More 12" restoration questions

[Guest Bret]

I don’t know if I’m trying to make a point or have someone confirm that my method is sound [no pun intended] or if a major correction to my thinking is in order. One more time into the breach, gentlemen, as I get in way over my head and fool around with the UniBox software and the measurements Ken made for the various drivers he tested for us then try to relate those to something else.

Opening UniBox, selecting DUT #6 for values, then looking at the “Closed Box” area under “Design by Vb and Q” - changing the Physical Vb to 41.9 liters (correct for the 10pi), leaving Absorption alone at 20 and Leakage alone at 15 (neither of which I know is reasonable), and “recalculating” the graphs; the software shows me that the Enclosure Resonance is 42.6Hz and the -3db point is 40.93Hz. This is certainly a low enough figure to be satisfying in all respects which is confirmed by listening.

Now repeat the procedure for DUT #2. The spreadsheet will go back to default values, so you have to reset the enclosure size and update the graphs, etc. NOW look. We get an Enclosure Resonance at 46.67 which at first glance doesn’t seem that dramatic. A 12db/octave roll-off from there to Fs is still going to produce some bass. But look at the -3db figure! It is at a whopping 71.12Hz ! We’re a full -6db somewhere right around 50Hz as opposed to 35Hz. What happened?

Is the variance caused by the Free Air Resonance problem? There appears to be a meaningful divergence in Vas numbers (the equivalent amount of air to replace the driver’s suspension).

Due to the differences in Fs, is the Qms (driver’s Q at resonance) reasonable?

Does such a large difference in Fs automatically imply the differences in Qms?

It’s almost as though they are two entirely different drivers.

My 10pi’s are considerably brighter, overall, than a friend’s. There also seems to be lower-midrange “hole.” Thinking I might get these speakers to “hang together” better, last night I replaced units DUT #4 and #7 with the AB Tech purchased replacements.

Several things happened to the “system” all at once. We spent hours listening and I’ll describe what happened in detail if asked. What I’m trying to get to the bottom of [no pun intended] is the reason for the complete loss of deep bass. It was gone; AWOL, missing entirely. What driver parameter is responsible for this?

Thanks for your indulgence.

Bret

[Pete B]

Hi Bret,

I've looked at the data a few times and it is clear that there is a large spread in the drivers because there are two driver types, and perhaps a mixed experimental or transition type. I've gone through the data and was never completely clear on which drivers these all are, except for the ones from the 10pi and the 303. Do you have the AR part numbers for these drivers?

DUT4 shows a DCR around 20 which I think is a typo, I corrected it to 2.5 as an estimate. Xmax peak should be 6 mm not 20 mm for all the non 303 drivers, does not matter for small signal analysis. Pn in the driver section should represent the large signal power level of interest but does not matter for small signal analysis. I'd estimate 60 W for RMS or 200 W peak analysis, it is simply the thermal limit of the driver/analysis, use 1000W if you simply want to investigate displacement limited large signal capability. Rs of .1 on the main Unibox page is correct for directly driving the woofers, but should be changed to the DC resistance of the woofer crossover inductor. This will raise Qts and Qtc and is necessary to accurately predict the low end response.

I noticed that DUT7 was completely rebuilt with a new cone, and it appears that they've altered the driver since the new moving mass is now 164.88 g. Remember that I mentioned that moving mass is more important than Vas or Fs as long as Vas is much greater than the box volume (high alpha) since the mass along with the box volume then primarily determine Fc:

http://www.arsenal.net/dc/dcboard.php?az=s...page=show_topic It seems that a moving mass (Mms) of about 100 to 110 g is about right. Note that Fs and Fc are inverse square law with regard to moving mass, that is Mms must be increased by a factor of 4 to reduce Fc or Fs to one half and small Mms differences do not shift Fs significantly. However, efficiency is highly sensitive to Mms so it is important in that regard:

------ moving mass (Mms)

DUT2 103 g

DUT3 104.36 g

DUT4 123.95 g

DUT5 107 g

DUT6 115.16 g

DUT7 164.88 g (reconned with probably the wrong cone)

DUT1 139g (actually a 303 woofer)

It's clear that some of the drivers, DUT2 and DUT3 (Bl about 11.75)for example have significantly stronger motors as compared to DUT5 and DUT6 (Bl about 10), which results in lower Qes - Qts - Qtc. DUT 7 is yet different and more like the 303 with a Bl of 13:

------ Bl ------- Ref Eff

DUT2 11.76 Tm .572 %

DUT3 11.78 Tm .531 %

DUT4 11.59 Tm .358 % (estimated used RDC = 2.5 ohms)

DUT5 10.10 Tm .391 %

DUT6 9.89 Tm .319 %

DUT7 13 Tm .289 % (was the voice coil replaced?)

DUT1 12.94 .320 % (actually a 303 woofer)

I would not usually be very concerned with Qms except when trying to raise the Q of a system, a low Qms will require more series resistance, and/or more damping material to be removed when trying to raise Qts/Qtc. DUT2 and DUT3 need a higher Qts but unfortunately their Qms is somewhat low:

------ Qms

DUT2 2.89

DUT3 3.11

DUT4 2.82

DUT5 7.92

DUT6 7.89

DUT7 9.3 (reconned)

DUT1 5.02 (actually a 303 woofer)

DUT4 seems to be close to the family of DUT2 and 3 but the cone mass is higher for some reason. It seems that DUT 2, 3 and 4 probably use a similar spider with their lower Qms and lower Vas as compared to 5 and 6.

Do the magnets all look the same?

Are they all ceramic types?

I'm thinking that AR might have moved to a stronger magnet material, or is the top plate thickness different? Are we certain that the motor geometry is the same, plate/VC height?

The drivers all have similar voice coil resistance, and inductance which suggests that there were no major changes to the voice coil wind.

