[Evolution of and Replacements for the AR 11/12" Woofer]

>As we know, AR was a true pioneer in the development of many

>innovations and technological advances in the science of sound

>reproduction. But even AR was not afraid to use the best

>conventional methods of the day, and anodized-aluminum voice

>coils certainly falls into that category. It’s not really an

>example of “prior art,” which refers principally to patents,

>but to a process resident in the public domain for many years.

> This is in contrast to Thiel and PSB (that you mentioned),

>companies that produce physically beautiful loudspeakers, but

>who have contributed little to the science of sound

>reproduction. Are there any significant contributions from

>these companies that have affected the entire speaker

>industry?

Hi Tom,

I'm fairly sure that black anodized aluminum former voice coils are common today in some of the better woofers but I don't think they were common that far back. In fact, I seem to be the first reporting it in an AR woofer, most others are not anodized. I've not done extensive research so I'm not going to get into a back and forth. I referred to it as prior art because most ideas such as this, as obvious and trivial as they might be, become patented. I have some experience with the patent system and it is a mess, it is pathetic but I won't get into the details.

Several here have commented on innovation and significant contributions to the art/science which is fine but not a requirement for every company producing speakers, from my perspective. I'm able to admire a company that simply applies the well known concepts of the day and produces a top notch product, especially at a fair price. PSB speakers test very well, have been praised by both the objectivists (Keele) and the subjectivists (several at Stereophile), and they sound outstanding for their price point to my ear.

I was referring to Thiele in the sense of Thiele and Small driver parameters. OK you made me look up the spelling, it is Thiele and I apologize to Mr. Thiele as I admire his work very much.

Concerning Theil speakers, they've certainly perfected first order crossover design and system designs that attempt to preserve wave shape. They've also perfected high performance coaxial speakers. These are not high priorities from my perspective.

Thanks for the mention of the source material, I think I have a few of those but do not play them often.

>>Many here talk of specs and the flat response of AR speakers.

>

>>The first version of the Stratus Golds were rated at +/- 1

>dB

>>from 36 Hz to 20 kHz on axis. This is hard to believe and

>>Keele was only able to confirm +/- 1.5 dB from 40 to 20 K,

>>still not bad and Keele does outdoor tests with spliced near

>>field woofer response if I remember correctly, so there's

>some

>>room for error. Both specs are certainly outstanding.

>>Pete B.

>>

>

>Pete, the spec of +/- 1dB from 36 Hz to 20 kHz, *on-axis,*

>does not impress me in the least. It is meaningless in the

>grand scheme of things. The speaker you mention is

>directional at most mid-and-high frequencies, and I believe

>has steep 24 dB crossover slopes. This will produce very flat

>on-axis response if the drivers are appropriately

>high-quality. A speaker such as this also has a relatively

>focused energy output, and does not have good integrated-power

>response. It might be very flat, on-axis, in the near field,

>but it will sound increasingly *dull* the further back in the

>listening environment you get -– even on the speaker’s axis.

>Placement of the speakers and the listening position is

>therefore highly critical for good results, and it allows only

>one person at a time to truly get into the “sweet spot,” and

>other listeners have to wait their turn, or are likely to get

>a different impression of the performance of that speaker. In

>short, the speaker takes on different personalities when one

>moves around the room. If one seeks that type of listening,

>why not simply listen through good-quality headphones?

A specification of +/-1 dB from flat and not a prototype response, over that wide a range for a production speaker is mighty impressive IMO. Many here are having trouble getting their -3dB point below 40 Hz for the AR 11/12" woofer. Few unequalized systems even offer the bass extension. It's interesting that 4th order LR crossover is claimed for the mid to tweeter XO in the PSBs and it is obtained with 2nd order electrical networks in combination with the driver's acoustical response. I believe they're approximately 4th order LR. The mid to woofer is claimed to be 3rd order, IIRC but looks to me to be approximately 4th order also. It is well known that 4th order LR crossovers do not have flat power response, however I never believed AR's claim of the importance of flat power reponse, since I first heard them and found them lacking.

I'm certainly familiar with the power response issue as it relates to crossover design and it has been covered for probably more than 30 years in the technical journals. I'm familiar with most of these papers and don't believe that a passive higher order network has been covered that provides both flat on axis and flat power response. It appears, based on looking at AR's published specs that some driver overlap is used to improve the power reponse, however I don't believe this provides flat *summed* on axis response. I also do not believe that this crossover with overlap and the goal of flat power response has been covered in the journals. There are many experts that place on axis response as the most important parameter, and I agree with them. This of course assumes that a reasonable crossover type is chosen so that the off axis response does not suffer significantly.

