[Danny Richie]

Okay, we hit on cabinets and connectors.

On to drivers?

[SoCalWJS]

Tweeter or Woofer first?

I know that it requires a synergy, but do you go with one or the other first, or strictly the whole picture?

I feel that the mid bass to the upper vocal range is the thing that I notice first in a speaker. I've run across many manufacturers that get this wrong IMHO.

There are several tweeter types that are good (but I am not a fan of the metallic cone ones as a general rule) and add enough of the upper frequencies and "air" or whatever you want to call it. If it's bright?

Fugetaboutit.

[Cheeseboy]

I thought we covered drivers with jtwrace.  They are either fast or loose.  They can be measured as such. 

I really liked the Vifa Ring tweeter in my old Signature Rocket 850's.  It seemed to be crossed over lower into the upper bass area and that worked well for me.  I don't remember the name of the driver but the tweeter in the old ADS products was one of my favorites as well.

[Jonathon Janusz]

Just trying to nudge back to the core of the discussion (as right now I don't have a lot more constructive to add being more a learner than teacher in this thread):

We have started to discuss drivers in laying the ground work for the discussion terminology and concepts so we're all (more or less) on the same page speaking a similar language.  Now, focusing on what makes the grade in a top shelf mini monitor in particular, a couple questions:

I'm going to go out on a limb and assume that pretty much anything goes as far as tweeter designs, as long as the tweeter fits in the basic mold of the size/shape of speaker we're discussing (meaning huge room filling horns, maybe not so much. . .)?

Regarding woofers, how big is too big to call it a "mini-monitor"?  By extension (ha!), I guess that requires us to also touch on how low in the bass frequencies is acceptable?  The wish list posted earlier started with a suggestion that mid-30's with the help of subs was the target overall, but how about in the speaker itself, without subs?  Is mid-30's out of line to expect out of a top-shelf mini monitor?  40's?  60's?  80Hz?

Another line of questioning (and as Danny has said he doesn't mind, I'm glad to use his N2X design as an example by which to kind of pick a bit of a fight, to use as a template for discussion of other drivers hereafter I hope): 

Straight to it - consider the BG Neo3PDR.  Would you consider this a driver to be used in a "top shelf" design, or better yet why would this driver NOT be considered to make the grade?  What in particular makes it either pass or fail the test?  Similarly, how about the M165X?  Same line of questioning.  How about the popular (almost ubiquitous?) RAAL?  Accuton?  Pick a flavor from ScanSpeak? 

Thoughts from the designers/manufacturers as to what made them choose a driver in particular in what they would call a "statement" design of theirs would be great insight too.  Was it just measurements?  Synergy with another driver or crossover scheme?  Enough customers storming the shop with pitchforks that even though the driver didn't really "measure up" in the lab, it would have been crazy marketing-wise NOT to use it in something, just to be able to say on a brochure "speaker model XXX using YYY tweeter and ZZZ woofer", YYY and ZZZ being most clamored-for models?

 

[Vapor Audio]

10 hours of making sawdust, then 3 hours of replying to customer emails ... wish I had energy to reply 

[persisting1]

10 hours of making sawdust, then 3 hours of replying to customer emails ... wish I had energy to reply 

That's a good problem to have 

[sharpsuxx]

Wow great thread...good to hear all of the engineers in on the same thread.  Lots of knowledge being thrown around here; just finished reading the whole thread up to this point real page turner.

Let me summarize what I have gathered out of it so far and try to frame the driver debate.

Staying consistent with a traditional monitor design, the Cabinet must be acoustically equal to zero.  No noise introduced from edge diffraction, resonance, or back wave reflections; essentially engineered out of materials that are both high in mass, and with certain acoustic absorption properties.  The cabinet should be made of multiple different laminated materials to better transfer acoustics into other forms of energy to provide the optimal cabinet.

The speaker terminal should be made with we'll say "care" so that they do not introduce any complications into the design by adding distortion of any sort.

Now to the drivers, I think from an engineering perspective it is necessary to outline what a realistic frequency response will be from a mini-monitor speaker.  You do not want to challenge the existing technologies by making individual drivers cover too wide of a frequency range and losing what they do well so these can obviously not be full range, I would say it is expected to need a sub blended with these speakers to truly make them full range.  Personally I think most good drivers can adequately cover the Midrange down to about the 60hz range without struggling mightily to produce sufficient SPL.  Then a few more questions emerge, 

Knowing certain engineering compromises will need to be made for a mini-monitors extension to improve quality through the functional range, should a subwoofer be assumed in the design of the monitor?

