ForA-9 Speaker Concept Based on the AR-9, by Pete B.

[Pete B]

Steve F.

Your comments here show that you know very little about driver selection. The M15CH001 is very similar to the upper midrange in the NHT3.3. You demonstrate with your arrogant comments that you know very little about speaker design and why this is in fact an upper midrange driver. You'll have to pay my consulting rate if you want to be educated in the matter.

Pete B.

>

>BTW, that SEAS M15CH001 midrange has a free-air resonance of

>54Hz, and looking at the supplied curve, it appears to respond

>cleanly down below 100Hz. This is clearly not an

>"upper-midrange"-only driver and it illustrates the difficulty

>of preconceiving a particular complex design and then

>searching around for the bits and pieces to fit the puzzle.

>The AR-9 used a 1 1/2" dome upper mid for reasons of wide

>dispersion. That was an engineering objective. The AR

>upper-mid dome will have far better dispersion in the range

>from around 3000-7000Hz than this SEAS will, which is why AR

>used that driver. It met their objective. The 8" lower mid was

>then the perfect solution to having a driver that was robust

>enough to take over from the woofs at a low-ish 200Hz and

>respond smoothly up the 1200Hz x-o point to the upper mid.

>Each driver was designed and utilized for a specific reason,

>to accomplish a specific task within the system.

[Guest Sean]

Very nice post Barry. Without completely re-designing my 9's ( HT mains ) and 90's ( HT rears ), i've done everything that you've stated above and slightly more. Obviously, i was not able to stagger the drivers / alter the crossover for proper phase / time coherence, change baffle width, play with a transmission line design, etc... but everything else was addressed.

To be specific, all of the internal wiring was replaced with Kimber, wired point to point and directly soldered. For the woofers, i used 8TC and everything above that is connected with 4TC. The crossover parts were all upgraded ( caps & resistors ) and consolidated down into one board. The padded attenuators and all of their associated wiring were all bypassed. If one studies and listens to the AR 9's frequency response / tonal balance, it is quite evident that these speakers produce the greatest output in the woofer region. As such, further attenuation of the lower mids, upper mids and / or treble region would only compound the existing tonal balance, shifting the speakers further away from neutrality. As such, doing away with all of the associated attenuation circuitry at those higher frequencies really got rid of a lot of unnecessary clutter and made wiring a LOT easier. I did retain the switches in the front of the cabinet for cosmetic purposes though.

Through careful layout and altering the way that the coils were situated, i was able to keep crosstalk via stray inductive coupling down to a minimum. This is true even though some of the parts are actually closer together physically than they were in the original design. Consolidating all of the parts down to one board and getting rid of all of the switching / wiring drastically reduced the length of the signal path with quite a few less connections. The upgraded and hand selected caps were yet another step forward in terms of transparency and resolution. The installation of suitable binding posts were also mandatory, doing away with the horribly "cheap & plastic" things that the factory used.

The baskets of the woofers and the mid-woofers were physically damped using multiple layers of "Liquid Nails". This DRASTICALLY improved performance via the reduction of ringing. This also helped to clean up some of the "harshness" that these speakers exhibit in the upper mids, as the basket on the 8's rings like a bell once excited. Damping the metal leads on all of the caps also helped in this area. If one doubts this, they should take hold of a "Solen" type cap and "flick" the lead with their fingers. The amount of ringing that eminates from that lead is highly audible. While soldering this lead to another part of the circuit does lower the resonant frequency due to increased mass, that ringing is still evident when listening. As such, i took steps to mechanically damp the capacitor leads at the point of entry into the body of the cap itself.

As far as the structure itself goes, additional bracing was added along with sound damping sheets at various points of the cabinet. The binding posts were moved to the bottom of the cabinet, where the lone crossover board now rests. Cones were added at all four corners, which easily elevates the cabinet enough to allow for clearance of the wiring / binding posts. I am still playing around with the type and amount of damping material used internally. When doing mods such as this, using the stock quantity and type of damping material will result in reduced low frequency extension due to the reduction of internal cabinet volume due to the use of additional bracing, etc... Finding a suitable balance that achieves good low frequency extension and proper damping without creating small ripples in the upper bass region can be a tricky thing though, hence my further experimentation.

The end result of all of this is what most would consider to be a completely new speaker. The sound is FAR more open sounding with improved transient response, greater transparency and improved midrange "liquidity". The "slightly thick" and "boxy" sound associated with AR's is completely gone without losing any of the tight and extended bass that they were known for. If anything, bass authority has improved due to the shorter and heavier gauged signal path that i've created. On top of that, the speakers now seem easier to drive as the sound no longer seems like it is "straining" to get out of the box at higher volumes. The 90's were modified in similar fashion and my center channel, which is a custom built piece using two AR 8's and an AR mid / tweeter plate, is also similar in design.

When i first started running these, i tried several different passively bi-amped combo's. Prior to my current installation, i had settled on running a Perreaux PMF-2150 ( 400+ wpc @ 4 ) into the upper section and a Perreaux PMF-3150 ( 500+ wpc @ 4 ) into the woofers. Anything less than this amount of power didn't do the speakers justice, even though they would play quite loud with a "decent" 250 - 400 wpc stereo amp. In order to get obtain the spl's that i was looking for though, most amps were straining to get there. The two Perreaux's did the best out of everything that i had tried up until that point and the bottom end with the 3150 driving the speakers was AWESOME to say the least.

After further experimentation, i found that a Sunfire Signature ( factory rated at 1200 wpc @ 4, actually clips at 1480 wpc ) was a better match sonically, but it just didn't have the "thunderous" bottom end that the Perreaux did. After sending the amplifier into thermal shut-down mode two times ( and popping an upper midrange dome ), i sent Sunfire Signature back to the factory for further modifications after discussing the situation with them. This specific amplifier now has the current limiters pulled from it, the thermal shut-off's raised and the power supply has been stiffened via additional filter capacitance. This gave me the bottom end that i was looking for. After getting the amp back from them and performing further listening tests, i've also replaced the internal speaker wiring and changed the capacitor in the negative feedback network. This helped to open up the mids and top end a bit more, increasing transparency and high frequency airiness.

Between this amp driving the 9's as my HT mains and a Sunfire Cinema Grand Signature feeding the 90's, my center and two low Q sealed and stuffed subs, my HT system can perform double duty for both music and movies with great power and finesse.

