Roy, Bret to name a few, insisted that some early AR-3as had the #7 choke (customarily used in the AR-2ax) rather than the #9 choke. From the beginning there was not a good explanation for this observation, and all the literature and drawings I have show *only* the #9 choke. I certainly had doubts that the AR crossover had never been changed, but to be sure I asked Roy Allison on probably three or four occasions (and I reported his comments to this forum) to reaffirm the notion that the crossover had never been changed.

Last night, while arguing (for the last time) with another forum member, I happened to notice an unpublished internal AR document written by Roy Allison (regarding the AR-3a) in 1970 that indicated that, “…there is only one aspect of these curves (describing *total* energy output curves) that seems to be cause for concern: an excess of output in the 400- to 500-Hz region. Other tests showed that this was a genuine small energy peak, and subsequently we made changes in the crossover network to reduce the output in the range by 2 dB.”

So I contacted dear Mr. Allison yet again, and he realized that he had forgotten the change to the crossover due to that response problem. He said that that there was an increase in the series-choke value. Most likely this is the #7-to-#9 choke issue; therefore, what John, Roy, Bret and others have noticed in the early series AR-3a is correct and intentional. The changeover to the #9 coil was to remedy a frequency-response anomaly (2 dB) with AR-3as that were produced somewhere after 1967 and before 1970, the date of the AR document. Therefore, as John, Roy and others have said, the #9 coil is the correct coil with all units, but the #7 coil was definitely in evidence in the early models. A day doesn't pass that we don't learn something new!

--Tom Tyson

[Guest Bret]

Hallelujah, Amen.

Thank you, Tom. It takes a big man to stick to his guns when he knows he’s right and an even bigger man to provide the documentary evidence that maybe he was just not quite as right as he thought he had been.

I, for one, have no problem saying publically that this proves every theory I had about why the inductors changed is wrong. In fact, they seem downright silly under the circumstances.

I’m glad I was wrong because this explanation is so much less destructive to my opinion of AR and its engineers. In fact, it is a positive boon. A minor problem (2db is pretty minor, if audible) was found and in an on-going effort to refine the product in keeping with a great tradition, it was fixed. Excellent. Perfect. Bravo!

Praise too should go to Mr. Allison for not being so afraid of ever having been wrong that he couldn’t admit the problem and get it fixed. Many would have said, “No, the original design is right!” or “So what?” or even, "Close enough." Who can blame him for a minor memory lapse from 35+ years ago? Not me.

This also shows that some of us have pretty extraordinary ability to hear, or we have learned to listen well, as the problem Mr. Allison identified is *precisely* what John has said he heard and later, I believe, measured. I suppose when you have an example of “right” it is possible to hear exactly what is wrong.

If it hadn’t been for Roy’s dedication to restoring so many of these classics and for his purchase of a woofer tester, and if not for John’s getting one of each to be able to do comparison listening and tests, the hard way, we might never have even realized the problem existed. Lastly, if it hadn’t been for arguing with PeteB, the solution might not have seemed so urgent. It's all been productive.

Ken was right when he talked about this place as a community and the . . . how did he put it; The collective brain-trust? THIS, friends, was quite a "poser."

Tom, I think this information belongs, somehow, prominently in the Forum Library. Thank goodness Mark has provided us with the means to distribute the information you and others worked so hard to find.

Three cheers for several people! Hip-hip. . .

[soundminded]

...and I know for most it will be a minor, even trivial point for most but I have to say it anyway. I wish they would have given the altered design a different designation such as AR3ax or AR3b. Why? So that they could be distinguished from one another and so that owners of the earlier version would have been aware of the change and had the opportunity to "upgrade" to the flatter response. Yes I know it also takes a big company to fess up to even this minor point but 2db is audible and contributed to the characteristic sound signiture that gave AR3a a distinctive personality. In a later day world this kind of change without a corresponding change in model designation would not have lived up to the spirit or letter of ISO 9000 which demands tight conformance to a prototype. Yes it is a minor point and only a handful of those who own these speakers and read this thread will rip open their speakers to find out which version they have. It still takes nothing away from the fact that it was a landmark loudspeaker and a remarkable achievement.