I would add between 1 and 1.5 ohms in series with woofers of type DUT2 or 3 to raise Qts/Qtc and remove about half of the damping material behind the woofer, none should touch the frame or magnet. This can be simulated in Unibox by adding 1 to 1.5 ohms in series with the inductor value and using the total for Rs. Use "walls covered" to approximate less damping material, or manually enter Qa = 10 as an intermediate value.

One could also add about 20g of mass to DUT2 and 3 to throw away some efficiency and get a lower Fs/Fc to provide better extension but the series resistance and damping removal are more important. This can be simulated in Unibox by lowering the entered value for Fs (on sheet "DDBase"), just a few Hz is about right, 25 to 23 for example, see how Mms changes and adjust if necessary.

To answer your question for simulating with Unibox, there is a small initial state problem, probably due to the data since I don't see it in my original Unibox of the same rev. number. You might have noticed, the Qa value of 20 corresponds to "walls covered" but the field says "heavy fill". Just hit "recal" next to damping and the correct value of Qa = 5 for heavy fill will be entered and used. If you do this with the bug in mind and hit "recal" the F3 for DUT2 is even worse at 78.7 Hz. Hit recal often when using Unibox.

Let's look at a few more cases and try to improve the situation with these lower Q drivers.

Rs of .1 is incorrect for all these simulations since it is important to include the Rdc of the crossover inductor to raise the effective Qtc of the system. I get an F3 of 57.54 Hz for DUT2 just by changing Rs to 1 ohm, remember to hit enter in the Rs field and recal in the damping area. Changing box damping to "walls covered" further lowers F3 to 51.52 Hz. Changing Rs to 2 ohms (assuming a 1 ohm/50W resistor is added) F3 becomes 42.46 Hz, and with Rs = 2.5 ohms F3 = 40.18 Hz.

Changing Fs to 22 Hz on the DDBase sheet for DUT2 represents increasing the moving mass by about 23 grams, remember to hit reload for the driver on the main page for this to take effect. F3 is then 38.37 Hz with Rs = 2 and "walls covered" damping. I'd expect this or the 2.5 ohm example to sound very close to the 10pi woofers and the original AR 11/12" woofer in the low bass. It may be a bit too heavy in the bass for some people's taste.

It is interesting that DUT5 and 6 have about 89.5 dB/2.83V/m with Rs = 1 ohm, and DUT2 and 3 need an Rs of about 2 ohms to provide about the same voltage sensitivity.

The 303 woofer (DUT1) is also close to the 10pi woofers when a value of 1 is used for Rs and "walls covered" damping, F3 is then 43.65 Hz and no extra resistor is needed.

I'd like to read about what you heard.

Pete B.

[RoyC]

Bret,

I tried new replacement woofers (then known as the "Tonegen" woofer) in AR-3a's and abandoned them long ago for the same reason. Tom T. provided an interesting history of the 12 incher and explained that it was gradually changed (less compliant) as time went on. My personal experience is that virtually any decently refoamed early era AR 12" woofer will produce better bass response than a new replacement in cabinets of the same volume and stuffing as the AR-3a, 11 or 10pi. Do you have the original woofers for your '10pi's.

Roy C.

[Guest Bret]

>Rs of .1 on the main Unibox page is correct for directly driving the woofers, but should be changed to the DC resistance of the woofer crossover inductor. This will raise Qts and Qtc and is necessary to accurately predict the low end response.<

Sure, now you tell me! ;-) Looks like I have a lot of going-back and restudying to do. For instance, I could use the “2 drivers in parallel” setting to see how they ought to act in my 9s, as well as changing to the resistance of the woofer’s crossover inductor (which, of course, I have no clue about).

>I've looked at the data a few times and it is clear that there is a large spread in the drivers because there are two driver types, and perhaps a mixed experimental or transition type.<

Ah, you share my confusion considering these numbers. What I’m about to tell you is very likely to come as a complete shock. Let’s ignore DUT 1 for the moment.

All of the rest of these drivers are 200003s. As far as I can tell, the motor portion of all six are identical.

Let me explain DUT #7. That was an experiment run by Tri-State. Mass was added to the on the back (hot melt with cone material stuck on it). Although I understood he was trying to raise the cone’s mass to drive the Fs down to18Hz (per AR literature), and clearly he got it down to close to 18Hz, but otherwise this doesn’t seem to work very well. I could probably remove most of that extra weight without destroying anything and bring this one to the same performance as DUT #4.

DUT 2,3,4, and 7 are all identical 200003 woofers from my AR-9s (circa ‘78).

DUTs 5 and 6 are the 10pi woofers. To borrow a phrase from Roger Daltry, “Now here’s the surprise:” #5 got a new surround and spider. DUT #6 got a new voice coil, a new surround, and a new spider (reusing only the cone material). The cone material may be very important. These 10pi’s were EARLY production models, with the fuses on the back. Originally these 10pi’s had the lighter-colored tweeters in them, not the soft cloth dome but some sort of pressed-paper as can be seen in the 10pi color brochure. Those were replaced early on at the dealer (by me) because one of them was dented and there was no way to work a dent out of that material, it would have cracked and split. By this time, all replacements were the soft cloth, no felt/foam. The point is that 10pi’s with this woofer cone material were not unique, but were rare and early. I wish you could feel this cone material. It is close to the pressed-felt on some old 2ax speakers.

DUT 5 http://www.aural.org/ar_hist/data/05/05_s.JPG

DUT 6 http://www.aural.org/ar_hist/data/06/06_s.JPG

Next, let’s tackle #2 and #3. These two drivers were sent to Simply Speakers for re-surrounding. The old, floppy spiders (from years of my playing Also Sprach Zarathustra, Berlioz Requiem, and Saint Sean’s organ symphony at “Earsplittin’ Loudenboomer” volumes). CLEARLY this shows what happens if you stick a surround on these speakers that’s just way toooooo. . . . something. You tell me. Too floppy? I was guessing way too stiff; otherwise I can’t account for the rise in Fs. Is it possible that the “stronger motor”numbers are just an artifact of a worn-out spider being held in place by surrounds that are too stiff ?