As such, and i can say this based on first hand experience, as good as the AR 9's are in stock form, they can still be made to sound much better. Those that own 9's and are thinking about changing speakers should really check into simply optimizing what they already have. Doing so will net them HUGE gains in performance AND save them gobs of money. Trying to replace a set of fully modified AR 9's with a speaker of similar performance / design integrity would cost them thousands upon thousands upon thousands of dollars. In all likelihood, those speakers probably wouldn't be as good of an over-all performer either. Sean

>

PS... I've performed many of the same type of mods to other speakers with similar results. The last pair that i really went to town on were my Father's Legacy's. These responded MUCH better to mods than the AR's did, but that's because they were a lesser design to start off with. As a side note, it is a little known fact that Bill Dudleston of Legacy fame was EXTREMELY influenced by the design of the AR 9 and 9 LS. If you doubt this, look at the similarities between the 9 LS and the original Legacy 1's. The 9 LS used a 12" front firing woofer with a smaller, ten inch down-firing woofer to supplement the bottom end. The Legacy 1's used a smaller cabinet with a 10" front firing woofer and a smaller, eight inch down firing design to supplement the bottom end. Given that very few speakers were using down-firing designs at the time, and the similarities in terms of being a four way, vertically aligned array, one can't help but see the influence that AR had on Dudleston's early work.

Many years ago, the prior owner of my AR 90's took them down to Legacy's showroom in Springfield to compare them to some Legacy models side by side. When Dudleston saw these speakers, he ended up taking them apart and studying their design while the customer waited. Obviously, his curiosity was pretty high when it came to AR designs.

To take that a step further, you can even see the similarities in the older Legacy catalogues to AR's marketing literature when they were selling the "Vertical series" ( 9's, 90's, etc... ) of speakers. If i could post pictures of both catalogues for you to examine, you would think that the same marketing team designed them. Not only are the cosmetics similar, but both try to offer quite a bit of technical info as to how & why the speakers are designed the way that they are. The fact that Dudleston went for quantity of bass via porting over the quality of bass that AR was able to achieve with their sealed designs was one of his major downfalls.

Having said all of that and after hearing the modified Legacy's and AR's, there's no comparison between the two. The AR's stomp all over the Legacy's. Then again, we are talking about comparing the original to a copy, so what did you expect ????

[kkantor]

Well, I assume you have seen the basic published data for each of the drivers you list, and that is why you picked them. They are each fairly well documented by their makers. In addition, you might want to try www.thielesmall.com. If you need any unpublished data about the Vifa or SS units, of course I will try and get it.

Since there are many, many, many good drivers around, and an infinite number of crossover circuits to connect them, it would be useful to have some specific goals in mind to start the selection process. It is doubtful that the general driver layout of the 9 has all that much to do with the sound. Something, yes, but maybe 30%.

My 2 cents:

I had just started at AR when the 9 was under development. A great deal of research was done on the audibility of "time alignment." This was also about the time when Preis and Bloom, et al, were researching the matter. While many things could be improved about the 9, or any speaker, I'm convinced that "time alignment" is not one of them.

My other 2 cents:

Steve is a pretty bright guy, who knows a lot about speakers. I suspect he misinterpreted your intentions in his first posts, but his latter post seems to correct this.

Ken Kantor

www.tymphany.com

www.d-s-t.com

[Guest Bret]

Sean,

Out of curiosity, what did you do with / about the 2500uF cap?

Bret

[Steve F]

Pete,

Geez! I went from "Thanks for your comments" to being "arrogant and offensive," all in the span of one day. Wow—-that’s a pretty dramatic fall from grace.

Before you carry out my execution, let me say my last words from a condemned man. I wasn’t questioning your or anyone else’s motives or abilities. I was simply trying to point out that people can and do have different feelings and approaches to these kinds of projects. If it please the court, your honor, I did say that "Both rationales are valid." I think that’s reasonable, no? People are allowed to have their personal take on a given situation.

As to my background, I have been "on the job" in this industry in technical/director positions for close to 30 years with many companies that you would be extremely familiar with. I’m sure many people you know have owned or currently own products for which I’ve been responsible. I do not divulge my identity or the companies I’ve worked for (or am currently working for), because I don’t want any of my comments on this forum to ever be thought of as having an ulterior motive—which they don’t. I was a hobbyist long before I was professionally involved in this business, and my involvement here is strictly as a hobbyist/historian.

That’s all I have to say in my defense. I’d like a nice NY sirloin, medium, with baked potato, butter, bacon and chives, with a glass of single-malt, neat, for my last meal. And hand me that blindfold before you shoot. I don’t want to watch.

Steve F.

[ar_pro]

Pete B. -

Please make an effort to read the posts more carefully - nobody's being "slammed", and there's nothing "arrogant" about dissecting an insupportable/incorrect statement.

Your initial offering posited "a modernized design following the AR-9 concept" - everything proceeds from that statement, and all comments/suggestions/opinions have flowed from there.

Maybe you could attempt to more successfully explain your goal (and what "excellent characteristics" and "keeping up" signify)...please feel free to make it as *technical* as possible, and we'll all try very hard to stay with you.

[soundminded]

A lot of speakers have been cloned or used as the starting point for someone else's deisgn. Some have been successful, and some not so successful. I think most people here would say that the Cello clone of the AR LST did not perform as well. OTOH, when Paul Klipsch marketed his Klipschorn in the 1950s, a lot of me too designs came out including the JBL Sheffield which used the folded corner horn concept but built it with some of the finest drivers available, JBL's own, not the drivers Paul Klipsch bought from University Sound. This is nicely explained on the Lansing Heritage Site. JBL moved on while Klipsch still has a version of Klipschorn at the top of its line fifty years later. There is nothing wrong with stealing, plagerizing, learning from other people's ideas. I've read Tim Holl's excellent description of the AR9 on this site several times and I strongly recommend that you read it if you haven't already. I think understanding the concept of a speaker is very important while duplicating its details is practically insignificant. There is usually more than one way to arrive at the same results.

Tim's explanation shows how AR addressed problems which other manufacturers never solved or probably even thought about and the elegant solutions AR arrived at. There are many other problems which were not solved and went completely unaddressed. Lessons learned from AR LST for example were apparntly forgotten or ignored in the AR9. I have yet to hear a loudpeaker which sounded "accurate" when installed the way the manufacturer built it and recommends installing it, AR9 included. IMO, AR9 out of the box falls far short of what I consider the most important goal of any high fidelity loudspeaker and that is the accurate reproduction of the tone of acoustic musical instruments on most commercially produced recordings. As I see it, when there is a discrepency between what you measure and what you hear, my instinct is to distrust the measurements because the measurement concept failed to take into consideration important variables which affect perceived performance. It took me five years before I found a solution which worked for me to satisfy my expectations of this speaker. Playing with the performance parameters is to me like adjusting a color television set, when you get the controls set just right, you know it because everything pops into place. It just sounds right. And every time I've had to replace a component in my sound system because it failed and I decided not to repair it, once a power amplifier, another time a cd player and even my equalizer itself, it took another two years of tweaking to get back to where I had been. What we hobbyists lack in megabuck laboratories, we make up for in enthusiasm and the availability of time. Experimenting with different drivers and crossover designs is to be expected. AR9 as we know it was not their first try at this project, it was their last and you can be sure there were many many alternative designs which were tried and failed, some of them probably pretty good.

