Pete B

https://audiokarma.org/forums/index.php?threads/♫♫-april-2nd-2022-spring-frankenfest-in-west-boylston-ma-♫♫.987408/ Are any AR enthusiasts or even CSP members planning to go this April? Pete Basel

Starting on page 96 in the .pdf: https://worldradiohistory.com/Archive-All-Audio/Archive-Audio/80s/Audio-1985-01.pdf Note that the larger box provides a 30 Hz Fc.

I would do it, just make sure that the EPI drivers mount in such a way that works with the grilles.

Yes directly wired and 12 dB/oct in the miniDSP with baffle step. I've also tried 24dB/oct but this is another application, not trying to be similar to the AR s - yet. If I get around to an AR like design I'll try to do it passive.

It does not have the truncated sides so I don't think it will fit in any of the AR-11 type boxes, but it does fit in a Large Advent utility cabinet that is about the same volume. Might fit in an AR-9 type box but I don't recall if there is any sort of restriction on the diameter. It provides the correct Fc in the same box volume, is 4 ohms, and I'd call it an acoustic suspension woofer. I've used it with a miniDSP crossover and have not determined if component changes are required. It is also a 4 ohm woofer. I'd estimate that it provides at least 3 dB more MaxSPL output due to higher Xmax and a bumped back plate.

I'm fairly certain that is Steve F on here so not really an outside source. If you read Steve Feinstein's bio, you'll see mention of working at BA and Atlantic Technologies, just as Steve F has mentioned here. And I've found a 12" that performs very similarly to the AR woofer in the same enclosure volume. I'll try to demo it at the next Frankenfest.

By the way, I studied that curve over 15 years ago and simulated the crossover in LAUD pointing out to the board that the system input impedance did not simulate correctly with the 2.85 mH woofer inductor, rather something like 1.9 mH was required. We finally determined that early AR-3a s had the 1.9 mH inductor, most of the later production had the 2.85 mH. Roy pointed out that he had worked on systems with both values. It was not easy to find this thread, and note that one of the forum updates scrambled the order of old threads: https://community.classicspeakerpages.net/topic/1270-anyone-have-ar-3a-measurements-system-drivers/ Note that there is peaking at the top of the woofer's response, I believe that the 2.85 mH should eliminate that and provide 1-2 dB more baffle step.

An old thread on this topic: https://community.classicspeakerpages.net/topic/8496-ar-mgc-1/#comment-104591

Yes exactly, 12dB/oct for the midrange and 12 dB/oct for the XO, cascaded (complex multiply or add dB s). For an ideal LR it should be 6 dB down at the crossover point.

That makes a lot of sense, I've mentioned this before but worth going over again. What we now call a Linkwitz-Riley 4th order crossover is the cascade of two 2nd order Butterworth (Q=.707) filters. Some refer to them as Butterworth Squared crossovers: https://www.aes.org/e-lib/browse.cfm?elib=3287 More on LR crossovers: https://www.ranecommercial.com/legacy/note160.html It seems that the AR engineers figured this out sometime before the LR paper was published since in the AR-3a/AR-11 we see an XO around 500 Hz with the mid having an Fc also close to that. Based on your measurements the AR-91 with a 700 Hz crossover also has the mid with an Fc very close to the XO point. Closed "box" dome drivers have a 2nd order highpass electro-acoustical transfer function and tuning that properly with a 2nd order crossover can provide a nearly ideal 4th order LR XO. You measured a slightly low Qtc but keep in mind that any resistance in the crossover to the dome mid will raise Qtc.

Ken was co-author on "A Psychoacoustically Optimized Loudspeaker" published in the AES. https://www.aes.org/e-lib/browse.cfm?elib=5230 I have this in my library and studied it years ago, IIRC it was the research behind the MGC speakers. Not sure why it was not listed in the references at the end of the Audio article. Here's the Audio article in .pdf format starting on page 37 in the .pdf, might be easier to get a clean print: https://worldradiohistory.com/Archive-All-Audio/Archive-Audio/80s/Audio-1985-07.pdf

Which ones did you measure Tom? Very curious, my AR-9 s only have one working upper mid.