Your discussion of near field (using this term loosly) as compared to other listening positions might apply to much larger rooms but I would not agree with your analysis in the typical home environment. Reflections from the boundaries around the speakers in the typical sized home environment are far too early to improve the sound in any way, indeed these early reflections tend to blur the early arrival. This is why the front of the NHT 3.3 is aimed in toward the listener to reduce side wall reflections, it is also why the signal is electronically delayed in the AR systems that have reflective firing augmentation speakers. We may also note that the design of the 3.3 forces the main drivers to be far out from the wall behind them which also reduces early reflections.

I've found in practice just the opposite of your theories claimed above, that systems with higher order in-phase crossovers sound good in most listening postitions. They have excellent polar response and the sound does not vary significantly in seated or standing positions, or over a wide angle in the horizontal plane. This is in contrast to first order systems that typically have issues with seated versus standing positions, not due to power response issues but rather their inferior polar response with non-coincident drivers. I don't believe that the non-flat power response to the degree seen with LR crossovers is an issue in typical home listening environments. We may just have to agree to disagree.

>Like it or not, we are influenced by the room boundaries and

>reflections, which are a balance of the *sum* of direct and

>reverberant fields. Even Baranek’s calculations imply that

>listeners are almost always in the reverberant field of the

>room for all frequencies being reproduced, and a

>wide-dispersion loudspeaker is necessary to support uniform

>and flat response in the reverberant field. Most of the AR

>speakers we mention in the forum have wide dispersion, and

>provide good integrated-power response in the typical

>listening room. Yes, we “talk of specs and the flat response

>of AR speakers,” but the reference is made primarily to the

>integrated-power energy of these speakers, which in the real

>world is more important than the simple in-situ on-axis

>frequency response.

Again with your points here we may just have to agree to disagree.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

It's been said that most good ideas have already been done, and I'd have to say that this AR woofer is an example of prior art with regards to my thoughts on black anodizing, I didn't know about it at the time.

I've usually seen aluminized paper used in smaller drivers and I'd think that a thin layer of aluminum on those Nomex formers would help quite a bit as far as heat sinking goes. On the other hand solid metal formers have the advantage that both sides serve as heat radiating area. One problem with pure aluminum is the difference in the coefficient of thermal expansion as compared to the copper windings, it does seem that they've gotten around this since aluminum is common these days. This is one reason that I like bronze type formers.

I'm curious to know what the source material was used for your drum test, live recordings often have a lot of subsonic material that should be filtered. Vocal pops, string initial impact and so on. Were you running the active filter for the 801s? It's an interesting challenge to try to reproduce uncompressed material at realistic levels.

Keele probably half joking, referred to a system I particularly like as the 10,000 Watt speaker. The original PSB Stratus Golds will handle his short term burst test at 10,000 W above 700 Hz. He runs a bridged Crown amp to do this. One has to wonder if the voltage rating of the capacitors is exceeded in this test. This system crosses and stays above the 120 dB mark (at about 4 kW bursts) from 70 Hz on up in Max SPL capability. I don't need this SPL level but it just shows how much head room is available and the robust construction.

Many here talk of specs and the flat response of AR speakers. The first version of the Stratus Golds were rated at +/- 1 dB from 36 Hz to 20 kHz on axis. This is hard to believe and Keele was only able to confirm +/- 1.5 dB from 40 to 20 K, still not bad and Keele does outdoor tests with spliced near field woofer response if I remember correctly, so there's some room for error. Both specs are certainly outstanding. I found this hard to believe and actually measured the Thiel and Small parameters of the Stratus Gold woofer, then simulated the response, -.9 dB at 36 Hz - impressive.

They claim hand selection of inductors and caps to a 1% tolerance, resistors are 2%. There are no exotic parts but I'd expect tight quality control on drivers. I'm just offering this as a comparison to a value engineered modern design.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

I've been aware of thermal issues for a long time and even stated that voice coils should be black anodized many years ago but this is the first I've seen in practice.

Those 10Pi's are a part of audio history, obviously, I'd guess that you've got you own AR museum there.

Yes most drivers will take a lot of short term power, it's interesting to look at the sine burst tests done in the old Audio reviews by Don Keele.

Pete B.