[Steve F]

This is great information Tom. Thanks for digging through it all. Since the new larger coil was used to attenuate the top end of the 3a’s woofer response, I’d bet that this was what AR meant when they answered in response to my 1974 inquiry as to the origin of the 3a x-o change from 575Hz to 525Hz.

AR stated to me that the change in specs was made to "reflect a change that had been made several years ago."

"Several years" before my 1974 question would certainly coincide with the time frame of the new coil.

I wonder if an analogous coil change was made in the AR-5, hence the change in the 5’s x-o from 650Hz to 550Hz?

Re: the non-model number change of the 3a—

All companies, in any industry, make continual small upgrades in product performance without introducing it as a "new" model every time. Should AR have done so? Maybe yes, maybe no, but there’s no doubt that the "3a" name had considerable marketplace acceptance by then and AR obviously didn’t want to take a chance with it. Remember, they made wholesale changes to the 2ax and still called it the 2ax.

Also, during the 1954-1974 Classic era, AR’s model number changes followed a very strict, consistent pattern:

"x" denoted a change concerning a cone driver (1 to 1x, 2 to 2x, 2a to 2ax, 4 to 4x)

"a’ denoted a change concerning a dome driver (3 to 3a, 2 to 2a)

This pattern held nicely up until 1974 and the 4x to 4xa, which did not involve a dome driver.

The coil change to the 3a did not involve any drivers at all. "3ax" or "3b" did not fit into their model number scheme at the time. This is not to say that they were correct from a marketing standpoint in not updating the model number; it’s merely an observation as to why they didn’t.

Steve F.

[soundminded]

"All companies, in any industry, make continual small upgrades in product performance without introducing it as a "new" model every time."

You're right of course Steve. Although some will change a last letter such as a -G, -H etc. I had several seemingly identical model Fedders air conditioners bought at different times with that kind of change to the designation of the model number.

Other manufacturers were much worse. JBL says of its D44000 Ranger Paragon on the Altec Heritage web site that of the approximately 1000 made, no two were exactly alike. The cabinets had no interchangable parts and they kept changing drivers. One version even had a built in amplifier (although that was specified as an option.)

The problem is that looking back at it years later, short of taking it apart and comparing it to drawings and specifications and talking to people who were there and testing their memories as we do here, you don't always know exactly what you've go. (You should have seen the disaster the GE tech had trying to install the replacement controls for my electric oven that the factory sent. He needed technical backup from his office and the unit was only 4 years old. They no longer even made the original controls and had run out of exact replacements.)

[charger3834]

Fantastic, everything fits now. This whole thing was a worthwhile endeavor.

[Carlspeak]

>

>

>

>

>........"The changeover to the #9 coil was to remedy a

>frequency-response anomaly (2 dB) with AR-3as that were

>produced somewhere after 1967 and before 1970, the date of the

>AR document........."

>

>--Tom Tyson

>

Wow, I'm a bit surprised but also disapointed that all the quibbling over the No. 7 vs the No. 9 woofer coil has boiled down to a 2 dB anomaly. Yes, I agree this is audibly noticeable, but just barely. I would think in the grand scheme of the audio spectrum produced during normal music playback that this wouldn't be noticed by the average listener. However, AR engineers must have noticed it during testing and decided to make the change.

BTW, I was glad to see the unusually large volume of AR posts tonight. Could it have something to do with Bush's speech?

Carl

[RoyC]

Tom,

Thank you (big time!) for solving the coil mystery, and assistance in helping me regain my sanity. In recent weeks, regardless of what I would choose to listen to, the AR-3a's would ONLY play the theme from The Twilight Zone! I was getting concerned.

Thanks again,

Roy

[RoyC]

Hi Carl,

It is not until a ceramic magnet woofer is installed in a cabinet with the #7 coil that the issue becomes annoying to my ear. Conversely the alnico works quite well in a #9 coil cabinet..

Another more obvious change along the way, was that the early AR-3a's are equipped with a slightly different style midrange without the center disk, in other words no "pupil" in the eyeball. There are probably some measurable differences between them as well.

(See attached photos of a 1968 and a 1975 midrange.)

It appears that subtle refinements/modifications were being made to all models. That info was proprietary, and as we have seen not very easy to nail down.

It would be prudent for those fixing these old beasts to take measurements and keep notes to ensure faithful restorations.

Roy