I have recently received a Steinway grand piano, a 5'-7" M built in 1927. It is a magnificent instrument. It sits at one end of my music conservatory, my AR9s sit at the other. Of all the recordings I own, only one has a piano which has the same distinctive sound of this one. Tonally when played through the AR9s as I have them set up, it is a dead ringer. But it does not have the same presence which IMO is due to the entirely different spatial radiating characteristics of the loudspeakers versus the piano. I think this is a very significant problem which few manufacturers have addressed. One who did was Bose in (what I consider) his flawed 901 design. Too bad he didn't explore the possibilities of his direct/reflecting principle much further. That's another one I intend to try, probably as a three way design. I've already experimented with my original 901s as a two way bi-amplified project which I've reported about here on the "other" board, and have gotten the tonal balance very close to the way I have the AR9s set up. Therfore my ultra ambitious project I'd call ARTD9901LST stealing the best ideas from all three of them and adding a few of my own. It would be bristling with drivers, about 23 in all. In my current conception I'd use four Tonegen 1259s side mounted in a 13 cu ft base enclosure of about 3' x 3' x 1-1/2'H. For upper midranges, I'd take a hard look at Hi Vi Research DMN-A 2" dome which Parts Express is offering at a remarkable discount of $28 instead of the list price of $118.50. I'd have to be inventive to build the annulus which creates the semihorn loading above 4khz. I'd use an array of these dome midranges and an array of tweeters as LST did and I'd have two back panels just as 901 did. I'd also make provision for adjusting the reflected hf level and contour to compensate for absorption due to room acoustics. Having the woofers near the floor while having the midrange/tweeters at ear level is one big advantage of this kind of tower design.

I have no problem with using multiple amplifiers, active crossovers, and equalizers. Professionals don't require or expect loudspeaker systems to be flat out of the box, only that they are equalizable and that's what I also think is important. For me a louspeaker system doesn't stand alone, it is part of an integrated system. Anyway, when I do get started, I'll let you know how I'm doing. Good luck on your project, sounds like you are also in the early stages.

[Guest radkrisdoc]

>I don't see your problem with the SEAS 8", I will resize the

>enclosure if necessary and I doubt it will be necessary. Have

>you done a simulation of it in 3 liters?

I think it is pretty obvious that the SEAS driver will need something more than 3 L to perform satisfactorily. By satisfactorily I mean a target Q of around 0.5 to less than 1. Since you may have access to modeling software more complicated than WinISD, please simulate an acceptable enclosure and tell me what volume you get. My guess is that it should fall around 20 to 35 litres. I think that is big if you are designing a speaker system within one single enclosure, as it will turn out to be one huge box when you accomodate for two woofers as well. That is exactly why the AR-9's 8" lower mid has a volume of (roughly) 3 liters.

>Do you know Fc and

>Qtc for the AR 8" in the AR-9 enclosure?

No, I do not have access to AR blueprints; I doubt anyone here has one of those 8" lower mids with original foam surround for us to measure. Therefore, the only option left is to make an educated guess. Since I thought everyone here would understand that situation, I did not take up that part and explain. It is left for us to guess, and my guess is, since the AR-9 was the flagship speaker from AR at that time, it should be well designed and if that is the case, Q should fall around .5 to 1 (I think it will be a max of .8, a Q of 1 does not have that good a transient response). As for resonance, the AR-9 manual states that the 8" lower mid has a resonance of 175 Hz in its enclosure. That is a fact which I am quoting from the manual. Even with that information, I dont think we need to duplicate it down to the minutest details for this design. All we need is a driver with practical volume considerations. Unless you are building separate enclosures for the woofers, of course. I think the 7" driver from Vifa's PL line is an excellent choice when you take volume/Q into consideration. It will fit in a small enclosure easily (less than 5L) and deliver good performance. We can, on the other hand use that SEAS midbass, but at the expense of increasing enclosure size by at least double that of the AR-9's LMR. Its a choice which you need to make, but I've made my point. My choice would be the Vifa 7" PL driver. I've even thought of using it in my AR-90, using a mounting adapter so there are no modifications done on the cabinet, keeping it original.

The AR-9 attaining idol status as stated by others here seems to be a statement that has outlived itself. Sure, it is an excellent design, but the drivers are all degrading, decomposing, rotting, falling off! If idol status has to continue, either us lot need to do something or the AR-9 series production started again. Something that is highly unlikely to happen should happen, something akin to a miracle. If not, fond memories will be all that's left.

As far as I am concerned, refoamed/restored drivers are just new drivers on old frames. If you go by the book and expect everything to be done the way it should be, then all refoamed drivers need to be measured, thiele-small parameters recorded and enclosures designed accordingly if the drivers do not have original characteristics after the refoam. Its on everyone's mind, how close to the original is a refoamed driver? I am not willing to even take the chance without knowing that. Kinda like sheep in a wolf's clothes. I would rather redesign or reverse engineer using factory fresh drivers. And that makes this whole thread interesting.

[tysontom]

>Hi Bret,

>

>I explained in Kris' UMD thread that small cone drivers

>provide much more VD and often have as good if not better

>dispersion because the dust cap acts as the radiator at high

>frequencies. This is supported by the fact that there are not

>many mid dome drivers available, only one 1.5" that I know of.

> There is little demand for them.

>

>Pete B.

>

>>Why wouldn't you use a dome upper midrange driver? The

>>problem of matching a 12" to a 1.5" dome is pretty-well

>>eliminated by adding the lower midrange driver, isn't it?

>>

>>Bret

>

Pete,

Are you trying to say that a (hypothetically) 4-inch cone driver (volume-displacement issues notwithstanding) with a domed dust cap would have "as good as if not better dispersion at high frequencies" as a 1.5-inch dome tweeter? Are you implying that the dust cap decouples from the cone, and the cone ceases to vibrate once you get into the highest frequencies?

--Tom Tyson

[roundhome]

Hello Mr Kantor

May I invite you over to the other forum i'm on its a quad one the best minds about quad and other formats on there.

Thank you

JIM

http://www.quadraphonicquad.com/forums/forum.php?

[Guest Sean]

"Out of curiosity, what did you do with / about the 2500uF cap?"

Bret, that cap and the 470 uF's used were a tough one that i had to make a decision about. What i ended up doing was using multiple caps to achieve those specific values. In the long run, i ended up with a 2000 uF & 400 uF electrolytic bypassed with an 80 uF and 24 uf metallized polypropylene caps for the 2500 uF cap. The 470 uF was replaced with a 250 uF & 100 uF electrolytic bypassed with an 82 uF and an 8.2 uF metallyized polyester cap. While this was "less optimum" than what i had wanted to do, the cost of going with all "high grade" caps would have been exhorbitant. On top of that and in case you haven't seen them, high grade poly type caps are absolutely HUGE compared to an electrolytic of similar rated value.

