[Aether Audio]

Friends,

Below is a little something to ponder.  In the past I've usually come right out and explained things in (almost excessive) detail.  This time I thought I'd have a little fun. 

The "newbies" to the hobby out there probably won't get it, but I'm sure there's a handful of "old birds" around these parts that will figure out the significance of the graphs below.  I'm just gonna set back and watch for a while and let those interested in doing so "guess at it."  After one or two of you "nail it,"  I'll chime back in to confirm.

These graphs are the result of the research I've recently undertaken in the development of the "new project" loudspeaker for the "unnamed company" I spoke about on the earlier Update thread I posted.  The resulting design is "nothing more" than a 2-way WTW design using two 6.5-inch woofers in a vertical array with a tweeter/6-inch waveguide arrangement nested in between them.

This new technology will be made the new standard offering of all Aether Audio "Millennial Reference Series" production, and former SP Technology products in the field of the "Mark III" vintage  will also be upgradeable to it.

OK... let the fun begin.  This first graph below speaks for itself with regards to the top Frequency Response plot.  And no... there was no "smoothing" of the graph used in generating any of the plots below (that's the first thing I expect to be accused of).  Assuming some "basic" knowledge on the part of the reader, the FR plot doesn't therefore take much insight to interpret.  A bit of "faith" maybe... but not much technical knowledge.

For those that might want to dig a bit deeper, take a good look at the Phase Response plot, and particularly the Group-Delay plot.  Question:  Based on these results, what might you suspect the result would be if the speaker represented in these plots were to attempt to reproduce a 1kHz square wave?    No... a square wave isn't music... it's MUCH more difficult to accurately reproduce than music.



Next we have a rather "ugly" looking plot.  Ahh... but "beauty is in the eye of the beholder."  Notice the cursor position at 527 Hz?  Why... just look at that "nasty" null.  it's at least -35dB!!!  That must sound horrible!  Well, it would... if you played the speaker that way.  But you see, I'll give you a little hint.  This plot was taken with the tweeter polarity reversed 180 degrees (terminal wires flipped) from that of the woofers.  So then... just what does this plot actually reveal? 

For those that truly grasp the significance of these plots, I need not say anything.  For the rest of you... you'll hear all about it soon enough... and before long, so likely will the rest of the audio world. 

Have fun! 
-Bob

[Albireo]

I clearly haven't been speaking to you often enough because I only heard about this through AC, but I guess that's clear evidence you really did get the tweeter down that low. Dang! And it appears that the resonant frequency of the tweeter is far above 30kHz? Whatcha got there, Bobbie? Is it "forever?" Oh yes ....

I'm curious to know what the 10-30 degree dispersion look like? Horizontal, of course: one of the best things about your waveguide designs is that you avoid all those nasty suckouts in the vertical direction that usual MTM designs suffer from. WTW, indeed.

In any event: congrats!

[bhobba]

Hi Bob and All

Thanks for posting that.  Nice bit of thinking cap stuff.

First as to the null when polarity is reversed thinking back to many many moons ago when I read Vance Dickinson that is a very very good thing - it indicates the phase response at crossover when it is not reversed is good - the bigger the null the better.  From what you posted it shows your phase response looks great.  And oh yea - since it is at 527 hz your crossover is now quite a bit lower.

For the square wave thing I need to think back to my Fourier theory (I did applied math not engineering so its how I look at these things) since your phase responce is so good above 1K and your amplitude response nearly flat the higher harmonics of the square wave will be reproduced correctly leading to very little effects like ringing etc so I expect the square wave to be reproduced very cleanly.

Don't know if its what you are after but its what sprang to my mind any way.  My conclusion is your new 'breakthrough' is something to do with the crossover and/or tweeter 'placement' to help time alignment at crossover and keep phase coherency.  Since Bob always used, IMHO, excellent crossovers I suspect its a technique to time align the tweeter better.  Then again it may simply be the crossover is now lower.  Maybe you are now using that great Seas air circ tweeter with perhaps some better time alignment technique?

Thanks
Bill

[DeadFan]

Hi Bob.

Don't understand it, but if it further improves your Revelation speakers, hope you can wack it in mine before their packed up. Don't want to have to buy'em again

Graham

[Aether Audio]

Abe & Bill,

Well... it looks like it didn't take you guys long - way to go! 