>I'm thinking that AR might have moved to a stronger magnet material, or is the top plate thickness different? Are we certain that the motor geometry is the same, plate/VC height?<

I haven’t put a caliper to it, but they look to be identical as far as size goes. Obviously I can’t comment on the magnet strength, but I find differences very unlikely. You are finding these differences in drivers from a single pair of AR-9s with drivers that look-alike.

>Let's look at a few more cases and try to improve the situation with these lower Q drivers.<

I can’t wait, but it’s going to have to wait. This is “brain-work” for me since I’m still grasping these concepts and working with unfamiliar relationships and tools; I’ll get it, it’s just taking a minute. I've read your suggestions a few times. They need time to perc.

I appreciate your taking the time to analyze, but most of all I appreciate your time and willingness to instruct.

Let’s continue this tomorrow if possible. I'm brain-dead for the night.

Thanks again.

Bret

[Guest Bret]

>Do you have the original woofers for your '10pi's.<

Unfortunately, no. A guy parted-out these 10pi's on eBay and only through persistence was I able to get him to sell me the cabinets that were in his garage (which I was afraid were going to get tossed, eventually). Originally I thought I would use Tonegen replacements in these and put my original drivers back in my 9s.

Obviously, that isn't going to happen. . .or if it did, the result would be most satisfying.

Bret

[Guest Bret]

>I'd like to read about what you heard.<

Thank you for asking. I am happy to share whatever I experienced.

The "Tonegens" (if that's what they are) are considerably louder and make lousy midrange drivers in these cabinets.

Background: My 10pi's are too bright. I replaced all the capacitors with Solen (except the 2500uF) and they got better as far as "balance" is concerned, but got harsh, too. As I say, balance was better, but still not right. There seemed to be a "hole" in the lower midrange, upper bass. (I think it may be a stuffing issue)

Thinking that I could overcome that problem by using what are clearly more efficient "Tonegens," I decided to remove the repaired 200003s (4 and 7) and install the replacements.

The cabinets did hang-together better in the lower midrange, upper bass area. In fact, they "over-hung together" to a pucker-factor of about 8. Leaving the tweeters at "0" and turning the midrange drivers to -3db made them more tolerable, but still throaty when "gutsy" was required. Turning the midranges down further would have been useless as the upper midrange was now sucked-out with the accompanying loss of detail.

What I didn't expect was the extent of the bass drop-out. The only way to get these speakers to skip a stylus across an LP would be to drop one of them on the corner of the turntable. The bass wasn't only "attenuated somewhat" it was out-cold. AWOL. True, the speakers were away from the walls, but they were flat on the floor in the 4pi position.

With the old, repaired drivers in them, this position was tolerable, but only just. There was so much bass and so much mud in the bass that you felt like you needed to scrape it off the bottom of your shoes.

With the Tonegens in there they still had plenty of "thump" but no "whuump." ZZTop kept it's "bang" and impact until a percussive effect should have made the couch shake. Not only did it not shake the furniture, it wasn't even audible.

Male vocals lost all their chestiness and authority. Listening to Ladysmith Black Mambazo's "Shaka Zulu" was wholey useless. It sounded like they all had something in their throats. No chest. The periodic gigantic "WHUUUUMM" in Fleetwood Mac's "Emerald Eyes" just wasn't there. At all. It was like the recording had been edited.

Piano sounded "far away" and unsatisfying. Supertramp's MFSL release of "Crime of the Century" was unlistenable.

Also on Fleetwood Mac's "Mystery to Me" is a recording of "For Your Love." A few times on that track Mick must hit a big drum with a loose head using a very soft mallet. It was audible, but not in the proper octave and it certainly didn't move the Earth as it should. It was unidentifiable without the low register cues.

And beyond mourning the loss of house-shaking low sounds, the midrange ended-up sounding colored and confused and at times like a fire-truck's horn.

We listened to the same music (repeatedly trying to get a firm grasp of a few key elements) before and after sticking the Tonegen's into the cabinet, so we didn't listen to a great variety of things; it wasn't necessary. There was no part of the "Tonegen" experiment that went well for the replacement drivers.

Without first having the benefit of doing any UniBox experiments I described the problem to my listening-buddy as "there is way too much in the 350-500Hz range, a huge bump somewhere around 70-100Hz and not much of nothin' below that."

I'm sorry if it seems I've gone all "Stereophile" on everyone, but this isn't a nuance I'm talking about so I described "big" because what's missing is "big."

They are too loud, they throw the whole cabinet out of balance, they have no umph, they don't seem "fast," just tight and stingy. Bizarrely they seemed to suck all the dynamics right out of the recording, even in frequency ranges where they don't operate.

In fact, I'm going to feel guilty about sticking them on eBay. I've got about $520 + freight in them and need some of that back but I really can't see representing them as replacement woofers for AR speakers and can't imagine that they have other uses. . . refrigerator magnets? Bulk tape erasers? Ballast?

And now, an editorial:

In my personal opinion, those who sell these as drop-in replacements should be ashamed of themselves. You and I know that *they* know better. The flat sided baskets turn these replacements into AR woofers the same as a Halloween mask turns a sweet little girl into a lion. These drivers are sheep in wolves' clothing and everyone who buys them, in my personal opinion, is getting fleeced. (everybody sing, "Money for Mutton" by Dire Strifes)

Personally, I think anyone who sells these under false pretenses should replace them with suitable drivers - which is what we asked for and paid-for and which these were represented as, right? I'd love to get a pair of these into a courtroom in a class action suit. Justice is blind, but she isn't deaf.