As such, i ended up with good quality electrolytics bypassed with poly's as mentioned. This gave me a reasonable combo of size and cost to work with. Performance wise, i'm 100% certain that using all poly's would have given me improved transient performance with increased transparency, but it would have come at the expense of low frequency extension. That is, the amount of box volume that 2500 uF's & 470 uF's worth of poly's would have eaten up would have been equivalent to adding several more internal braces, which i had already done and lost cabinet volume from. I just didn't think it was a worthwhile trade-off and as mentioned, the cost was also a factor.

As mentioned above, each cap was hand selected. I did this after each cap was brought up to max rated voltage and cycled several times. This allows the cap to fully form and "settle in". I am a firm believer in "break in" of passive components and this pre-conditioning allowed me to see how each cap would have ended up working after extensive usage in a circuit. It also allowed me to achieve specific values taking into account mass production parts tolerances, etc... That is, if you add up some of the rated values mentioned above, they don't total up to the value that they were supposed to replace. In all actuality though, they do equal that value and with a very tight tolerance due to hand selection.

What i did was to purchase a large quantity of each value needed and used those that demonstrated the least loss with the closest values to what i needed. Comparing the charge and discharge rates and noting the measured value of capacitance after forming allowed me to pick and choose the cream of the crop. Since i was also doing a set of AR 90's and a similarly designed center channel at the same time, buying in bulk made not only a lot of sense, but was also more economical for me too. Not only did i measure each cap individually, i also measured them in ganged form when trying to achieve "oddball" values. In doing so, i was able to achieve uniform values for each specific application from speaker to speaker.

After performing the testing that i did with these caps, it was a rather eye opening experience. That is, anyone that tells you that "a cap is a cap" knows not of what they speak. The differences between various types of caps ( electrolytics, poly's, etc... ) and even within the same type of cap from the same manufacturer was quite obvious. In plain English, the amount of internal loss ( dielectric absorption ) and series resistance ( ESR ) between a "high grade" cap and an electrolytic is staggering to say the least. That is, electrolytics are SO slow and lossy that their use in anything but a power supply circuit is extremely questionable to me. If it were not for the problems mentioned above i.e. size and cost of other higher grade caps, i would never use electrolytics again.

I used to have a link to a page where an EE had measured and charted the linearity of various types of caps across the audible bandpass, but i can't locate it right now. All i will say is that there are advantages to combining various types of caps in order to achieve a higher overall level of linearity with less of a given sonic signature from any one type of cap. In effect, the performance advantage of one type of cap can help to "band-aid" or compliment the flaws of another type of cap and vice-versa. The end result is two non-linear devices working together in order to achieve a higher level of linearity than either could have achieved by themselves.

As a side note, the crossover schematic of my 9's did not match that of the word document that is shown here at the Classic Speakers website. My speakers had an extra inductor in each of them. Needless to say, had i followed the schematic posted here and / or was trying to duplicate that crossover without having one in front of me to study, i would have been in bad shape. I can't remember the specifics, but i would HIGHLY recommend that one verify exactly what it is they are working with PRIOR to dis-assembly and / or rebuild. Sean

>

[kkantor]

OK, thanks. I'll take a look around.

>Hello Mr Kantor

>May I invite you over to the other forum i'm on its a quad one

>the best minds about quad and other formats on there.

>Thank you

>JIM

>http://www.quadraphonicquad.com/forums/forum.php?

Ken Kantor

kkantor@gmail.com

www.tymphany.com

www.aural.org

[kkantor]

If I remember correctly, the NHT 3.3's upper-mid was derived from a SEAS MPC12VC. This is a 4.5" mid with a conventional motor on it. Wicked flat, and really low distortion. The actual effective radiating area is 62 cm^2, making it the equivalent of a 3.5" dome. The NHT unit wasn't quite SEAS stock... there were some changes to adhesives, etc, designed to improve power handling.

>selection. The M15CH001 is very similar to the upper midrange

>in the NHT3.3.

>

[soundminded]

From the PDF files, it appears to me at first glance that the most suitable crossover frequency for the Seas upper midrange and the Vifa tweeter is 3 khz. These are the recommended frequency response limits by both manufacturers. Although the upper mid extends above this frequency, it appears that its on axis response is starting to rise slightly while its off axis response is falling suggesting this is where it is starting to become directional. Any comments on this. I don't see anything magical about the 7.5 khz crossover of AR9 except that it was probably most suited for their choice of drivers.

BTW, I don't see anything wrong with sitting down with a bunch of catalogues and picking out drivers as likely candidates for early on fleshing out the plans for executing a concept. Unless you are going to design your own drivers, you'll have to do this sooner or later anyway.

[Pete B]

Hi Ken,

Sorry, I should have been more specific, I meant for the original AR-9 drivers. I thought maybe you knew who might be holding the old measurements and data. This would be helpful for all of these classics.

I agree about the driver locations, and would probably keep the driver locations approximately the same, but would move the mids and tweeter closer, and lower the UMD to tweeter XO to 3 to 4 kHz.

I tend to agree with you about time alignment and have read many of the papers.

I will take your comment about Steve under advisement!

Pete B.

>Well, I assume you have seen the basic published data for

>each of the drivers you list, and that is why you picked them.

> They are each fairly well documented by their makers. In

>addition, you might want to try www.thielesmall.com. If you

>need any unpublished data about the Vifa or SS units, of

>course I will try and get it.

>

>Since there are many, many, many good drivers around, and an

>infinite number of crossover circuits to connect them, it

>would be useful to have some specific goals in mind to start

>the selection process. It is doubtful that the general driver

>layout of the 9 has all that much to do with the sound.

>Something, yes, but maybe 30%.

>

>My 2 cents:

>

>I had just started at AR when the 9 was under development. A

>great deal of research was done on the audibility of "time

>alignment." This was also about the time when Preis and

>Bloom, et al, were researching the matter. While many things

>could be improved about the 9, or any speaker, I'm convinced

>that "time alignment" is not one of them.

>

>My other 2 cents:

>

>Steve is a pretty bright guy, who knows a lot about speakers.

>I suspect he misinterpreted your intentions in his first

>posts, but his latter post seems to correct this.

>

>Ken Kantor

>

>www.tymphany.com

>www.d-s-t.com

>

[Pete B]

I notice that several people here seem to look for exact answers, what I say is generally true and yes what you state below agrees with my thinking. Note that it is true for carefully designed drivers that I find of interest, there are probably exceptions out there but those are not what I'm talking about.

My statement was aimed more at domes, currently available, that I would consider using. If I had to use a dome UMD today it would probably be the Morel 2", or I might look at the large 28mm, low Fs, dome tweeters from Scan, Usher, or NorthCreek. Most of those tweeters have 1mm one way Xmax and might have higher VD than the 1.5" Scan dome. The 1.5" Scan is .5mm Xmax IIRC. The likely best dome mid is the Morel and this 2" is what I had in mind when I made the statement.