@JKent The 12 TPI Mixed primitive linen looks closer to the original to me, but I wonder what it looks like in person. What do you think? I also wonder how they get the weave so straight on the original grills from the factory, anyone?

It has been a while since I read the papers on this speaker, there were publications in the AES. It would be great if someone who really knows could tell us about any frequency response shaping, and delay range for the ambience speakers. A MiniDSP could probably provide what's needed but not sure about the range of delay.

I'd like to throw in a bit about what I think went on in the design of the AR-3a. The emphasis was wide dispersion through the use of a small dome mid and tweeter. Dome mids do not generally have much Xmax and as a result also low volume displacement. This results in rising distortion with lower frequency and higher output. They probably would have liked to crossover at 300 Hz but then the distortion would have been higher for the target max output level. Let's remember that the 3a dome was 1.5", very small, and an educated guess for the Xmax is 1 mm. The PSB Stratus Gold has a 5" cone (6.5" frame) with 3mm Xmax allowing for a 200 Hz crossover point - and plenty of headroom. Aadams asked why not cross lower in the AR-9, the woofer inductors are huge in the 9 and they'd have to be even larger for a lower passive XO. Ligs posted a measurement of the 3a woofer distortion and noted that it was higher at 300 Hz. I believe that the test was of the full 3a system and at 300 Hz it is likely that there is enough drive to the midrange driver to cause it to add mostly distortion to the measurement. In fact, if I am correct this supports my first point in this post. Some may say that 100 dB is well beyond the requirement for any sane listening level but note that Keele tested the PSB Stratus Gold and Wilson WATT/Puppy as providing peak outputs above 150 Hz well above 110 dB often hitting 120 dB. Well into hearing damage range but they could do it.

I enjoyed the read Tom, and the extra detail makes the reader feel like they are getting an inside perspective. Nice work, I actually printed it out in color to add to my collection.

What about this one? Is there a match?

Those are a mess but with your skills I bet they will be looking fine in the end! Your 901 project was amazing!

Good to see you on here Tom, I see only one AR-2 with the rectangular cutout, is there another somewhere?

How close is the AR-11 to the 3a, anyone heard them side by side?

Here's a .pdf of the October 1960 High Fidelity magazine with the AR-3 review, a bit easier to read, starting on pg. 55 in the .pdf: https://worldradiohistory.com/Archive-All-Audio/Archive-High-Fidelity/60s/High-Fidelity-1960-Oct.pdf

Info: https://audiokarma.org/forums/index.php?threads/♫-post-pandemic-frankenfest-is-on-in-west-boylston-ma-saturday-october-23-2021-♫.958389/

The 8 ohm and 4 to 10 uF capacitor are in series across the woofers, the cap impedance goes down when the woofer impedance goes up at high frequencies. It flattens the rise to about an 8 ohm load at high frequencies so that there is no inductive "kick back" from the woofers. Only do this when biamping.

Fantastic job on those! I'd like to try the Simple Green cleaning, is the procedure post on the forum? Pete B.

Driving a woofer alone is a very inductive load and a zobel across it makes a lot of sense. It does not have to be a perfect match, I'd put 8 ohms, 50W in series with 4uF across the woofers. Was it the woofer or tweeter amp that failed. The other issue is a very light load allows that amp to go deeper into clipping (slowing down BJT output devices that can result in cross-conduction). Any load will help avoid this so 22 ohms 50W across the mid-tweeter might help in that regard. Was the system driven hard when it failed? Was it a 400 or 700 amp that failed? There was a post on DIYaudio where a top name in audio was testing a few amps with long (20ft?) speaker cables into speaker loads and two of them where showing oscillations due to the tweeter inductance. .1uf in series with 10 ohms cured it and should probably be used in all speakers to avoid such issues.