Bret

PS - "I want my, I want my, I want my low H Z"

[Guest Bret]

Thanks Pete,

To continue:

>I would add between 1 and 1.5 ohms in series with woofers of type DUT2 or 3 to raise Qts/Qtc and remove about half of the damping material behind the woofer, none should touch the frame or magnet. This can be simulated in Unibox by adding 1 to 1.5 ohms in series with the inductor value and using the total for Rs. Use "walls covered" to approximate less damping material, or manually enter Qa = 10 as an intermediate value.<

Good grief ! That does flatten things out nicely, doesn't it? Wow. I'll experiment. Now, I usually find that removing a lot of stuffing results in a hollow "ring" I attribute to inadequate damping, but I won't know until I try. Cool. Thanks. But, doesn't that lower Re? So am I going to be able to get this change, or am I just "simulating" a situation I can't obtain?

>One could also add about 20g of mass to DUT2 and 3 to throw away some efficiency and get a lower Fs/Fc to provide better extension but the series resistance and damping removal are more important. This can be simulated in Unibox by lowering the entered value for Fs (on sheet "DDBase"), just a few Hz is about right, 25 to 23 for example, see how Mms changes and adjust if necessary.<

Got it. I only had to read it fully awake. I see that Mms increases to about 139g. So if I were able to uniformly add mass in order to drive down the Fs, I get a lowering of Fc as well. It looks like I could get #2 and #3 to "act right" or at least approximate acting right.

I talked to Mr. Miller today. He's willing to entirely recone #2 and #3 IF he has cones and voice coils for the AR 12" in shop. I'm supposed to call tomorrow afternoon to see if he does. Clearly I can "do no harm" if I'm going to have these rebuilt so before I send them away I'm going to do what you and UniBox suggest for the learning experience. I may even give these things a whirl with the Tonegens (although we don't know degree).

Is any inductor size unimportant because we aren't trying to filter low frequencies in this case?

>Rs of .1 is incorrect for all these simulations since it is important to include the Rdc of the crossover inductor to raise the effective Qtc of the system. I get an F3 of 57.54 Hz for DUT2 just by changing Rs to 1 ohm, remember to hit enter in the Rs field and recal in the damping area. Changing box damping to "walls covered" further lowers F3 to 51.52 Hz. Changing Rs to 2 ohms (assuming a 1 ohm/50W resistor is added) F3 becomes 42.46 Hz, and with Rs = 2.5 ohms F3 = 40.18 Hz.<

This is certainly worth trying. Thanks for doing the work. I would never have thought to raise Rs to include the choke, but I see why that, and all the other crossover electronics, must be considered. It drives home the point about changing ESR by using new capacitors. (that doesn't mean my opinion has changed about the necessity of changing the caps)

>I'd expect this or the 2.5 ohm example to sound very close to the 10pi woofers and the original AR 11/12" woofer in the low bass. It may be a bit too heavy in the bass for some people's taste.<

I wish I had an all-original pair to compare to, but I'll give it a good test-listen. Judging by these numbers I ought to try 2.125 or so ohms.

>It is interesting that DUT5 and 6 have about 89.5 dB/2.83V/m with Rs = 1 ohm, and DUT2 and 3 need an Rs of about 2 ohms to provide about the same voltage sensitivity. <

Do you believe this might be due to the old spiders on #2 and #3?

Bret

[RoyC]

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.

http://www.arsenal.net/dc/user_files/301.jpg

http://www.arsenal.net/dc/user_files/302.jpg

[Guest Bret]

>the last 2 should work<

They do. It's easy to identify the 9 woofer of the two.

Simply Speakers claims that *their* 12" replacement is an AR build. I have my doubts about this, but if it were true it opens an opportunity to find an "original" set of specs.

Bret

[Pete B]

Hi Bret,

It's taking me a little while to digest your description and thoughts but I'll try to comment. I feel a bit like a detective here, and I'm not sure were thinking along the same lines. As a detective I'm wondering why only 2 of your AR-9 woofers went out for (the same) service. Did that side get overloaded?

My conclusion at this point is that AR made changes to the 200003 woofer and just kept the same part number probably to avoid having to produce the older woofer. They probably designed all their current and new systems to match the newer spec driver and just figured close enough as a replacement for the older systems. This is not that uncommon in manufacturing to avoid having to stock and produce a lot of different parts. I've often heard of early and late production of the same part being different, not that it's good practice. That's even how some refer to the parts.

Your DUT2 and 3 drivers may be just fine for an AR-9. The AR-9 has that special circuit to raise the Q and impedance of the system so they may be just fine. Do you find that your 9s are not sounding as good as they did before the work was done?

Is it well known among AR people that early and late production AR-9s sound different as far as deep bass goes?

I'm not sure what your goals are for the 9s, and are you satisfied with DUT5 and 6 in the 10pi's? Also, I'm not sure what your goals are in general are you thinking that the woofers should swap because they have the same part number? This might not be a reasonable goal, without altering damping and adding resistors. I'm certain that the magnet strength is different.

Why not have Miller do DUT4 and 7 which seem to need work, or do you think that any one of DUT2,3,4,7 currently has about the right T&S parameters for the AR-9?

Another question do all these drivers have the protective screen in place? It could alter Qms.

You mention the soft cone, I do have old AR2ax woofers with the very soft cone and I don't think it's helping anything, it will tend to flex more under high excursion and cause a loss of output. The advantage is that it's mass combined with its softness will roll off the high end fast allowing for a simpler and cheaper crossover when used high into the mid bass (500 to 1 kHz). You mention the floppy cone and Roy mentions the floppy suspension, I think the only reason why these drivers work well is that another part such as the magnet is different in these older drivers.