Your exactly correct about the square wave response and yep... crossover is 500Hz (527Hz to be exact in this case, although that will vary a bit from unit to unit).  THAT's from the little 6-inch waveguide as is used in the Timepiece Minis too.  In the past the larger 10-inch waveguide as used in the Timepiece 3.0 and the larger models always managed to allow me to crossover a bit lower than the Minis.  Specifically, the TP3.0s could be crossed at 600Hz while the Minis were crossed at 800Hz.

Now... the question is: Does that mean then "if" all things being equal, I can cross this new tweeter over at 300Hz in the TP3s?  Haven't done the testing yet to see, but if so... YIKES!!!   We're talking about a 1-inch ring radiator tweeter going all the way down to 300Hz?  That's just frick'n nuts.    It's too early to tell yet as I'm just speculating... but we already know it will be at least 500Hz.  Most would say that's still nuts! 

Not only that, but the goofy thing goes out flat past as far as I can measure to 25kHz.  The spec graphs state it doesn't start rolling off until about 30kHz and still has decent output at at 40kHz.  The only real pain in developing the system was that due to the fact the tweeter is a ring-radiator, I HAD to develop a bullet shaped phase-plug that mounts in front of the tweeter diaphragm and at the center of the waveguide.  Without the phase-plug there was a nasty null of almost -30dB at 14.5 kHz.  Come to find out after doing some research, they all pretty much have that problem and have to use a phase-plug to get rid of it. Even though it was a pain and complicates things, still it was well worthwhile.  Even though it was my first attempt at metal casting, it turned out pretty good... if I do say so myself. 

Oh... and please don't ask which tweeter it is.  Our agreement with the other company we're building this first design for prohibits me from stating publicly.

Now, I'm sure there are quite a few guys out there thinking that a 500Hz crossover is gonna stress the heck out of the tweeter, even though I managed to get it to go down that low.  Trust me... it's not stressed in the least.  I've included another graph below that should give you an idea as to what I mean.



The curve with the "big humps" is the tweeter mounted in the waveguide/phase-plug assembly before any crossover filters are added.  The second curve with the flat response is the final response after the full high-pass filter is applied.  As you can see, I have to attenuate the output by -20dB a 1.5 kHz, and to a lesser degree on either side of that frequency.  At the low end the attenuation is at least -10dB across the whole range.  Even though -10dB = 1/100th the drive current, some might still think that's too much juice and might result in too much diaphragm excursion.  What the graphs don't show though is the filter part values and the actual current vs. frequency.  I never got around to printing that graph out and I'm not about to give out the part values, but rest assured... the tweeter is getting very little drive current.  In fact, it would be just "coasting along" at SPLs that would certainly destroy pretty much any other tweeter.

The novice might be asking what all this amounts to right about now.  Well, the distortion arising in any driver is primarily the result of diaphragm excursion.  The more excursion there is... the more distortion there is as well.  While there are secondary sources of distortion in any driver, excursion related distortion almost always dominates and by a wide margin.  The upshot is that if you can significantly lower excursion related distortion, you're gonna end up with a much better sounding speaker at lower volumes and one that offers far greater dynamic range capability.  That's been just one of the hallmarks of our designs from the beginning, but now with this new tweeter/waveguide design... 

So that's the scoop.  All Millennial Reference Series products will employ this new tweeter/waveguide from this point forward.  Older units depending on vintage can be upgraded to it as well.  We're making some other changes in the near future as well, so unfortunately we're also gonna have to raise prices again before long too.  Even still, for world-class performance such as this, our prices will still be at least 1/3rd that of any equivalently performing system... assuming you can find one at all.  I mean, how many companies out there are offering a 2-way speaker that crosses over at 500Hz (or lower) and that offers almost unlimited dynamics, extreme low-level resolution, controlled dispersion, has virtually no diffraction artifacts, and disappears and sounds as seamless/coherent as the best single-driver designs?  Heck... we won't even talk about the bass performance either! 

Take care, 
-Bob

PS.  Graham... we'll build them into your Revs and Abe... I'll get some dispersion graphs before too long as well.

[konut]

To those that know how difficult this is to achieve, this is pretty exciting news.

So that's the scoop.  All Millennial Reference Series products will employ this new tweeter/waveguide from this point forward.  Older units depending on vintage can be upgraded to it as well.   

Does this apply to the Mini? What would be the cost of a retrofit?