If so which is the better roll-off to a sub at 60hz?  Steeper - sealed design, more gradual - ported design?  (I think this is the last question about cabinets)

As far as characteristics of the woofer I think low moving mass is important, very nimble and resolving drivers for the midrange with plenty of impact through the functional range.  For the tweeter, speed should be equally important resolve detail but without being fatiguing. 

Coherence and refinement of the sound are the most important to me for an uber speaker, so really it is a combination of selection of the tweeter and midrange and implementation with the crossover that will really create the voice of the speaker.  The cabinet should be as close to a net 0 as it at can be, no coloration or resonances, these just add to the complexity of properly implementing the drivers.  I am not as familiar with which OEMs make the best drivers for these functions but given a no compromises approach what drivers would you guys (Speaker Designers) select for your dream speakers, cost being no object, but given the constraints of size, volume, and given a bass extension target of 60hz (i.e. assuming your customer will implement your speaker with a well blended subwoofer.)

Sorry for the longwinded post but 8 pages got me thinking.   

[Ric Schultz]

Please check out Coincident audio and read all their lit on their speaker designs.  Very enlightening.  They use custom high efficiency Accuton drivers with just one part in series with the signal (custom 10 gauge OFC litz coil or Mundorf silver/gold cap) and time aligned and in their top of the line jobbie the woofer/tweet box is isolated from the woofs below.  They use braced double refined MDF (Rangerboard) for the speaker boxes.  I am sure they will sell you the top "mini monitor" from their all out speaker separately and you could mount this on top of Dannys servo woofs.....Wow!  That would be good!  94 db sensivity, so you can use those SET or Pass labs class A jobbies.  Check out high-endaudio.com for that guys evaluation of the speakers and how he stacked them for even better perfomance....and also biamped them.

[nickd]

The Accuton drivers always amaze me with their transparency and speed but. IMOP, they lack tonal color and "weight" (for lack of a better word) of treated paper drivers. The "perfect" driver (if one exists) has the qualities of both. The Scanspeak revelator series is good but they always seem to sound thin. Maybe a tilted power response

It would be nice to play with a pair of the ATD woofers that Magico used to use. last time i checked they cost about $300. ea. I guess thats why I never ordered them. That Ti. composite cone seems to be fast, well damped and has a very "natural" tone.

Anyone do a nice build with those yet?

[morganc]

Please check out Coincident audio and read all their lit on their speaker designs.  Very enlightening.  They use custom high efficiency Accuton drivers with just one part in series with the signal (custom 10 gauge OFC litz coil or Mundorf silver/gold cap) and time aligned and in their top of the line jobbie the woofer/tweet box is isolated from the woofs below.  They use braced double refined MDF (Rangerboard) for the speaker boxes.  I am sure they will sell you the top "mini monitor" from their all out speaker separately and you could mount this on top of Dannys servo woofs.....Wow!  That would be good!  94 db sensivity, so you can use those SET or Pass labs class A jobbies.  Check out high-endaudio.com for that guys evaluation of the speakers and how he stacked them for even better perfomance....and also biamped them.

What they don't tell you is that Mr "high-end audio.com" is none other than best friends with the owner of Coincident. I don't mean to derail the subject, just wanted to point out what some may consider a very relevant conflict of interest. 

[Vapor Audio]

Ric, I don't mean to pick on you but it seems like you have a tendency to hone in on these what I call theoreticals.  Things like time alignment, 1st order networks, etc.  All those things really have nothing to do with the ultimate success of a design, ALL that matter is the sum total of execution. 

I've heard fantastic speakers that boast things like 'transient perfect' ... and I've heard terrible ones boasting the same buzzword.  Coincident IMO is a perfect example of how focusing on a theoretical like 1st order networks can kill a design, because again IMO all their stuff I've heard wasn't enjoyable at all.  And speaking specifically about the pure ref extreme, that Accuton driver is not suitable for 1st order slopes.  I've used it, have a new design coming out with it, and know it pretty well.  A 1st order network apparently was singled out as a design goal, and this mindset overruled proper evaluation of the drivers best application.  Again, that's all my opinion ... but I see that time and time again.  Designers let these theoreticals dictate the design, which predictably leads to subpar results. 