I came to this conclusion over 20 years ago when I obtained detailed driver specifications for most of the SEAS product line. SEAS does not have dome mids in their product line today so I'll use similar examples from Vifa. Vifa only has a 3" mid dome and here's the FR at 0, 30, and 60 degrees off axis:

http://www.d-s-t.com/vifa/data/freq_d75mx-31-08.gif

Here's a Vifa 4" cone driver:

http://www.d-s-t.com/vifa/data/freq_m10md-39-08.gif

Here's the Scan 1.5" dome which is not what I had in mind, but I'll show for reference, note that the off axis response may appear better because of the rising high end on axis:

http://www.d-s-t.com/scs/data/freq_d3806_820000.gif

Here's a Scan 28mm dome tweeter just as another point of reference:

http://www.d-s-t.com/scs/data/freq_s2905_970000.gif

The point is that small cone drivers are not nearly as bad off axis as the diameter of the cone might suggest. Note the destructive notch in the 60 degree off axis response at about 4.5 k, 15 k, 10 k, and 28 kHz for the 3", 4", 1.5" and 28mm drivers this suggests that the effective piston diameter is smaller for the cone driver than the actual cone diameter. It can also be seen that notching begins at about 7 k, and 19 kHz for the 4" and 28mm drivers.

I'd plan to use a 3 to 4 kHz crossover between the UMD and tweeter in this version. I usually design for a 2 to 4 kHz XO with higher order designs.

Yes, I know the SEAS UMD has a phase plug.

Pete B.

>Pete,

>

>Are you trying to say that a (hypothetically) 4-inch cone

>driver (volume-displacement issues notwithstanding) with a

>domed dust cap would have "as good as if not better dispersion

>at high frequencies" as a 1.5-inch dome tweeter? Are you

>implying that the dust cap decouples from the cone, and the

>cone ceases to vibrate once you get into the highest

>frequencies?

>

>--Tom Tyson

[Pete B]

I posted this concept because I think it is a good starting point and most of the driver selection comes from drivers that I'd already looked at in the past. It could lead to nearly a clone, or improved clone, not sure what the right wording is for this, or a highly improved version. I mentioned quad amping, digital XOs, and room correction which are just a few possibilities.

You seem interested in reflected or ambient sound and perhaps some ideas could be incorporated from the AR-MGC where the reflected source is delayed in time as it should be. This direction is not currently a major interest of mine but I have looked at it in the past. I have the AES article on this, BTW, you might find it and the AR-MGC of interest. There are crossover schematics for the MGC on this site.

I might as well address some of the basic advantages. There is no doubt that it will have improved bass response since these drivers are well known in the industry and there is good agreement with simulation when reliable driver data is used. Several asked what problem does this solve, and we've heard the ongoing problems from Bert and Roy of finding replacement woofers for the AR-9. We also just heard of someone burning out an AR-9 woofer and therefore there are at least some users who need a more robust low end. These drivers will provide that in thermal capacity, freedom from bottoming, and longer life due to their operating well below their maximum throw. The midranges use advanced low distortion motor design (several have been confirmed through independent testing) and were chosen for this reason.

I agree with you concerning the "out of the box" voicing of AR speakers and it is probably one reason why I've never bought any AR speaker products. Several have pointed out the live vs. AR speaker demonstrations which may have been impressive but a slightly different voicing is required IMO when speakers are moved from a stage or open air, into the typical home environment. I would also plan to voice them for use away from the wall in order to provide better imaging.

Good to hear about your Steinway, do you or one of your family members play? I would think so, but you never know, they sure do look nice! I've seen your posts that IMO idealize what goes on in the studio, whereas my experience and several recording engineer friends confirm that studio work is "all over the place" nearly all recordings are EQed, compressed, limited, chorused, reverbed, doubled and more. I also see audiophile purists who seem to think the original recording is some sort of perfectionist production which couldn't be further from the truth for the majority of the available recordings. There are no standards and therefore studio work and mastering is more of an art. I believe that this is one of the significant weaknesses in home audio reproduction today. This ties into your Steinway because you could make your own recordings, move the AR-9s near the piano and attempt some live vs. recorded tests.

Yes, I read the AR-9 writeup years ago, as well as Allison's AES articles. I put much of his measured data into a spread sheet for analysis. I've not heard all of the speakers following this design practice, but I have heard Allisons on several occasions and IMO they do not have any clear advantage based on what I heard. This is not very reliable since I heard them without being able to experiment with position and so on.

Pete B.

>A lot of speakers have been cloned or used as the starting

>point for someone else's deisgn. Some have been successful,

>and some not so successful. I think most people here would

>say that the Cello clone of the AR LST did not perform as

>well. OTOH, when Paul Klipsch marketed his Klipschorn in the

>1950s, a lot of me too designs came out including the JBL

>Sheffield which used the folded corner horn concept but built

>it with some of the finest drivers available, JBL's own, not

>the drivers Paul Klipsch bought from University Sound. This

>is nicely explained on the Lansing Heritage Site. JBL moved

>on while Klipsch still has a version of Klipschorn at the top

>of its line fifty years later. There is nothing wrong with

>stealing, plagerizing, learning from other people's ideas.

>I've read Tim Holl's excellent description of the AR9 on this

>site several times and I strongly recommend that you read it

>if you haven't already. I think understanding the concept of

>a speaker is very important while duplicating its details is

>practically insignificant. There is usually more than one way

>to arrive at the same results.

>

>Tim's explanation shows how AR addressed problems which other

>manufacturers never solved or probably even thought about and

>the elegant solutions AR arrived at. There are many other

>problems which were not solved and went completely

>unaddressed. Lessons learned from AR LST for example were

>apparntly forgotten or ignored in the AR9. I have yet to hear

>a loudpeaker which sounded "accurate" when installed the way

>the manufacturer built it and recommends installing it, AR9

>included. IMO, AR9 out of the box falls far short of what I

>consider the most important goal of any high fidelity

>loudspeaker and that is the accurate reproduction of the tone

>of acoustic musical instruments on most commercially produced

>recordings. As I see it, when there is a discrepency between

>what you measure and what you hear, my instinct is to distrust

>the measurements because the measurement concept failed to

>take into consideration important variables which affect

>perceived performance. It took me five years before I found a

>solution which worked for me to satisfy my expectations of

>this speaker. Playing with the performance parameters is to

>me like adjusting a color television set, when you get the

>controls set just right, you know it because everything pops

>into place. It just sounds right. And every time I've had to

>replace a component in my sound system because it failed and I

>decided not to repair it, once a power amplifier, another time

>a cd player and even my equalizer itself, it took another two

>years of tweaking to get back to where I had been. What we

>hobbyists lack in megabuck laboratories, we make up for in

>enthusiasm and the availability of time. Experimenting with

>different drivers and crossover designs is to be expected.