[Albireo]

Abe & Bill,

...

PS.  Graham... we'll build them into your Revs and Abe... I'll get some dispersion graphs before too long as well.

No doubt I didn't say anything that was new to anyone, but the original post on XO/tweeter/dispersion was me. Thanks for the info though, and congrats again.

- KJ

[ooheadsoo]

Yep, that was Albireo up there, but trust me, I'm drooling right now.  Very VERY interesting stuff!

[Aether Audio]

Kuan Ju & Abe,

DUH - sorry guys... I got your monikers mixed up.   Just another sign of old age setting in I guess. 

Konut,

Yeah... all of the Minis ever built were of the newer waveguide vintage, which means they can all be upgraded.  It's the old black textured "2.1" vintage stuff that can't be.  Those old waveguides had a higher "Q" gain profile, so instead of trying to attenuate 20dB as I am now (which is one tough nut to crack to begin with), I'd be looking at trying to attenuate somewhere near 30dB using this new tweeter/waveguide/phase-plug system.  That just ain't gonna happen.  Talk about trying to hold a bubble down under water with the tip of your finger! 

I haven't really had time to figure out a cost of an upgrade yet.  In fact, I'm gonna have to build one for each model first so i can work out the parts needed.  This first project used two woofers, so tweeter shelving attenuation will have to be worked out to match a single woofer, plus the woofers are different than in the Minis too... so the low-pass is gonna be different too.  I am working on the Minis first though, so I should have some kind of idea in about a month or so.  I will say that based on my own sense of costing (I tend to think everything is too expensive), seeing that this new design is FAR more complex than ever before... it's not going to be what I would call "cheap."  Just to give you an idea, this first design has 35 parts in the entire crossover.  By my counting, that's a LOT of parts for a 2-way design.  Yeah, some of those are parallel caps to get the final values, but still... that's a LOT of frick'n parts!

In the mean time, keep  and...

Take care, 
-Bob

[Russell Dawkins]

do mine eyes deceive me or is that 90 dB/W/M shown on the top graph on this page? Is this speaker designed with a sub in mind, or is intended to be run full range? That's a pretty high sensitivity for a long excursion capable 6.5" driver - even a pair - isn't it?

Good going, Bob.

[konut]

35 PARTS???? HOLY RUBE GOLBERG Batman!  Does that include Black Boxes too?  I had a bad feeling that if if I had to ask how much.......     Oh well, I'll burn down that bridge when I come to it.

[bhobba]

Bob this is unbelievably fantastic news.  Getting the crossover this low is simply mind-boggling and the benefits distortion wise will undoubtedly be there for all to hear.  Not knowing exactly what the new tweeter is isn't really an issue - it was just a guess it may have been the saes air circ.

I am really really itching to get my new mini once my money situation abates bit.  It will fully abate when I get my super in November but in the interim I may be able to give Bob the go ahead to complete them (I already have made a down-payment but asked Bob to hold off until my monetary situation is better) and take delivery a bit sooner.  Basically it may require me to go into a bit of debt so I will need to think this through.  My sister reckons do it - you only live once and since I am now retired enjoy it as much as possible.  But I am a bit cautious about these things by nature.

When people ask me why I am so enthusiastic about your speakers I explain that one of the most critical  - if not the most critical - area of reproduction is the midrange.  You have developed a technique using waveguides to allow low distortion tweeters to play this critical region.  Now you have it even lower it can only make things better.

Again Bob congratulations on a job well done. 

Thanks
Bill

[isaeagle4031]

This is my first post here, but I have been watching Bob and his waveguides for a bit.  Being as I only live 10 miles from his shop helps too.    Gotta get out there and havea talk with him about some of this new stuff!  Amazing and genious Bob!

PS Bob in case you forgot who I am, the Name is Chuck Morris

[Russell Dawkins]

Bob - one question I keep forgetting to ask; what are the advantages, from your perspective, of going lower and lower with the crossover frequency to the HF driver?

[Aether Audio]

konut,

I can build the BBs into the enclosures, but as of now that's not standard.  What with all the parts in this new design, I'm not looking forward to the job of finding room for them inside either.   

Bill,

As always... thanks!    And I hope things work out for you.  Believe me when I say - "I understand." 