>It appears that the late-70s designation of the 200003/1210003

>woofer (AR-9/AR-11B) you show is consistent the advent of the

>aluminum voice-coil former. The anodized-aluminum former was

>more robust (up to a point) and would likely dissipate heat

>better than the first versions. Ironically, the voice coil I

>showed with the flattened former was hit with peaks in excess

>of 1 kilowatts. That will usually do it, although the

>AR-10Pi/Neil Grover live-vs.-recorded drum demonstration by

>AR/C. Victor Campors in the mid-1970s used the ill-fated

>Dunlap-Clarke Dreadnaught 1000 amplifier driven to full power

>much of the time before failing completely. I don't believe

>that the woofers suffered any damage (I own one of the

>original pairs of 10Pis used in the demonstration, and it has

>all the original drivers according to the date codes), but

>there were some tweeters destroyed. So the woofers will

>definitely take a lot of peak power before damage.

>

>One way to help in removing the glue on the woofer flange is

>to use denatured alcohol to soften the glue, and then

>carefully scrape away. It's slow, laborious work, but a clean

>surface is mandatory for good results.

>

>--Tom Tyson

[Evolution of and Replacements for the AR 11/12" Woofer]

Hi,

Yes interesting to be able to count the turns right off the picture.

It's sort of detective work to figure out what's going on with these woofers, interesting.

>Hi there

>

>Isn't digital photography a wonderful tool?

>

>I never even spent 5 minutes really ever looking at a voice

>coil before, but with these pictures and the feedback it

>certainly makes this more interesting.

[Evolution of and Replacements for the AR 11/12" Woofer]

I quickly counted 55 (per layer) turns in the picture of the Nomex voice coil above, and 51 for the VC I'm dealing with here, I didn't double check these and I might be slightly off but here's the math using the total number of turns:

2.37 ohms * 110/102 = 2.56 ohms

The shorter coil gives the driver less ideally linear throw and slightly more before bottoming.

It seems that the windings are better packed in this driver than in the Nomex picture where fairly wide gaps can be seen in the windings, near the AR lettering in the picture for example. Tighter packing provide more wire length in the B field and therefore a higher Bl.

There's clearly an economic driver to not have to repeatedly replace drivers and I think the replacement units might have used a better voice coil.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

Here are pictures of AR 200003 woofer, date code 561 7838 (RDC of 2.37 ohms) which was damaged by bottoming, and I believe the voice coil extended past the pole piece and over the top and stuck there. This caused part of the voice coil to bend in toward the pole piece. I've been referring to this as woofer #2 from an AR-11.

Assembly rear view:

http://members.aol.com/basconsultants/AR11WCRV.jpg

Voice coil rear view, note black anodizing, vent holes, black glue on spider, rear of dust cap can be seen, it is cloth with what seems like black silicone sealer coated on the back. the back of the former is mostly bent out toward the windings from bottoming, part bent in toward the pole piece and broke off when bent back out. The voice coil now fits perfectly into the magnet assembly, there is no binding at all:

http://members.aol.com/basconsultants/AR11WVCR.jpg

Voice coil side view, note that windings are in perfect condition, there are no shorts to the former. Note the bronze color on the top of the former, is this a treatment for better adhesion to the cone and spider?

http://members.aol.com/basconsultants/AR11WVCS.jpg

Spider separation from cone, glue near cone looks like epoxy, spider is porous:

http://members.aol.com/basconsultants/AR11WSPB.jpg

Cone front view:

http://members.aol.com/basconsultants/AR11WCFV.jpg

If this was a pro driver in heavy use I'd replace the voice coil, but I want to keep this driver as close to original as possible to maintain original performance and so that I can measure it. I'll probably run a thin bead of epoxy at the end where the former is folded up. I'd expect that the woofer will tolerate slight bottoming even in this condition, and if it bottoms hard and the coils come loose that would be a good time to replace the voice coil. Are people measuring the DC resistance of these drivers? It seems from the measurements on this site that 2.5 ohms was the typical average but this driver is lower was the lower value also common? Does anyone know the wire guage and number of turns? I suppose I could count them.

I notice that the pole piece is not extended past the top plate, this is often done to reduce inductance modulation, provide a fringe field that is more similar to the lower fringe field and to prevent the voice coil from going past the top.