I think the late 200003 used a stronger ceramic magnet, note that DUT4 has about the same Bl as 2 and 3, 7 could be a measurement error, or it could have the wrong voice coil (different l (as in Bl)- length of wire in the gap). There are different grades of ceramic magnets so this is a reasonable theory. Ceramic magnets can also demagnitize with use, heat, and high stress, so this is also a possibility but I lean toward the differing magnet material theory.

Did you mark your woofers before they went out for work, could they have gotten mixed up with someone else's when they went out for service?

We don't know the T&S parameters for the Tongens, correct? Anyone have measurements of the Tonegen parts? Do we think that the Tonegens are a direct replacement for the late 200003?

We could assume that they're close to the newer 200003s and see how my suggestions work with them. The Unibox analysis that we've done so far is only valid below about 100 to 200 Hz, we have to consider the rest of the crossover and cone breakup modes as we move into the upper range of the woofers. I believe the crossover for the 9s is too complex for Unibox, you can try just the DC resistance of the crossover as a rough approximation. I can do a full AR-9 woofer simulation in CALSOD if I get the necessary data, if your interested, not sure how soon I can get to it anyway.

[RoyC]

Did Simply Speakers give you a part number? When I first looked into replacements in the early 90's, nobody had them. I received my Tonegens directly from AR in '94 after waiting for a "backorder" to be filled. Interestingly AB Tech and Simply Speakers got them in stock right after I received them...

[Pete B]

>Good grief ! That does flatten things out nicely, doesn't

>it? Wow. I'll experiment. Now, I usually find that removing

>a lot of stuffing results in a hollow "ring" I attribute to

>inadequate damping, but I won't know until I try. Cool.

>Thanks. But, doesn't that lower Re? So am I going to be able

>to get this change, or am I just "simulating" a situation I

>can't obtain?

I would at least keep the walls lined, which should avoid serious ringing if you even want to remove that much. Or leave the mid/tweeter end stuffed and remove about half behind the woofer, or try your own ideas, you'll only find what works by experimenting. The resistor should raise the input impedance of the woofer as seen by the crossover. Remove the hot lead to the woofer and insert the resistor between the lead and the woofer terminal. I typically have several .5, 1, and 2 ohm 10W resistors handy for experimenting, you can bring clip leads out if you can sneak them out under the woofer frame. You mention the change in Re, yes the loading of the crossover will be slightly higher and a shunt resistor could be placed across the crossover leads as shown below:

-----------------Rs (series) --------- +

XO output | Woofer

Rp (shunt)

|

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

Note that I computed these values taking into account

the voice coil inductance and the fact that the impedance

should match around the crossover frequency. This does not

apply to the AR-9s with their special crossover:

Rs Rp

1 20

1.5 15

2 12

2.5 10

>

>>One could also add about 20g of mass to DUT2 and 3 to throw

>away some efficiency and get a lower Fs/Fc to provide better

>extension but the series resistance and damping removal are

>more important. This can be simulated in Unibox by lowering

>the entered value for Fs (on sheet "DDBase"), just a few Hz is

>about right, 25 to 23 for example, see how Mms changes and

>adjust if necessary.<

>

>Got it. I only had to read it fully awake. I see that Mms

>increases to about 139g. So if I were able to uniformly add

>mass in order to drive down the Fs, I get a lowering of Fc as

>well. It looks like I could get #2 and #3 to "act right" or

>at least approximate acting right.

Yes this might be a better idea if you had a way to measure the input impedance of the system so that you could see exactly where Fc is before and after adding mass. I don't usually do this, I offer it

in case anyone wants to experiment and to demonstrate it in Unibox. Mortite on poly cones comes off easily, are these all paper?

>I talked to Mr. Miller today. He's willing to entirely recone

>#2 and #3 IF he has cones and voice coils for the AR 12" in

>shop. I'm supposed to call tomorrow afternoon to see if he

>does. Clearly I can "do no harm" if I'm going to have these

>rebuilt so before I send them away I'm going to do what you

>and UniBox suggest for the learning experience. I may even

>give these things a whirl with the Tonegens (although we don't

>know degree).

I'm not sure there's anything wrong with 2 and 3. I hear Miller does good work but I doubt that it will solve the problem unless he knows how to rebuild them to "old" specs, including a trick to weaken the magnet. I can understand wanting to do this just to rule anything else out. It would help if Miller measures his drivers and knows the typical T&S parameters for AR-9 woofers with his replacement parts.

>Is any inductor size unimportant because we aren't trying to

>filter low frequencies in this case?

Not sure I'm following, the inductor will effect the higher end of the woofer range where it crosses over to the mid. Roy speaks of two values for old and newer AR-3a systems, this could be to match the different woofers, or to alter the voicing. We've not dug into the upper range of the woofer, yet. I'd need a crossover schematic for the 10pi to do this.

>>Rs of .1 is incorrect for all these simulations since it is

>important to include the Rdc of the crossover inductor to

>raise the effective Qtc of the system. I get an F3 of 57.54 Hz

>for DUT2 just by changing Rs to 1 ohm, remember to hit enter

>in the Rs field and recal in the damping area. Changing box

>damping to "walls covered" further lowers F3 to 51.52 Hz.

>Changing Rs to 2 ohms (assuming a 1 ohm/50W resistor is added)

>F3 becomes 42.46 Hz, and with Rs = 2.5 ohms F3 = 40.18 Hz.<

>

>This is certainly worth trying. Thanks for doing the work. I

>would never have thought to raise Rs to include the choke, but

>I see why that, and all the other crossover electronics, must

>be considered. It drives home the point about changing ESR by

>using new capacitors. (that doesn't mean my opinion has

>changed about the necessity of changing the caps)

>

>>I'd expect this or the 2.5 ohm example to sound very close to

>the 10pi woofers and the original AR 11/12" woofer in the low

>bass. It may be a bit too heavy in the bass for some people's

>taste.<

>

>I wish I had an all-original pair to compare to, but I'll give

>it a good test-listen. Judging by these numbers I ought to try

>2.125 or so ohms.