Chuck,

It's been a while, but sure I remember.  Give me a call whenever you get a chance and stop on by.  I always look forward to folks visiting.  Heck... we're neighbors for Pete's sake! 

Russell,

Thanks again for the support my friend!    You know me... I'll do my best to answer any questions, although you may regret asking! 

OK then, what are the advantages of pushing the X-O lower?  Well, there are several the way I see it.

First of all there's the issue of Time-Domain performance.  "One driver does all" is the theoretical ideal in that regard.  I'd reference the "square wave" comments above.  Getting the fundamental tones and all of their harmonics to line up in time represents one very difficult step closer to perfection.  Although even my designs don't perfectly achieve this goal, now at least we're looking at virtually ALL of the harmonics of a given instrument remaining intact with respect to each other upon arrival to our ears.  Seeing our hearing is most sensitive in the "above 1kHz" range, my reasoning is that by pushing the X-O this low, at least we've covered the most critical area.

Then there's the issue of woofer/midrange cone distortion (break-up modes).  A 1-inch tweeter diaphragm is simply too small for the typical "standing waves" (that we see develop in woofer cones) to propagate in.  If you study the break-up modes that arise in most woofers, you'll find that they look very similar to the standing waves that develop in acoustic spaces (listening rooms).  Sure, their wavelengths and decay times are shorter, but often their frequency and decay characteristics are much alike - especially the shape of their "envelopes."  I once saw a review by Bruce Bartlett of an ETC measurement he did on a famous brand electrostatic speaker.  The frequency and decay envelope looked just like the ETC response (RT60) of a typical reverberant room.

Well, all that "hash" is nothing more than standing waves developing in the membrane and bouncing around.  The edges where the membrane is supported are the "walls" where the energy just reflects back across and around the periphery of the membrane until it finally "decays" away.  As in a typical room, the number of "modes" is almost unlimited as you not only have reflections occurring across the major dimensions of the membrane, but also "orthogonal" modes where they travel around the periphery and what have you.  I guess that's why I've never cared for planer speakers... they're just too under-damped for my tastes.  They may be "fast" in responding to an initial impulse, but then there's a ton of energy that comes out of them later in time that has nothing to do with the original signal.

Sorry, that's not meant to be a slam to planer lovers, that's just my take on them.  Although they aren't typically as bad in that regard, most cone drivers have essentially the same problem.  Seeing then that a 1-inch tweeter diaphragm is simply too small physically for such modes to develop (in the midrange frequencies), if we can get it to cover more of the midrange (by lowering the crossover frequency) we should have a lower distortion reproduction. 

Don't get me wrong, a tweeter diaphragm can (and virtually always does) have the same problem, but such artifacts are "pushed" well up out of the critical midrange area into the +5kHz region.  Although, in the specific case of a ring-radiator there are even fewer such modes and they reside even higher yet in frequency due to the fact that they don't use a simple dome diaphragm.  Also, seeing that we use rigid, "piston-like" aluminum woofer cones, this second advantage isn't as big of a deal.  These cones don't flex in the way most paper, plastic and other types do, so they don't suffer from the break-up modes nearly as much to begin with.  They do "ring like a bell" out around 5 kHz, but that's well above where we use them.  In fact, in this newest case that's a full decade above our crossover frequency.  Nevertheless... every little bit helps.

Finally, the third major advantage of lowering the crossover as far as you can go is that it reduces Inter-Modulation Distortion that arises within the woofers as a result of woofer cone excursion.  The more any woofer "pumps," the more its lower frequencies "modulate" the higher frequencies it's trying to reproduce at the same time.  To be sure, if you can actually "see" your woofer cones moving while reproducing music, you can figure you're hearing a significant amount of IM distortion as well.

IM distortion is particularly nasty in that the spurious artifacts generated are not harmonically related to the original tones that gave rise to them, so it definitely doesn't impart any form of "euphonic" effect.  Well, if we then lower the crossover, the woofers don't operate as high into the midrange to begin with... so "modulation" and hence IM is gonna be reduced.  At first listen, most folks will notice this reduction in distortion as the most prominent effect.  They may not have a clue as to "why" the speaker sounds better, but it's clearly audible regardless.

So... that's my story and I'm sticking to it.  You'll know better to ask another question next time. 

Take care, 
-Bob

[bhobba]

So... that's my story and I'm sticking to it.  You'll know better to ask another question next time. 