If something like transient perfect is a design goal, you can set it up front.  But you must know how to choose components that will work inside the parameters set.  For example, transient perfect 1st order designs need a tweeter with a very extended bottom end and a low fs, and a midrange that is very well behaved above it's passband and has a very low inductance.  Those things are harder to find that you might think, and end up again making your choices for you.  To have a midwoofer with a very well behaved top end usually means a highly damped cone, and a tweeter with very extended bottom end usually means more moving mass.  So to fit your predetermined goal of transient perfect, you're ruling out many drivers that may well perform better than the ones that work with your theoretical ... and in the process giving up performance.  Or even worse like Coincident did, you're ignoring those facts and using drivers not meant for the theoretical goal anyway.

Now back to the topic of the day, drivers.  Specifically the Accuton stuff.  I have a new 3-way design that I knew what I wanted, the resolution and refinement of the Cirrus and the dynamic jump and excitement of my Arcus.  That's hard to combine, and few speakers I've ever heard do.  But in the process of working with midrange drivers, and drivers in general, I've found that two very important things for choosing drivers that do have that excitement is high motor force and low moving mass.  Also an extremely low inductance motor.  The Neo Accutons do that, they have ridiculously high Bl and low moving mass,  also a dead flat inductance curve.  The AudioTech SDKM is in the same boat, although with a hair less motor force.  But in evaluating midrange drivers for the Joule I tried the C90, and found it to be a bit lifeless and non-engaging.  Very accurate, just not the fun factor I was wanting.  Also with the ferrite motor Accuton C173's, same thing albeit a bit more life than the C90.  With the C173-6-90, neo motor, fabric surround, dedicated midrange ... wow!  That thing has crazy jump to it, lifelike dynamic jump and incredible transparency. 

Once I had chosen my midrange I let it tell me how it wants to be used, not the other way around.  I didn't try to stuff it into some network topology I had decided upon before-hand, I tried a lot of combinations until I found the magic.  You have to let the drivers tell you how they want to be used. 

Another observation I shared with Danny on the phone the other day is that drivers like the AudioTech stuff, because they're each made by hand by the man himself, just have tighter tolerances. Everything fits in alignment better than high end Scan or Seas stuff, and the AudioTech stuff is dead on every time.  There's never a unit that sways from the baseline, like there is with anything made by even partial automation.  The result of that is AudioTech can use a more compliant suspension, spider and surround, because things are in better alignment and they don't need the soft parts to keep things from banging around.  That more compliant suspension results in a driver that gets moving easier.  The AT woofers always sound full even at whisper volumes, ScanSpeak woofers do not.  Those small improvements are what I look for when I choose drivers, unless I have a price point to meet. 

And although there's no suspension in the RAAL, you get the same level of attention.  IMO the RAAL is simply the finest tweeter I've ever heard, and it's not a close comparison.  I had (and still have) shelves full of tweeters that didn't make the cut for my designs, and now that the RAAL is my paradigm I can't find anything that makes me consider for a second using anything else.  I guess I should point out however that I used the Beyma TPL-150/H in my Arcus and love it also, crazy fun to listen too, and compared to any dome blows them away in resolution ... but it's still not as resolving and refined as the RAAL.  The Beyma was used because it does something the RAAL doesn't, cross over at 1000hz ... and 120db 

There, still made piles of sawdust today AND had time to reply!

[Ric Schultz]

What does time alignment have to do with theories?  All you have to do is move the tweeter up and back in relationship with the mids and you will hear and measure a difference.  Same with 6db per octave xovers.  I don't favor any sort of xover.....I favor transparency.  I don't have a theory.  I just know that all parts distort the sound.  Purity is the goal.  So, I use the absolute fewest parts I can to do the job.  Too little parts and you get distortion and overlap.  To many parts and you get veiling and lack of dynamics.  Looks to me as if the Accuton midrange you mention could be used with just a coil.  Very little peaking at the top.  The only way to know would be to try it with just a coil and then add a cap and see which one measures and most of all sounds best.  In the end the sound is all that matters.  The Coincident speaker you slam (maybe you have not heard their best) was raved about in Absolute sound and the very knowledgable guy at high end audio seems to think it is world class.

[jtwrace]

I wanted, the resolution and refinement of the Cirrus and the dynamic jump and excitement of my Arcus.  That's hard to combine, and few speakers I've ever heard do.

My Abbeys do just that!   

The Coincident speaker was raved about in Absolute sound and the very knowledgable guy at high end audio seems to think it is world class.

That's meaningless...seriously, because one person said it sounds good in a magazine which is for profit we should believe them?   