Woofer #1 where the cone can be moved freely has some odd characteristics from my perspective. The spider is very loose, yet the cone moves slowly as if there's some resistance, it feels like air flow and I wonder if it's air leaking slowly through the spider which is porous. This would probably act as a combined compliance/resistance element but would be non-linear since the air volume is small. Vents in the woofer frame below the spider might eliminate this effect and improve the driver.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

Thank you gentlemen for all the information. The common things that normally come to mind to remove glue are lighter fluid which is Naptha and WD-40 which has solvents and oil. Obviously work in a well ventilated area and follow the warnings if anyone tries these. I'd worry about these solvents getting into and damaging the masonite ring. Acetone probably works very well as it seems silicone sealers and glues have it. I cleaned the second woofer with Isopropyl alcohol and an old tooth brush which helped but still did not make it an easy job. I don't know what sealers are used on masonite and thought that alcohol would be safe then remembered that shellac based sealers have denatured alcohol as the solvent. The masonite looks fine in any case and probably uses a modern sealer.

It seems that both masonite rings are attached to the metal frame with clear silicone sealer/glue as it is a hard rubbery substance. The foam edge to masonite bond seems to be much more like rubber cement but of a thinner consistency, I don't know what type of glue this is. The foam edge to cone seems to be a PVA type glue.

I have a fresh bottle of the common black glue that's used for dust caps and so on and it appears to be silicone based and has an acetone smell. This seems to be what was used for the dust cap and lead in wires on the AR woofer. Theres a small amount of this black glue where the spider attaches to the voice coil, but where the spider broke away I see what looks like two part epoxy closer to the joint, which makes sense to me since a strong high temperature glue is needed here. The spider to masonite ring seems like a PVA glue. These are my observations and I welcome more input on the glue types.

Rich, did you notice if any of the woofers with aluminum voice coils were black anodized on the inside?

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

While the date code for this woofer indicates that it was manufactured in the 38th week of 1978 it does have a red tag and I wonder if the replacement woofers, or some of them, were rebuilt units.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

This was the second woofer that I've been referring to as #2 and it has the red service replacement tag, here's what I said about it:

"The second woofer (#2) was damaged when driven with excessive power (300W), and the voice coil rubs in such a way that the cone sticks wherever you leave it. ... It is marked: 561 7838 and has an RDC of 2.37 ohms."

I'll try to post those pictures soon. I'm working on cleaning off the old rotted foam and glue from the frame and masonite, anyone have suggestions for a good way to do this? Any suggestions for a solvent?

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

The foam from RSSOUND arrived today with hand written answers to my questions on the receipt. Seems they mainly do business with large customers, but also provide this semi-automated service for small customers. The foam looks like a good fit and is 5/8" wide which I believe is correct for this woofer.

I unsoldered the lead in wires on woofer #2, unglued the spider which allows the entire cone, spider, and voice coil to come out as a unit. The voice coil has a tarnished bronze look on the outside top near the cone, the appearance on the inside is black anodized, and the edge has a silver aluminum color. I believe the former is aluminum but I can't explain the bronze look on the outside. The voice coil did bend in and the windings are in perfect condition. I was able to flare the former back out and it should be fine.

I notice now that the spider is stretched at high excursions and I believe this is what causes the floppy spider that some people mention. The spider was also starting to come unglued where it meets the cone/voice coil joint. I believe this was also caused by over excursion.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

Thanks for the pictures Tom, very interesting.

This has me thinking again about this second woofer, I now believe after taking another look at it that the VC former bent in toward the pole piece when it bottomed and is what's scraping and causing it to stick. I've only seen this happen once before with an aluminum former and the VC would have been fine if it had bent out away from the pole piece. Formers should be flared out or even completely formed so that they're more tolerant of bottoming, IMO.

I don't have a lot of experience with Nomex formers but at this point I prefer the metal types that I've seen.

It would be good if we had replacement part numbers for the internal parts of these drivers, spider, VC, cap, cone, etc. so that we can specify them to driver rebuilders and get more consistent results. We could write a specification for the rebuild process calling out specific parts to be used, adhesive types to use, certification testing, and so on. This is the way to get consistent results.

NuWay has extensive offerings and online catalogs.

http://www.nuway-speaker.com/html/products.htm

I'm waiting for foam edges from RSSOUND and have not been getting much response to my emails. This happened with the Advent surrounds that I ordered and I understand that he is busy running a business, he did deliver and did answer my important questions in the end. What I like about RSSOUND is that the foam was a perfect fit and more compliant than other standard 10" foam surrounds that I've seen. I have a few other 10" drivers that I wish I'd used RSSOUND's Advent edge on. Is Rick from RSSOUND a member here?

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

Original owner tells me that one woofer was replaced, he did notice the magnet difference and the AR tech told him it was not a problem.