I'd be very curious to hear what you think, 2.5 with reduced damping, might be a good place to start to see if you can get them to a point where there's too much bass, then work back.

>>It is interesting that DUT5 and 6 have about 89.5 dB/2.83V/m

>with Rs = 1 ohm, and DUT2 and 3 need an Rs of about 2 ohms to

>provide about the same voltage sensitivity. <

>

>Do you believe this might be due to the old spiders on #2 and

>#3?

I believe that it is due to a difference in magnet strength, but it shows that the new driver plus resistor essentialy "clones" the behavior of the older driver. The force of the motor is Bli, where i is the VC current, B is too high so we lower i by adding resistance.

Pete B.

[Pete B]

RoyC provided this info in another thread:

#5280, "RE: Evolution of and Replacements for the AR 11/12" Woo..."

In response to Reply # 10

Hey Pete, The AR#9 inductor (2.85mh) was used with this woofer in the '3a and by my measurement with an inexpensive LCR meter its .9+ ohms. Some of us have found that an AR#7 inductor (1.9mh) was used with the cloth surround woofer in some early '3a's and that measures .7+ ohms. Roy

Thanks Roy,

We would use .7 ohms for systems that use AR#7 inductors, and .9 for AR#9 inductors when simulating in Unibox where I used 1 ohm as an example. One could add another .1 for hookup wire resistance. More should be added if a higher output impedance amp is used, tube or whatever.

[Guest Bret]

>Did Simply Speakers give you a part number?<

Sure, it's right on their website. . .

1210003-2

"Factory AR Part" they say. So I said. . ."Okay, who manufactures these "AR" drivers?" The answer was "AR." My disbelief runs deep.

Bret

[Guest Bret]

>As a detective I'm wondering why only 2 of your AR-9 woofers went out for (the same) service. Did that side get overloaded?<

No. When Ken offered to do this I thought it was important for us to know the results of more than one shop's repair work and so sent them to two different places. I did mark them, but not with a marker (stupid of me), I put labels on them. Not one of them came back with the label still on it.

>Did you mark your woofers before they went out for work, could they have gotten mixed up with someone else's when they went out for service?<

It's entirely possible that #2 and #3 are different.

>Your DUT2 and 3 drivers may be just fine for an AR-9. The AR-9 has that special circuit to raise the Q and impedance of the system so they may be just fine. Do you find that your 9s are not sounding as good as they did before the work was done?<

The two drivers in question have not been installed in anything since I received them back. Since they have an Fs 25% higher than the others, I thought they probably weren't worth trying.

>Is it well known among AR people that early and late production AR-9s sound different as far as deep bass goes?<

Not that I'm aware of. In fact, I don't understand the reason for the question.

>I'm not sure what your goals are for the 9s, and are you satisfied with DUT5 and 6 in the 10pis?<

#4 and #7 were in my 10pis. They seemed to belong. 5 and 6 came from 10pis and back into 10pis they went. Yes, they are very satisfying.

>Another question do all these drivers have the protective screen in place? It could alter Qms<

No, they do not.

>I think the late 200003 used a stronger ceramic magnet, note that DUT4 has about the same Bl as 2 and 3, 7 could be a measurement error, or it could have the wrong voice coil (different l (as in Bl)- length of wire in the gap). <

The fact that they are alike would make some sense. They were all manufactured about the same time. Units 2, 3, and 4 do not have new voice coils, 7 does.

>We don't know the T&S parameters for the Tongens, correct? <

Correct.

>Do we think that the Tonegens are a direct replacement for the late 200003?<

They are sold that way, but I no longer consider them a suitable "drop-in" replacement.

>I can do a full AR-9 woofer simulation in CALSOD if I get the necessary data, if your interested, not sure how soon I can get to it anyway.<

Since the 9 is bi-ampable, the woofer section is really pretty simple, if ingenius.

>Why not have Miller do DUT4 and 7 which seem to need work, or do you think that any one of DUT2,3,4,7 currently has about the right T&S parameters for the AR-9?<

Other than the extra mass added to #7 (which I may be able to remove), what spec seems out of line on 4 and 7?

Pete, let's see if I can cause an "ah ha!" for you. The 10pi woofers are a friend's and they go in his 10pis. He bought Tonegen replacements and installed them. We hadn't heard his speakers in years and years before installing the Tonegens. I hadn't had anything worth hearing since just after my son was born - the 9s went to storage (climate controlled) with rotten surrounds. I bought replacement drivers for my 9s. The opportunity came for Mr. Kantor to measure AR drivers and I was willing to have my old ones repaired and sent to him for his tests. I also sent the friend's 10pi drivers. I sent my drivers two different places on purpose, so we could compare repair jobs. When the repaired drivers were returned to us, we put the original 10pi drivers back into the cabinets from whence they came, removing Tonegens. The result was that we could no longer listen to the Tonegens (his speakers just keep getting better and better). I put a pair of the repaired 9 drivers into my 10pi cabinets, rather than remove the Tonegens from my 9s and put them in the 10pis.

In searching for the answer to the coil question posted here, I decided to remove the Tonegens from my 9s, took some pictures, made a note or two. I then removed the repaired units from the 10pis and dropped in the Tonegens which I had just taken from the 9s. My 10pis went from sounding unbalanced on the bright side to sounding awful on all sides, but especially the bass. Later, I installed the repaired drivers into the 9s and listened. No comparison. The repaired drivers blow the Tonegens into the weeds in every important respect - I'll bet my 9s are reaching lower by half an octave or more. So, despite the "problems" with units 4 and 7, they are in every respect superior to the Tonegen replacements in either a 10pi or 9 system.