>AR9 as we know it was not their first try at this project, it

>was their last and you can be sure there were many many

>alternative designs which were tried and failed, some of them

>probably pretty good.

>

>I have recently received a Steinway grand piano, a 5'-7" M

>built in 1927. It is a magnificent instrument. It sits at

>one end of my music conservatory, my AR9s sit at the other.

>Of all the recordings I own, only one has a piano which has

>the same distinctive sound of this one. Tonally when played

>through the AR9s as I have them set up, it is a dead ringer.

>But it does not have the same presence which IMO is due to the

>entirely different spatial radiating characteristics of the

>loudspeakers versus the piano. I think this is a very

>significant problem which few manufacturers have addressed.

>One who did was Bose in (what I consider) his flawed 901

>design. Too bad he didn't explore the possibilities of his

>direct/reflecting principle much further. That's another one

>I intend to try, probably as a three way design. I've already

>experimented with my original 901s as a two way bi-amplified

>project which I've reported about here on the "other" board,

>and have gotten the tonal balance very close to the way I have

>the AR9s set up. Therfore my ultra ambitious project I'd call

>ARTD9901LST stealing the best ideas from all three of them and

>adding a few of my own. It would be bristling with drivers,

>about 23 in all. In my current conception I'd use four

>Tonegen 1259s side mounted in a 13 cu ft base enclosure of

>about 3' x 3' x 1-1/2'H. For upper midranges, I'd take a hard

>look at Hi Vi Research DMN-A 2" dome which Parts Express is

>offering at a remarkable discount of $28 instead of the list

>price of $118.50. I'd have to be inventive to build the

>annulus which creates the semihorn loading above 4khz. I'd

>use an array of these dome midranges and an array of tweeters

>as LST did and I'd have two back panels just as 901 did. I'd

>also make provision for adjusting the reflected hf level and

>contour to compensate for absorption due to room acoustics.

>Having the woofers near the floor while having the

>midrange/tweeters at ear level is one big advantage of this

>kind of tower design.

>

>I have no problem with using multiple amplifiers, active

>crossovers, and equalizers. Professionals don't require or

>expect loudspeaker systems to be flat out of the box, only

>that they are equalizable and that's what I also think is

>important. For me a louspeaker system doesn't stand alone, it

>is part of an integrated system. Anyway, when I do get

>started, I'll let you know how I'm doing. Good luck on your

>project, sounds like you are also in the early stages.

>

[Pete B]

I've stated for years online that my preference is 2 to 4 kHz mid to tweeter in most systems. I was certain that I'd be moving this crossover point and the plan was for 3.5 kHz.

Re: sitting down with a bunch of catalogs is fine if that is what a person actually does. Today even those with a modest lab can do extensive testing.

Pete B.

>From the PDF files, it appears to me at first glance that the

>most suitable crossover frequency for the Seas upper midrange

>and the Vifa tweeter is 3 khz. These are the recommended

>frequency response limits by both manufacturers. Although the

>upper mid extends above this frequency, it appears that its on

>axis response is starting to rise slightly while its off axis

>response is falling suggesting this is where it is starting to

>become directional. Any comments on this. I don't see

>anything magical about the 7.5 khz crossover of AR9 except

>that it was probably most suited for their choice of drivers.

>

>BTW, I don't see anything wrong with sitting down with a bunch

>of catalogues and picking out drivers as likely candidates for

>early on fleshing out the plans for executing a concept.

>Unless you are going to design your own drivers, you'll have

>to do this sooner or later anyway.

[soundminded]

I haven't played the piano since I was a child but I'm thinking of taking it up again. My sister plays and teaches violin, piano, viola and can also teach cello. I am very familiar with the sound of these instruments and have become a bit adept at knowing who made them by the way they sound. This Steinway is in a room which is fairly symetrical. The acoustics affecting the speakers is similar to the acoustics affecting the piano. It is not practical to place the speakers next to the piano. Besides the speakers are almost certainly far more sensitive to their exact placement than the piano is. I'm also familiar with the sound of most other "acoustic" instruments, especially clarinets as I played one of those as a child as well. I've been priveleged to hear a lot of live music in my life and to be surrounded by musicians at one time or another. I've also been priveleged to hear some priceless violins close up on many occasions including a Guarnari del Jesu and an Amati. It is astonishing how loud the Guarnari can play.

The drivers you have selected as a starting point seem to be of very high quality and should result in a very fine loudspeaker. It would not surprise me at all if the ultimate result is comparable or in some ways superior to AR9. You not only have the advantage of the AR9 paradyme to learn from but 27 more years of advances in all of the technologies from computerized mathematical modeling, measurement, and availability of hardware at a time when great strides were made in all of them.

"I've seen your posts that IMO idealize what goes on in the studio, whereas my experience and several recording engineer friends confirm that studio work is "all over the place" nearly all recordings are EQed, compressed, limited, chorused, reverbed, doubled and more."

If I've given that impression then I've communicated in a way where my views have been badly misinterpreted. They say you don't want to see laws or sausage being made. I'd add recordings to that. I thought I said many times that they don't put all of those knobs, switches, and sliders on mixing consoles for nothing, they get used a lot. The recording engineers who use them to sell sound for a living don't have any of the reservations audiophiles do about twiddling those knobs. IMO most forms of music with the possible exceptions of classical and jazz are "manufactured sound" made to conform to what is most commercially sellable. Many of those so called artists sound awful live without their engineers and equipment. I think there is a more honest effort to "document" performances of classical music and jazz but the engineers are not above twiddling knobs there either if they feel that it is in the service of creating a product which more closely resembles their idealized conception of a live performance than what they've gotten down on tape. I think in general, collectors of classical music are more demanding that way.

My brief experience with Roy Alison's products have not impressed me either. Perhaps I didn't give them a fair chance. I also don't have a particularly high opinion of the voicing of the Original Advents, a very popular speaker. The best of the lot at the time seemed to me to be KLH 6. That speaker came closest to what I heard live.

"Several have pointed out the live vs. AR speaker demonstrations which may have been impressive but a slightly different voicing is required IMO when speakers are moved from a stage or open air, into the typical home environment."

I was priveleged to attend two such demonstrations, one with a guitarist and one with an nickelodeon, both at consumer trade shows in NYC and both using AR3s (the Nickelodeon also used AR4s for a comparison as well.) I considered that these highly contrived demos showed what was possible but their real world performance reproducing commercially made recordings fell far short of expectations. I had a chance to do a fair amount of comparison of KLH 17 with AR3 while I was still in school and while the AR3 was far superior in the lowest bass, the voicing of the KLH 17 sounded much more to me like live music in the rest of the range.

"This ties into your Steinway because you could make your own recordings, move the AR-9s near the piano and attempt some live vs. recorded tests."