Pete B.

>Woofer #1 on the right in the top picture, left in the bottom

>picture is obviously larger, that is not a shadow in the

>picture. Woofer #2 has a red tag stating: "Service

>Replacement Unit For Further exchange the original purchase

>Bill of Sale required" As I understood it the woofers had

>never been replaced, does this tag indicate that it is not the

>original woofer? I'll have to check with the original owner

>since he has a good memory of what went on with these

>speakers.

[Evolution of and Replacements for the AR 11/12" Woofer]

Hi Tom,

I'll weigh them at some point if I get a chance. The former must extend past the windings somewhat as with most voice coils, do you know what that dimension is?

Thanks,

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

The plot thickens, different magnet diameters.

Tom Tyson wrote "Can you provide pictures?" I've been thinking that I should measure every dimension on each of these woofers but did not get around to do it. I did put them down side by side and what do you know, one magnet has a larger diameter, enough that it's easily seen by eye. Here are the pictures:

http://members.aol.com/basconsultants/AR11WMAG.jpg

http://members.aol.com/basconsultants/AR11WSID.jpg

Woofer #1 on the right in the top picture, left in the bottom picture is obviously larger, that is not a shadow in the picture. Woofer #2 has a red tag stating: "Service Replacement Unit For Further exchange the original purchase Bill of Sale required" As I understood it the woofers had never been replaced, does this tag indicate that it is not the original woofer? I'll have to check with the original owner since he has a good memory of what went on with these speakers.

Woofer #2 has a magnet diameter of 6.04 inches, while woofer #1 does not fit inside the caliper which stops at 6.185", I'd say it's about 6.2".

I measured the thickness again Woofer #1 is right on the money at .75" if not .751", while Woofer #2 is over .74" but closer to .74 than .75". This is starting to split hairs but clearly woofer #2 has a smaller magnet. The back plates are both 5.5".

Woofer #1 is the one that did not fail and it would tend to have less bass excursion due to the stronger damping and reduced output in the low bass.

It's also interesting to note that the woofer that seems to be original is the stronger version, when the literature states that the AR-11 should have the weaker driver. Knowing what goes on in engineering and production this doesn't surprise me - the stories I could tell.

They probably used more than one supplier for magnets. Any other theories?

I'd appreciate it if others could measure the diameter of the magnets and report back. Dimensions of the square magnet version, and the top and bottom plates would also be helpful.

Pete B.

Here's a view of the front:

http://members.aol.com/basconsultants/AR11DYNT.jpg

>--Tom Tyson

>

>Pete and Roy: have either of you found examples of the

>200003-woofers in which you have noted different Bl or magnet

>structures? I know the Bl varies a bit from woofer-to-woofer,

>but I still have not seen (or are aware of) a different magnet

>structure for this woofer series. Can you provide pictures?

[Evolution of and Replacements for the AR 11/12" Woofer]

Hi Roy,

I notice that your 71 woofer has an indentation in the round magnet but I don't know how much of a difference that might make. What is the part number for those probably not 200003?

Does anyone know if the -1 on the 200003 part number indicates the square magnet?

I didn't know that there was a square magnet version of the 200003-1, this is the first I've seen of it:

http://www.kujucev.com/ar/ar12inch_a.jpg

The square magnet and plates might make the difference in Bl from 10 to 11.75 and this supports my theory that there were differences in early and late production. Now we know there were differences, the question is if they explain the differing Bl. The square magnet made to standard dimensions might just provide a bit more material/strength.

Bret were DUT 2,3,4,7 square magnet versions by any chance?

Pete B.

>Hi Tom and Pete,

>

>The only magnet differences I've observed have been between a

>'71 version and all the rest. I've attached photos. The '78,

>'74, and '72 versions have the same dimension's as Pete's

>'78.

>

>The Tonegen's magnet actually has a smaller diameter by about

>half an inch than the others, with all other dimensions being

>the same.

>

>I believe it was Bret that noticed differences in the

>published BL specs in the 10pi.

>

>Pete, Millersound said that the '78 spiders were shot and

>replaced them. I didn't think to turn them upside down as you

>did. Maybe mine were the same as yours and I just thought they

>were "stiffer" because they were bottomed out..They did sag.