I went to a Community Symphonic Band concert tonight and made note of a couple of things, sonically. First, a bass drum shakes the auditorium even when struck lightly. Second, a tuba shakes the auditorium when played moderately. Third, a barritone shakes the auditorium. Fourth, a trombone vibrates my chest. And so on and so on. Everything that played below the frequencies of a trumpet shook my seat, my chest, and the rest of me to greater and lesser degrees. Even the toms on a trap set used for a couple of pieces had physical impact. (I was also amazed by the lack of brightness in even a piccallo or glockenspiel from my seat) I do not understand how anyone who says they want the faithful reproduction of live music can consider the frequencies below 60Hz "teenage stuff," but those above 15KHz "all important." The big waves are a very real part of the "experience" of live, unreinforced, acoustic-instrument, music and the occur at *moderate* volumes. I was distracted by my son's running his fingers over his program, yet things shook.

It also occurred to me that no matter how great my speakers are or how many watts I throw at them, they will never have the huge energy of a simple concert bass drum, even at a distance, but especially not at low volumes.

Extension into the sub-sonic arena isn't a luxury, it is a necessity. For whatever reason, the Tonegen drivers do not seem to want to go there regardless of cabinet, stuffing, or crossover.

My 9s are eventually going to get 4 original, semi-matching, 200003s. I'm going to have to purchase a pair of used woofers for my 10pis and have them brought as close to original spec as skilled repair people can manage.

Bret

[Guest Bret]

>I think the late 200003 used a stronger ceramic magnet, note that DUT4 has about the same Bl as 2 and 3, 7 could be a measurement error, or it could have the wrong voice coil (different l (as in Bl)- length of wire in the gap). <

Could this be an artifact of not getting the voice coil at "0" in the gap?

In other words, could glueing up a new surround with the voice coil either too much into, or not far enough into, the gap, cause this?

Bret

[RoyC]

I'm with you in the disbelief department, Bret. The # is the same as that on my old Tonegen with the exception of the "A" at the end. The Tonegen was/is considered to be the authorized replacement and was in fact used in the AR-3a "limited" in the early 90's which sported a new crossover and 8 ohm l-pads instead of pots. I bet its the same animal, and probably not worth the dough to find out until we have more info..

Roy C.

[Pete B]

We might want to see if Miller can recone a Tonegen with the older 10pi parts. It would be interesting to see what can be salvaged from the Tonegens. Probably just needs a new soft spider. The cost is not reasonable given what other 12's go for.

Do any of the 11/12" AR/Tonegen woofers have a bumped back plate?

I believe that they stiffened the suspension to improve centering, and to reduce bottoming. Is bottoming a problem when driven hard?

The floppy suspension might not have returned the voice coil to center reliably as the driver aged. I would shoot for loose enough to get an 18 to 20 Hz Fs.

It's too bad that the odd frame is required in these speakers, will the 9's take a standard 12" woofer?

[Pete B]

I took a rough estimate of the AR-9, simulating half of it so that only one woofer is used, the FR response is the same as with 2 woofers. I estimated 70 liters (a bit more than half of 120 to account for the stuffing at the top), used a 3 ohm source impedance as an approximation to the clever crossover (this is analogous to 1.5 for 2 woofers), walls covered damping with the 303 woofer, Fb = 32.68 Hz, F3 = 29.92 Hz, there's slight peaking indicating that the Q is just over maximally flat. I believe that these would work well but I'd want to also do a CALSOD simulation.

Are replacements for the 303 woofers available? Do they have a bumped back plate?

[Pete B]

Magnet info:

http://www.dextermag.com/Page82.aspx

[Pete B]

>Is any inductor size unimportant because we aren't trying to

>filter low frequencies in this case?

This is a good question, yes that's right the inductance value doesn't matter at these low frequencies say below 100 Hz, but the resistance does as we've seen.

[Guest Bret]

>Mortite on poly cones comes off easily, are these all paper?<

Yes, they are all paper.

>It would help if Miller measures his drivers and knows the typical T&S parameters for AR-9 woofers with his replacement parts.<

Mr. Miller assures me that he knows nothing of the T&S parameters and is merely a very careful speaker reconer. . . I think part of the key is that he has the right parts.

>Not sure I'm following, the inductor will effect the higher end of the woofer range where it crosses over to the mid. <

That is exactly what I was saying. We didn't consider the inductor for that reason.

>I'd be very curious to hear what you think, 2.5 with reduced damping, might be a good place to start to see if you can get them to a point where there's too much bass, then work back.<

I'll see what I can do. It might be a while before I can get those resistors.

Bret

[Pete B]

These replacement drivers are expensive and I can understand the frustration of them not being right and the lack of bass extension. I'm wondering if fixing the suspension stiffness problem and perhaps adding a bit of series resistance would be a complete cure. It's possible that bringing the low bass up, and the upper bass down is all that's needed but there could be cone breakup issues with the Tonegen, probably not but it's something to think about and measure if possible.

Wondering if these Tonegens can be salvaged in any way.

About the harshness with your 10pi's, harshness is usually too much output in the 2 to 5 K range where our ear is most sensitive, the opposite of the sweet sounding BBC dip. You mention also a hole, this sounds like a driver out of phase, could the mid be out of phase causing a hole in the lower crossover and in phase or a peaked up response in the upper crossover? It's worth looking into.

>>I'd like to read about what you heard.<

>

>Thank you for asking. I am happy to share whatever I

>experienced.

>

>The "Tonegens" (if that's what they are) are considerably

>louder and make lousy midrange drivers in these cabinets.