It is unlikely that this idea will work. The acoustics of the room will be almost impossible to exclude from the recording and therefore a double acoustic would make the comparison unfair. The piano would have to be recorded in an anechoic or near anechoic environment. I think that this is what AR did and it also explains the mistake other companies made when they tried similar experiments. And this is the point. Even with live insturments in a room at home, the effect acoustics play on what you hear is enormous. Most instruments spatially radiate their sound entirely differently from the way loudspeaker do and the subjective effect is profound. The single exception which comes to mine is the human voice whose higher frequency components are highly directional just like most loudspeakers (I think that's why their intelligibility is so diminished when they talk facing someone else.)

I'm glad to see you didn't give up on this thread. After it started out, it seemed to be quickly heading in the wrong direction.

[Pete B]

Hi Ken,

I just noticed how many DST products are being mentioned here and the idea for a demonstration model for DST and Tymphany products just crossed my mind. Good excuse for building a demonstration model and going to trade shows. Ken, is there a Tymphany woofer that mounts in place of a 12", or could an adapter be easily made? I don't think I've seen any details concerning these new woofers. Digital crossovers and quad amping make rapid prototyping relatively easy so that new products could be easily incorporated as they come out:

Patented dual concentric ring radiator, with waveguide tweeter:

Scan R2904/700000 4 ohm, 1" ring radiator, SD-2 neodymium motor, non resonant aluminium chamber, multiple chamber low compression design, machined aluminium faceplate and phase plug

http://www.d-s-t.com/scs/data/foto_r2904_700000.jpg

4.5" upper mid, (or perhaps the 1.5" dome) low loss linear suspension, sliced paper cone, compact linear motor:

Scan 12M4631g00, 4 ohm:

http://www.d-s-t.com/scs/data/foto_12m_4631g00.jpg

6.5" long throw lower mid, low loss linear suspension, sliced paper cone, SD-1 motor:

Scan 18W4531g00, 4 ohm:

http://www.d-s-t.com/scs/data/foto_18w_4531g00.jpg

Is there a Tymphany woofer that could be used?

Pete B.

>An ambitious project, please let me know how I can help!

>You'll never quite make another AR9, but I don't think that's

>the point. Speaker building is fun, and there is much to be

>learned from the process. Who knows... maybe even a new

>"classic" speaker design will result.

>

>http://ar.us.biz.yahoo.com/bw/050310/95961_1.html

>

>http://stereophile.com/news/0301405tymphany/

>

>

>Ken Kantor

>

>kkantor@gmail.com

>

>www.tymphany.com

>

>www.aural.org

[kkantor]

Boy, I'm having a hard time tracking messages. I think I need to go back to the "old" BBS format.

1- OK, I understand your question about the AR driver specs. I do have data on the 12", but have long since lost track of the others. I'll ask around. I might even have some vintage replacement drivers in storage, but I can't promise that. If I find some, I'll test them and post here. Also, if anyone has a collection of Speaker Builder back issues, I seem to remember the 9's drivers having been analyzed at some point.

2- DST has always avoided demonstrating complete speaker systems at sales events, as do most larger OEM suppliers. It's a tricky issue when you have many customers with many different design philosophies. Besides, there are many commercial designs, from B&O to Sonus Faber, that use the DST stuff, available for easy reference.

3- Have you checked out www.madisound.com's discussion group? It's a great resource (with appropriate mental filters applied). Also, I'm thinking that they often have contents and regional events at which you could show off your baby.

4- Nothing from Tymphany for the aftermarket yet, but thanks for the consideration! Everything currently under development is for specific OEM contracts. Plus, the 9 had a very particular and (to me) extremely successful approach to the bass. Probably the design's most enduring aspect. For this reason, my first instinct is to go with dual, sealed 12's. I realize this last stuff is editorial on my part, but I can't resist....

-k

[Pete B]

Thanks very much for anything you can do concerning old AR driver specs didn't mean to cause you any work. I don't remember the AR-9 article in Speaker Builder, I'll look into it.

I wasn't sure if companies like DST avoided it or not, they're products are used in enough systems so I can see their perspective.

Yes I'm familiar with the Madisound board and plan to go to some local events. Don't know if I'll even follow through on this particular design.

I was thinking your own company might want to have some sort of demo platform, just throwing the idea out there. I'm sure you've thought about it yourself.

I have a good feel for what's going on with the 9s bass XO circuit and will probably simulate it at some point.

Thanks,

Pete B.

>Boy, I'm having a hard time tracking messages. I think I

>need to go back to the "old" BBS format.

>

>1- OK, I understand your question about the AR driver specs.

>I do have data on the 12", but have long since lost track of

>the others. I'll ask around. I might even have some vintage

>replacement drivers in storage, but I can't promise that. If

>I find some, I'll test them and post here. Also, if anyone

>has a collection of Speaker Builder back issues, I seem to

>remember the 9's drivers having been analyzed at some point.

>

>2- DST has always avoided demonstrating complete speaker

>systems at sales events, as do most larger OEM suppliers.

>It's a tricky issue when you have many customers with many

>different design philosophies. Besides, there are many

>commercial designs, from B&O to Sonus Faber, that use the DST

>stuff, available for easy reference.

>

>3- Have you checked out www.madisound.com's discussion group?

>It's a great resource (with appropriate mental filters

>applied). Also, I'm thinking that they often have contents

>and regional events at which you could show off your baby.

>

>4- Nothing from Tymphany for the aftermarket yet, but thanks

>for the consideration! Everything currently under development

>is for specific OEM contracts. Plus, the 9 had a very

>particular and (to me) extremely successful approach to the

>bass. Probably the design's most enduring aspect. For this

>reason, my first instinct is to go with dual, sealed 12's. I

>realize this last stuff is editorial on my part, but I can't

>resist....

>

>-k

[Pete B]

Correction, must have been tired or a memory lapse most of the big 28mm dome tweeters are .5 mm Xmax one way, not 1 mm as I stated below. The Scan 28mm are .4 mm one way and the 1.5" is also .4 mm one way and it would therefore have significantly higher VD than the 28 mm tweeters.

Pete B.

>If I had to use a dome UMD today

>it would probably be the Morel 2", or I might look at the

>large 28mm, low Fs, dome tweeters from Scan, Usher, or

>NorthCreek. Most of those tweeters have 1mm one way Xmax and

>might have higher VD than the 1.5" Scan dome. The 1.5" Scan

>is .5mm Xmax IIRC. The likely best dome mid is the Morel and

>this 2" is what I had in mind when I made the statement.

>

>Pete B.

[Pete B]

You make many assumptions about how the 9s or 90s must be, I've learned from measuring commercial products that general rules of thumb that your mentioning are not the norm. The Qtc of the original driver in the original enclosure might be very high to partially compensate for the baffle step. You say that the refoamed AR 8" is an entirely new driver which I strongly disagree with. The significant differences are the compliance or Vas and possibly Qms when a poor refoaming job is done, otherwise the driver will be very close to original. Vas is not critical when the enclosure compliance dominates and Qms is not usually very critical. I suggest you measure your driver in free air and in the system, otherwise your guessing at the correct driver parameters. If I was to continue more along the clone path I would want to measure the AR 8" driver. You seem to have very specific goals of finding a drop in replacement, which is not my goal for this design concept but I'm willing to help. I asked you for the DC resistance of the drivers and inductors but got no response.