>

>Roy

>

[Evolution of and Replacements for the AR 11/12" Woofer]

Thanks for the info Roy. I believe that ceramic magnet materials have been getting better and better over the years. I noted that the Bl values were grouped around 10, 11.75, and 13 in this thread:

http://www.classicspeakerpages.net/dc/dcbo...5393&page=#5399

I then also found different gauss values for the 11 and 10pi which supports this theory in the AR literature as reported in post 5795 of this thread:

http://www.classicspeakerpages.net/dc/dcbo..._id=&page=#5797

Having worked in engineering I can tell you that things do not always go as planned, the marketing literature could be what was planned and prototyped. It's possible that they decided it was too much trouble to stock the two different drivers and went into production with a common part. However, we do see that samples seem to show different values which I believe is simply early 1974 time frame production, and late 1978 time frame where magnet materials just got a bit stronger. It's possible that around the time the 10pi and 11 were being designed and planned that they had stock of both woofers and the logical thing was to put the stronger magner version in the 10pi and weaker in the 11, until they ran out. Then they would all get the stronger version. This is pure speculation on my part but it is how things go in the real world.

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

Hi Tom,

Thanks for the info, I prefer not to shim as you also mention and this is how I've always done it. I think the longer coil causes it to bind more easily than most, and I do notice that if I apply equal pressure I can get smooth motion. I use the method you mention of testing the motion while the glue is wet.

I noticed that both cones have 3013 F stamped on the back in white. Do you know the source for the cones?

The top and bottom plates measure .50" thick but are very slightly tapered at the edges. The magnet is .74" thick.

I'll test the Bl once I get these refoamed, they should be the lower strength given that they're out of AR-11s.

Do you have a preferred source for parts such as edges and voice coils?

How about ink to restore the black cone? One is faded more than the other. Anyone know what type of ink was used originally?

The only difference between these and the ones Bret shows here:

http://www.classicspeakerpages.net/dc/user_files/305.jpg

are that these have the date code stamped on the back, have decorative black foam around the outer rim, do not have the screen, and the masonite does not go as close to the screw holes. The cone, dust cap, frame, magnet, top and bottom plates look the same.

Pete B.

>>Pete and Roy: have either of you found examples of the

>200003-woofers in which you have noted different Bl or magnet

>structures? I know the Bl varies a bit from woofer-to-woofer,

>but I still have not seen (or are aware of) a different magnet

>structure for this woofer series. Can you provide pictures?

[Evolution of and Replacements for the AR 11/12" Woofer]

Hi Roy,

I've got a pair of AR 11/12" woofers out of AR-11s from the late 70s, both part #200003 (MADE IN U.S.A.) without any dash number. These do not have the screen material over the back and they do have the small hidden masonite ring. The original owner of these 11s informs me that no work has ever been done to the woofers, and they've never been replaced.

The first woofer (#1) out of what I'll call system #1 seems in good shape and only needs replacement of the foam edge. This one is marked: 561 7830 and has an RDC of 2.49 ohms.

The second woofer (#2) was damaged when driven with excessive power (300W), and the voice coil rubs in such a way that the cone sticks wherever you leave it. I believe it may have bottomed also since it sounds as if some windings are loose. It is marked: 561 7838 and has an RDC of 2.37 ohms.

The date codes suggest that these were manufactured in the 30th and 38th weeks of 1978.

The only thing that does not seem to agree with your findings is that the spider is so loose that the weight of the cone causes it to drop when the magnet is face down. I thought the spider had sagged and lost center until I flipped the woofer over magnet up and now it sagged the other way. It just seems that the spider is so loose that it will not return the cone to center without help from the foam edge. I've not seen this in most woofers with deteriorated edges. I also note that there does not seem to be much clearance around the voice coil and it easily rubs with the edge missing. I'm thinking that it is probably important to shim the voice coil when refoaming these drivers, do others agree?

Does anyone know if the glue is water soluble to remove the dust cap, and also the spider outer edge? Does Miller use new dust caps when he refoams?

Pete B.

>Roy posted this information in another thread:

>

>The other two versions are from the late 70's (AR-9) and

>possibly early 80's (from an unknown model). The observable

>differences in these from the '74's are noticably stiffer

>spiders, a smaller hidden masonite flange ring, a (now

>deteriorated) foam facing on the flange and the absence of the

>basket screen (on the early 80's version).

>

>The Tonegen (#1210003-2A from 1994) seems to have the stiffest

>suspension of all and of course no basket screen. The two 70's

>woofers are labeled #200003 and the later one (80's?) without

>the basket screen is a #200003-1. With the exception of the

>screen, the late 70's 200003 and the later 200003-1 seem

>identical and have the same stiffer spiders. They need

>re-foamimg so I cannot comment on sound yet.