>

>Background: My 10pi's are too bright. I replaced all the

>capacitors with Solen (except the 2500uF) and they got better

>as far as "balance" is concerned, but got harsh, too. As I

>say, balance was better, but still not right. There seemed to

>be a "hole" in the lower midrange, upper bass. (I think it may

>be a stuffing issue)

>

>Thinking that I could overcome that problem by using what are

>clearly more efficient "Tonegens," I decided to remove the

>repaired 200003s (4 and 7) and install the replacements.

>

>The cabinets did hang-together better in the lower midrange,

>upper bass area. In fact, they "over-hung together" to a

>pucker-factor of about 8. Leaving the tweeters at "0"

>and turning the midrange drivers to -3db made them more

>tolerable, but still throaty when "gutsy" was required.

>Turning the midranges down further would have been useless as

>the upper midrange was now sucked-out with the accompanying

>loss of detail.

>

>What I didn't expect was the extent of the bass drop-out. The

>only way to get these speakers to skip a stylus across an LP

>would be to drop one of them on the corner of the turntable.

>The bass wasn't only "attenuated somewhat" it was out-cold.

>AWOL. True, the speakers were away from the walls, but they

>were flat on the floor in the 4pi position.

>

>With the old, repaired drivers in them, this position was

>tolerable, but only just. There was so much bass and so much

>mud in the bass that you felt like you needed to scrape it off

>the bottom of your shoes.

>

>With the Tonegens in there they still had plenty of "thump"

>but no "whuump." ZZTop kept it's "bang" and impact until a

>percussive effect should have made the couch shake. Not only

>did it not shake the furniture, it wasn't even audible.

>

>Male vocals lost all their chestiness and authority. Listening

>to Ladysmith Black Mambazo's "Shaka Zulu" was wholey useless.

>It sounded like they all had something in their throats. No

>chest. The periodic gigantic "WHUUUUMM" in Fleetwood Mac's

>"Emerald Eyes" just wasn't there. At all. It was like the

>recording had been edited.

>

>Piano sounded "far away" and unsatisfying. Supertramp's MFSL

>release of "Crime of the Century" was unlistenable.

>

>Also on Fleetwood Mac's "Mystery to Me" is a recording of "For

>Your Love." A few times on that track Mick must hit a big

>drum with a loose head using a very soft mallet. It was

>audible, but not in the proper octave and it certainly didn't

>move the Earth as it should. It was unidentifiable without the

>low register cues.

>

>And beyond mourning the loss of house-shaking low sounds, the

>midrange ended-up sounding colored and confused and at times

>like a fire-truck's horn.

>

>We listened to the same music (repeatedly trying to get a firm

>grasp of a few key elements) before and after sticking the

>Tonegen's into the cabinet, so we didn't listen to a great

>variety of things; it wasn't necessary. There was no part of

>the "Tonegen" experiment that went well for the replacement

>drivers.

>

>Without first having the benefit of doing any UniBox

>experiments I described the problem to my listening-buddy as

>"there is way too much in the 350-500Hz range, a huge bump

>somewhere around 70-100Hz and not much of nothin' below

>that."

>

>I'm sorry if it seems I've gone all "Stereophile" on everyone,

>but this isn't a nuance I'm talking about so I described "big"

>because what's missing is "big."

>

>They are too loud, they throw the whole cabinet out of

>balance, they have no umph, they don't seem "fast," just tight

>and stingy. Bizarrely they seemed to suck all the dynamics

>right out of the recording, even in frequency ranges where

>they don't operate.

>

>In fact, I'm going to feel guilty about sticking them on eBay.

> I've got about $520 + freight in them and need some of that

>back but I really can't see representing them as replacement

>woofers for AR speakers and can't imagine that they have other

>uses. . . refrigerator magnets? Bulk tape erasers? Ballast?

>

>

>And now, an editorial:

>

>In my personal opinion, those who sell these as drop-in

>replacements should be ashamed of themselves. You and I know

>that *they* know better. The flat sided baskets turn these

>replacements into AR woofers the same as a Halloween mask

>turns a sweet little girl into a lion. These drivers are

>sheep in wolves' clothing and everyone who buys them, in my

>personal opinion, is getting fleeced. (everybody sing, "Money

>for Mutton" by Dire Strifes)

>

>Personally, I think anyone who sells these under false

>pretenses should replace them with suitable drivers - which is

>what we asked for and paid-for and which these were

>represented as, right? I'd love to get a pair of these into a

>courtroom in a class action suit. Justice is blind, but she

>isn't deaf.

>

>Bret

>

>PS - "I want my, I want my, I want my low H Z"

[RoyC]

Hi Pete,

Coincidentally I called Millersound this afternoon and thats essentially what I was told. The stiffer spider saved AR "alot of hassles" as amplifier power grew and listening levels increased. It unfortunately altered the bass response when used in older systems designed with the soft spider.

He suggested replacing the spider as well. I will be delivering some woofers to him soon and will have him install spiders equivalent to the early 12 inchers in the Tonegens. We shall see ("hear" I should say).

Roy C.

[Pete B]

The higher Bl with for DUT7 is hard to explain, the very similar Rdc of the VC and Lvc of the coil indicate that it is the right coil. But it is possible that a newer VC is made with thinner high temp insulation or some newer fabrication method that packs the windings tighter, that would provide more turns in the gap for a higher l. Probably not just a theory. Because the coil is much longer than the gap centering doesn't change the small signal parameters, it is important when driven to large excursion where with an offset coil one direction will have greater linear throw.

>>I think the late 200003 used a stronger ceramic magnet, note

>that DUT4 has about the same Bl as 2 and 3, 7 could be a

>measurement error, or it could have the wrong voice coil

>(different l (as in Bl)- length of wire in the gap). <

>

>Could this be an artifact of not getting the voice coil at "0"

>in the gap?

>

>In other words, could glueing up a new surround with the voice

>coil either too much into, or not far enough into, the gap,

>cause this?

>

>Bret