[corndog71]

I remember seeing a cartoon showing how time alignment of a tweeter and woofer could be mimiced by simply reclining in your chair the same amount.

[J Fallows]

High motor force and low moving mass, like a compression driver.

[Danny Richie]

I can't wait to add some juicy stuff to this discussion but will have to come back to it. I am not making saw dust though. I am packing up orders that are going India. It is a pallet load of product. I love all the orders but this one is going to take many days to pack.

What to look for in a driver or how to choose a driver, coming...

And I want to jump into the driver alignment and crossover order discussion...

[Russell Dawkins]

High motor force and low moving mass, like a compression driver.

I think effective radiating area comes into the equation, too, which is where larger cone drivers and horns make their appearance.

[Vapor Audio]

What does time alignment have to do with theories?  All you have to do is move the tweeter up and back in relationship with the mids and you will hear and measure a difference.  Same with 6db per octave xovers.  I don't favor any sort of xover.....I favor transparency.  I don't have a theory.  I just know that all parts distort the sound.  Purity is the goal.  So, I use the absolute fewest parts I can to do the job.  Too little parts and you get distortion and overlap.  To many parts and you get veiling and lack of dynamics.  Looks to me as if the Accuton midrange you mention could be used with just a coil.  Very little peaking at the top.  The only way to know would be to try it with just a coil and then add a cap and see which one measures and most of all sounds best.  In the end the sound is all that matters.  The Coincident speaker you slam (maybe you have not heard their best) was raved about in Absolute sound and the very knowledgable guy at high end audio seems to think it is world class.

Let me try and explain a little better.  Before I do, I just want to make clear this is just sharing of ideas and experiences.  I hope we can continue to do so without anybody feeling stepped on. 

I don't think you're getting my gist of the term theoreticals.  Things like time alignment are just that, in theory an advantage.  However in reality it's only an advantage if exectued properly, and there are a lot of caveats as I alluded to before.  The reality is time alignment is in large part a myth.  With real world drivers the phase is directly associated with the frequency response, so only a non-existant driver with dead flat FR from 0 to infinity would have no phase swing.  Real world drivers have large phase swings, that's just the way it is.  So when you time align, what are you trying to accomplish?  All that matters is phase alignment, and with real drivers that phase alignment can only be accomplished in a window, typically around the crossover point.  Outside that window there's really nothing you can do about it.  With 6db crossovers you're greatly exasserbating the problem because the areas outside that window where phase isn't in alignment aren't attenuated as much, so the phase differences are more audible, and integration suffers. 

Accomplishing phase integration isn't a difficult thing to do with drivers that aren't "time aligned" using asymmetrical crossover slopes.  The good thing about that approach is you don't rule out a large pool of drivers for possible use as you must with 6db slopes, most any drivers can be integrated perfectly with asymmetrical slopes. 

The bottom line is you don't try and fit drivers into your idea, your theoretical goal.  You let the drivers tell you what they want, because if you try and tell them ... they ain't listening. 

What is time alignment anyway?  The acoustic center of woofers changes with frequency, same with tweeters.  And when you align, are you setting the tweeter back further to compensate for it's faster rise time?  And what if the listener has a chair that's at a different height than you designed for?  The term time alignment is often used and near always misunderstood, and in reality the common use of what it means doesn't exist. 

This idea of every component adds coloration is counter-productive as well.  Back to the Accuton example, you can't stuff a driver into a 6db network and just expect it to work. Drivers that do work well in 1st order slopes are few, more often the designer is determined to use it that way and ignores evidence to the contrary.  So you stick your tweeter in a 1st order network, and are determined to keep the network as "pure" as possible, so you leave out the impedence compensation network.  For many tracks it sounds ok, and you think you're good to go.  Then you hit a track with brass or a female vocal right on top of the tweeters resonance, and it's like somebody stuck an icepick in your brain.  What happened?  You ignored reality, and instead chose to adhere to meaningless thoeretical advantages. 

Do all components add distortion as you say?  Maybe, maybe not ... I don't personally think so.  What I do know is that you can't leave parts out that are needed and still get results.  Yes, you keep it as simple as possible, but no less. 

[Vapor Audio]

My Abbeys do just that!   

I have to disagree, having heard them on 3 separate occasions.  Actually I think once was a Nathan, not an Abbey.  Yes, they have the dynamics ... but not the resolution or refinement.  At least not at the level I personally want.  If there is a compression driver that does have that, I haven't heard it yet. 

[Vapor Audio]

Can't believe no new replies this morning