>

>

>Roy C.

[Evolution of and Replacements for the AR 11/12" Woofer]

Roy posted this information in another thread:

Here's a link:

http://www.classicspeakerpages.net/dc/dcbo...id=&page=6#5409

Here's his post:

Since my negative experience with the replacement (Tonegen) woofer in AR-3a's, I've been obsessed with the differences between the various versions of AR 12 inchers. I have collected 3 versions to date not counting the original cloth surround version and the Tonegen. The earliest from a '71 AR-3a has a very soft spider and can be identified by an easily observable masonite ring upon which the foam surround is mounted. I also have identical woofers from 1972 and 1974 AR-3a's.

The other two versions are from the late 70's (AR-9) and possibly early 80's (from an unknown model). The observable differences in these from the '74's are noticably stiffer spiders, a smaller hidden masonite flange ring, a (now deteriorated) foam facing on the flange and the absence of the basket screen (on the early 80's version).

The Tonegen (#1210003-2A from 1994) seems to have the stiffest suspension of all and of course no basket screen. The two 70's woofers are labeled #200003 and the later one (80's?) without the basket screen is a #200003-1. With the exception of the screen, the late 70's 200003 and the later 200003-1 seem identical and have the same stiffer spiders. They need re-foamimg so I cannot comment on sound yet.

Although I've done most re-foaming myself in the past, I plan to have some work done by Millersound due to his reputation and to use as a source of comparison.

Roy C.

[Advent experiment - mod - much better sound IMO]

I made a few typos when I wrote this up, here are the corrected values:

C1 was swapped between the 2.2K and 4.7K version, and the C2=.016 uF in the 6dB case should be .16 uF.

Some asked for component values for other BSC step sizes, here

they are:

I also simulated the design taking into account the load

impedance and the values are slightly different. I just used

standard values in the early prototype and the BSC center

frequency was slightly higher than the original target.

These values are for a BSC center of 400 Hz:

IN O--------- R1 --------------O OUT

|--- C1 ---| |

|

C2

| ----------> RL

|

|

R2

|

GND O---------------------------------O GND

2.2K Version, these values are for RL = 22K:

R1 = 2.2 K, C1 = .01

C2 R2

1dB .027 15K

2dB .047 7K

3dB .082 4.7K

4dB .1 3.5K

5dB .12 2.6K

6dB .16 2K

4.7K Version, these values are for RL = 47K:

R1 = 4.7 K, C1 = .0047

C2 R2

1dB .012 33K

2dB .022 15K

3dB .039 10K

4dB .047 7.5K

5dB .056 5.6K

6dB .068 4.3K

Caps should be film type, mylar, poly, etc.

Pete B.

>Some asked for component values for other BSC step sizes,

>here they are:

>I also simulated the design taking into account the load

>impedance and the values are slightly different. I just used

>standard values in the early prototype and the BSC center

>frequency was slightly higher than the original target. These

>values are for a BSC center of 400 Hz:

>

>IN O--------- R1 --------------O OUT

> |--- C1 ---| |

> |

> C2

> | ----------> RL

> |

> |

> R2

> |

>GND O---------------------------------O GND

>

>These values are for RL = 22K:

>R1 = 2.2 K, C1 = .0047

> C2 R2

>1dB .027 15K

>2dB .047 7K

>3dB .082 4.7K

>4dB .1 3.5K

>5dB .12 2.6K

>6dB .016 2K

>

>These values are for RL = 47K:

>R1 = 4.7 K, C1 = .01

> C2 R2

>1dB .012 33K

>2dB .022 15K

>3dB .039 10K

>4dB .047 7.5K

>5dB .056 5.6K

>6dB .068 4.3K

>

>Caps should be film type, mylar, poly, etc.

[Advent experiment - mod - much better sound IMO]

Hi Ski,

Sorry I missed your post here for a while, good to hear that you tried it. I'm wondering which circuit you built, the 4.7K or 2.2K version or one with less than 6 dB of taper?

Basically, to make the calculations simple we'd like to ignore the source impedance and load impedance, and if they're 10X the circuit level impedance they can usually be ignored. So the 2.2K design wants to have a load of 22K or higher. The simulated designs take the load into account and therefore it should be close, but it is not critical since it is already much higher than the circuit level impdances. It will not blow up, with the wrong load or source impedance. However, a 10K load is rather low, it will reduce the number of dB in the correction. I said 6dB was probably too much so I think you'll be fine with the 10K load, I'd guess you'll get about 3-4 dB for the 2.2K ohm circuit.