I thought you mentioned that you purchased the factory replacement 8" drivers, have you measured them?

You mention the Vifa 7" PL and there are many other possibilities, the Scan 7" drivers come to mind and Peerless HDS are just a few. We need to determine the voltage sensitivity of the original driver so that we can determine which drivers are applicable.

I'm not worrying about the Qtc of the SEAS 8" until I know what the original 8" in the 9 was. Since Fc is very close to the crossover frequency the second order acoustical rolloff is a dominant part of the complete transfer function. If the driver Qtc is too high an overdamped electrical filter can be used to compensate. This can be optimized in CALSOD.

Pete B.

>>I don't see your problem with the SEAS 8", I will resize

>the

>>enclosure if necessary and I doubt it will be necessary.

>Have

>>you done a simulation of it in 3 liters?

>

>I think it is pretty obvious that the SEAS driver will need

>something more than 3 L to perform satisfactorily. By

>satisfactorily I mean a target Q of around 0.5 to less than 1.

>Since you may have access to modeling software more

>complicated than WinISD, please simulate an acceptable

>enclosure and tell me what volume you get. My guess is that it

>should fall around 20 to 35 litres. I think that is big if you

>are designing a speaker system within one single enclosure, as

>it will turn out to be one huge box when you accomodate for

>two woofers as well. That is exactly why the AR-9's 8" lower

>mid has a volume of (roughly) 3 liters.

>

> >Do you know Fc and

>>Qtc for the AR 8" in the AR-9 enclosure?

>

>No, I do not have access to AR blueprints; I doubt anyone here

>has one of those 8" lower mids with original foam surround for

>us to measure. Therefore, the only option left is to make an

>educated guess. Since I thought everyone here would understand

>that situation, I did not take up that part and explain. It is

>left for us to guess, and my guess is, since the AR-9 was the

>flagship speaker from AR at that time, it should be well

>designed and if that is the case, Q should fall around .5 to 1

>(I think it will be a max of .8, a Q of 1 does not have that

>good a transient response). As for resonance, the AR-9 manual

>states that the 8" lower mid has a resonance of 175 Hz in its

>enclosure. That is a fact which I am quoting from the manual.

>Even with that information, I dont think we need to duplicate

>it down to the minutest details for this design. All we need

>is a driver with practical volume considerations. Unless you

>are building separate enclosures for the woofers, of course. I

>think the 7" driver from Vifa's PL line is an excellent choice

>when you take volume/Q into consideration. It will fit in a

>small enclosure easily (less than 5L) and deliver good

>performance. We can, on the other hand use that SEAS midbass,

>but at the expense of increasing enclosure size by at least

>double that of the AR-9's LMR. Its a choice which you need to

>make, but I've made my point. My choice would be the Vifa 7"

>PL driver. I've even thought of using it in my AR-90, using a

>mounting adapter so there are no modifications done on the

>cabinet, keeping it original.

>The AR-9 attaining idol status as stated by others here seems

>to be a statement that has outlived itself. Sure, it is an

>excellent design, but the drivers are all degrading,

>decomposing, rotting, falling off! If idol status has to

>continue, either us lot need to do something or the AR-9

>series production started again. Something that is highly

>unlikely to happen should happen, something akin to a miracle.

>If not, fond memories will be all that's left.

>As far as I am concerned, refoamed/restored drivers are just

>new drivers on old frames. If you go by the book and expect

>everything to be done the way it should be, then all refoamed

>drivers need to be measured, thiele-small parameters recorded

>and enclosures designed accordingly if the drivers do not have

>original characteristics after the refoam. Its on everyone's

>mind, how close to the original is a refoamed driver? I am not

>willing to even take the chance without knowing that. Kinda

>like sheep in a wolf's clothes. I would rather redesign or

>reverse engineer using factory fresh drivers. And that makes

>this whole thread interesting.

[Pete B]

>The drivers you have selected as a starting point seem to be

>of very high quality and should result in a very fine

>loudspeaker. It would not surprise me at all if the ultimate

>result is comparable or in some ways superior to AR9. You not

>only have the advantage of the AR9 paradyme to learn from but

>27 more years of advances in all of the technologies from

>computerized mathematical modeling, measurement, and

>availability of hardware at a time when great strides were

>made in all of them.

The soundcard based PC measurement systems have made it possible for people with modest labs to do as good and sometimes better than the big guys. It's a new ball game as compared what one could do 15 years ago with modest equipment.

>"I've seen your posts that IMO idealize what goes on in the

>studio, whereas my experience and several recording engineer

>friends confirm that studio work is "all over the place"

>nearly all recordings are EQed, compressed, limited, chorused,

>reverbed, doubled and more."

>

>If I've given that impression then I've communicated in a way

>where my views have been badly misinterpreted. They say you

>don't want to see laws or sausage being made. I'd add

>recordings to that. I thought I said many times that they

>don't put all of those knobs, switches, and sliders on mixing

>consoles for nothing, they get used a lot. The recording

>engineers who use them to sell sound for a living don't have

>any of the reservations audiophiles do about twiddling those

>knobs. IMO most forms of music with the possible exceptions

>of classical and jazz are "manufactured sound" made to conform

>to what is most commercially sellable. Many of those so

>called artists sound awful live without their engineers and

>equipment. I think there is a more honest effort to

>"document" performances of classical music and jazz but the

>engineers are not above twiddling knobs there either if they

>feel that it is in the service of creating a product which

>more closely resembles their idealized conception of a live

>performance than what they've gotten down on tape. I think in

>general, collectors of classical music are more demanding that

>way.

OK, yes I had the wrong impression based on a few of your posts, I agree with most of what you say above.

>"This ties into your Steinway because you could make your own

>recordings, move the AR-9s near the piano and attempt some

>live vs. recorded tests."

>

>It is unlikely that this idea will work. The acoustics of the

>room will be almost impossible to exclude from the recording

>and therefore a double acoustic would make the comparison

>unfair. The piano would have to be recorded in an anechoic or

>near anechoic environment. I think that this is what AR did

>and it also explains the mistake other companies made when

>they tried similar experiments. And this is the point. Even

>with live insturments in a room at home, the effect acoustics

>play on what you hear is enormous. Most instruments spatially

>radiate their sound entirely differently from the way

>loudspeaker do and the subjective effect is profound. The

>single exception which comes to mine is the human voice whose

>higher frequency components are highly directional just like

>most loudspeakers (I think that's why their intelligibility is

>so diminished when they talk facing someone else.)

Close micing is the standard way to minimize room effects. A carefully placed cross pair of mics can sound very good.

Pete B.