No way Bob - all this technical stuff is friggen great.  Keep it up.  As I have posted elsewhere I am not a bit fan of this listening is everything stuff.  You and your ears can be easily fooled IMO (and sadly I have been).  Not only must it sound good but it must be backed up by hard nosed engineering.  Well done Bob.

Thats not to say listening tests are useless.  For example I am very impressed with Hugh Deans amplifiers at Aspen.  He starts out with well engineered designs but does the final 'tweaking' by ear.  Of course that final tweaking is crucial in setting his amplifiers apart from the rest.

Thanks
Bill

[bbchem]

Bob, when might we here news of the new speaker design?? When might they be available for purchase, I have been selling stuff in hopes to get back my Minis, but will wait for the new company?? 

Bill

[Aether Audio]

Bill,

To be honest, I'm not sure when this other company will be officially releasing the product for sale.  We're still getting the loose ends together on the little details like boxes and logo plates, etc.

In the mean time, all Aether Audio products will be offering this same tweeter/performance.  We just got a new batch of MDF in and are already making Timepiece 3.0 enclosures for several orders we have.  They will be built using this new tweeter/waveguide as well.

Also, seeing we are offering all our models in a satin painted finish, we are building an extra set or two while we're at it just in case we do get an order.  Now that we've caught up on all of the old orders except the 2 pair of Revelation Grand Masters, we are finally free to work on new orders for the first time in almost a year.

Seeing we offer the painted finish version now, all hardwood orders will be more of a "special order" case and incur the lead time needed to get the wood.  In the past that was always a hassle waiting to get the stuff delivered.  With the "paint over MDF" enclosures, we won't have that problem so lead times will be a lot shorter than in the past.

Oh yeah... and we use to have "bugs" in the CNC parts programs that slowed us down too.  We didn't have the correct software for our machine for the longest time after we got it (figures - bought it used  ), and never had the funds to buy it outright.  So... my son had to "hack" another program designed for a different machine to even be able to use the thing.  Well, his hack-job just barely worked, so once a program was written it took FOREVER to de-bug as it generated erroneous G-Code that had to either be deleted or changed manually (via notepad)... one line at a time.  

OK, so certain parts programs might have as many as 1,000 lines of code and he'd have to check every line to make sure it was correct.  Well, once a program was written... he wasn't too motivated to go back and change it if we discovered the part wasn't coming out correctly.  So, instead he/we would just re-machine the parts by hand on a table saw or router table.  Needless to say, that added a LOT of time to the whole process.

Technically, we should have gone back and changed the programs instead, but by the time we discovered the problems we would often already be behind schedule.  Besides, not realizing until we went to assemble the enclosures, a few times we had cut a whole bunch of parts that all had the same problem and we couldn't afford to throw out the material and start over - especially with the hardwood parts.  It wasn't long and the whole thing snowballed and we found ourselves so far behind that we never thought we'd get caught up.  Due to the above and the lack of new orders resulting from the economic recession, the combination spelled the end for SP Tech.

Since SP Tech shut down though, my son has had the time to fix the software to where it generates good G-Code now and it doesn't have to be de-bugged anymore.  If he does need to make a change in a part, it might take all of an hour from beginning to end.  Before it would take the better part of a week.  Back under SP Tech, a lost week due to a single part change just wasn't an option, so on we plodded at a snail's pace... and with ever decreasing funds to work with as we went along.

But... thank God... those days are now over.  Heck, we've made more progress in the last month than we did in practically 6-months under SP Tech.  We've already shipped one batch of the new project speakers, have started and are well under way on another, and have also started making Tinmepiece 3.0s.  With the funds the new project speakers are going to net us, we'll actually be able to build enclosures in advance of receiving any orders - at least the smaller enclosures anyway.  In fact, as I said... we've already started doing so. I don't think we're going to commit the resources to building any of the larger models at first though, until we get further ahead of the game.

In closing, that's the scoop reagrding both the new project speakers and our own production.  If anybody wants a pair of Continuum 2.5s or Revelations, you're still going to have to wait a while, but Continuum A.D.s, TP 3.0s and TP Minis will have a pretty quick turn-around... assuming you order a painted finish.

Take care,
-Bob

[bbchem]

How would you compare the smallest Aether series to the Timepiece Minis?

Dynamics, imaging, sound stage?

Has anyone bought a pair and reviewed them??

Bill