The 800 ohm source impedance is a bit high but if we take a look at the circuit we see that the pre amp output impedance is in series with the first resistor R1 in the latest schematic. You can subtract the Pre Amp output impedance from R1 to and use the new value for R1. For 800, and the 2.2K circuit the new value for R1 is 2200-800=1400 ohms or 1.4K.

The book I usually recommend is Horowitz and Hill: The Art of Electronics, it's expensive so you might want to take a look at your library first. The very simple AC analysis is short caps, open inductors at high frequency (HF), short inductors, open caps at low frequencies (LF). Try redrawing the circuit, once for LF again for HF and compare.

I'm very curious to hear your description of the change in sound once you get the circuit sorted out.

Pete B.

[Advent experiment - mod - much better sound IMO]

Some asked for component values for other BSC step sizes, here they are:

I also simulated the design taking into account the load impedance and the values are slightly different. I just used standard values in the early prototype and the BSC center frequency was slightly higher than the original target. These values are for a BSC center of 400 Hz:

IN O--------- R1 --------------O OUT

|--- C1 ---| |

|

C2

| ----------> RL

|

|

R2

|

GND O---------------------------------O GND

These values are for RL = 22K:

R1 = 2.2 K, C1 = .0047

C2 R2

1dB .027 15K

2dB .047 7K

3dB .082 4.7K

4dB .1 3.5K

5dB .12 2.6K

6dB .016 2K

These values are for RL = 47K:

R1 = 4.7 K, C1 = .01

C2 R2

1dB .012 33K

2dB .022 15K

3dB .039 10K

4dB .047 7.5K

5dB .056 5.6K

6dB .068 4.3K

Caps should be film type, mylar, poly, etc.

[AR 310 HO]

A friend recently upgraded his system and brought over his AR 310 HO speakers that he removed. These are 3way MTM vented floorstanding speakers and I'm aware that they're nothing like the early ARs. Anyone know where crossover schematics can be found online?

Pete B.

[Evolution of and Replacements for the AR 11/12" Woofer]

I just noticed that the driver motor flux densities are given in the AR literature for several models and there are at least two different values:

11/12" woofer:

AR-10pi 9,800 gauss

AR-9 9,800

AR-11 8,200

[Evolution of and Replacements for the AR 11/12" Woofer]

Thanks Tom for the detailed information.

I've got the old AES article on the AR-1 which confirms the 43 Hz for Fc and shows a frequency response plot for 180 deg radiation, amplifier damping factor of 1. This is a very low damping factor and I'm not sure if the woofer was directly driven. Qtc will be lower with a high damping factor amplifier.

Was there a woofer inductor in the AR-1? Does anyone have the DC resistance for the inductors used in the various designs?

It is interesting that Villchur offers that "A system with a Q of about one has the flattest response curve and does not ring appreciably." He also refers to "The well-known family of universal Q curves" provided in Beranek's Acoustics. I just had to take a look in my copy and at least the curves are accurate but the interpretation is a bit off. A Qtc of one is loosly speaking "flat" "give or take" in that the response crosses zero dB at resonance, but there is some peaking above resonance and it is not monotonic. Strictly speaking it is not perfectly flat.

We say in modern control theory and filter design that a Qtc of .707 has a maximally flat amplitude response (AR). It is also know as a Butterworth alignment. This alignment has a monotonic rolloff in that the response always has a downward slope which is not true for responses with peaking. It is the highest or maximum Q value that has no peaking in the frequency response or that is monotonic and this is why it is referred to as maximally flat.

It is also universally accepted in modern control theory and filter design that a Qtc of .5 is critically damped. This is also known as a Gaussian alignment. It is the highest Qtc value that has no ringing in the transient response (TR). Qtc values below critically damped are referred to as overdamped and also do not ring, and above as underdamped and always have some ringing.

Qtc Output at Fc Features Name

-------------------------------------------------------------

<.5 Overdamped TR

.5 - 6 dB Critically Damped TR Gaussian

>.5 Underdamped TR

.707 - 3 dB Maximally Flat AR Butterworth

1 0 dB 0 dB at Fc peaking above Fc Chebyshev

>.707 Peaking in AR Chebyshev

Beranek also suggests that Qt should be unity for "flat" response down to the lowest frequencies. Many would say that strictly speaking this is an error since it is not monotonically flat and there is some peaking.