[TomW16]

Hi Dave,

I have been lurking here for years patiently waiting for the Ellis 3-Way to materialize and from what I understand there could be 2 different speakers in the works:

i) existing 1801b speaker retrofitted with a 10" woofer, which would be backwards compatible for existing 1801b customers should they choose to add the woofer.

ii) new 3-way speaker with OW1F tweeter, W18EX001 mid and potentially a TC Sounds 12" woofer.

There has not been too much action on these topics in the last little bit so I thought I would create this thread to help me (and others) understand the status of both speakers.  This would also help me determine the time frame that I have to save up the required money and purchase the kit.      

I hope that you are doing well and that your hectic life is manageable and fun.

Thanks.

Tom

[EProvenzano]

Inquiring minds want to know...
 

[TomW16]

Hi Dave,

In case my original e-mail was too subtle, I was hoping to obtain the status of both 3-way speakers from you.  I.e., time lines to kit launch, prices, etc.  I wanted to clarify just in case you were thinking that I was going to update the group with the status.  

Take care,

Tom

[David Ellis]

I have "valid" woofers for the 10" speaker, but it remains behind the following projects.

1.  Craig Dorval's 1801 speakers.

2.  A deck on my home.

3.  Wiring in the basement.

I haven't heard from TC sounds for about 45days regarding the 12" woofer, but during our last chat I told them not to rush.  I did some testing with an Accuton driver (C95).  The performance was surprisingly good.  The Accuton's better cone, better motor airflow, and tight tolerances might "trump" the SEAS shorting rings and phase plug.  Some testing with an Accuton midrange will follow if the Accuton C95 proves better than the W18.

However, I am not sure if this will work.  The Accuton might not be sensitive enough.  I figured out how to use Neo magnets in the W18EX to get another 1.5db sensitivity.  Another possibility is a custom driver from Accuton or SEAS.  This would be expensive.

After limited time with the Accuton C95... my CLIO measurement jig got a bit whacky.  I still haven't completely solved this.

I have learned of a reputable gent that refoams drivers http://layneaudio.hypermart.net/repair.htm .  Given this knowledge, the TC sounds driver might have a foam surround.  Everyone agrees that foam is better.  It is better damped and lighter than rubber, but it needs to be replaced every 10-20 years.

Dave

[David Ellis]

This would also help me determine the time frame that I have to save up the required money and purchase the kit.  

I really don't know which one will get done first.  I am more excited about the 12" woofer 3-way.  Dynamics performance IS directly related to sensitivity.  If the  TC woofer gets finished relatively soon, this project will happen first.

Dave

[TomW16]

Thanks for the status update Dave.  I am leaning towards the 12" woofer 3-way at this point (due to the perceived/potential better soncs) so if you finish that one first, all the better for me.  Take your time as I need to save up my pennies.

I'll be watching this thread to learn about the speakers' status periodically.

Take care,

Tom

[hubert]

Dave, you said:

everyones agrees that foam is better damped and lighter than rubber...but it needs to be replaced every 10-20 years.

IMHO, it exists some thin rubber surround that approachs the weight of a foam equivalent; Scan-Speak uses such stuff on their drivers, particularly on their "revelator" series. Furthermore  it could also be used some dampening varnish that doesn't dry with time; well known speakers in France "Apertura" (Christian Yvon) uses successfully such varnish on their custom "Davis" drivers.
In addition, I'm not sure that foam surround on such 12"woofer does better perform than rubber on frequencies less than 200-300hz. If we look at the drivers market, foam is used almost only on high sensitive midranges with high fs.
And...intellectually and subjectively, I would hesitate to buy a speaker with a driver surround that I know it will have to be replaced even in a few years. For my cheap car stereo yes, for a 2000 dollars speaker kit  pair not.
 

[David Ellis]

If we look at the drivers market, foam is used almost only on high sensitive midranges with high fs.

Why do you think this is?

IMHO, it exists some thin rubber surround that approachs the weight of a foam equivalent; Scan-Speak uses such stuff on their drivers, particularly on their "revelator" series.

There are huge variations among the rubber used in surrounds and their thickness.  Given the same weight, a foam surround will be better damped than rubber.

The better foams seem to last about 10 years.  Some report 20 years, and folks in humid/salty environments might only get 5-7 years from their surround.

I'll make a few statements of belief here.  These statements are based on input from very knowledgable industry folks.  I welcome anyone of any degree of authority on the matter to disagree.

1.  Foam is better damped than rubber.

2.  Foam is lighter than rubber given the same general application.  

3.  There are only 2 downsides to foam.
     a.  It will eventually have to be replaced (at @10 years)
     b.  It's less durable than rubber for abrasion

Summary,

The only valid reason to use a rubber surround is durability.  

When attepting to obtain the very highest level of sonic performance, foam surrounds are better.

There is one potential "hole" in these comments.  Some folks may WANT a heavier surround and subsequently a heavier cone.  This will indeed lower the Fs of the driver.  My response to this is that if more weight is desired, it should be added to the moving assembly it should be added to the cone.  A heavier cone can & should have greater strength/dampening.

And...intellectually and subjectively, I would hesitate to buy a speaker with a driver surround that I know it will have to be replaced even in a few years. For my cheap car stereo yes, for a 2000 dollars speaker kit pair not.

Yes, this is an upside-down issue.  You'd think that for $2k the speaker would also be more durable too.  Unfortunately, the best surround will require about $100 of maintenance every 10 years.  I am indeed concerned about this, but ONLY from a marketing perspective.  I'll use whatever sounds the best.

An analogy... Consider the 2000 BMW M5 built with nice lightweight forged pistons.  The have a very low silicon content and expand considerably in the cylinder when warm.  This is the downside to a really great set of forged pistons.  Their very light weight necessiates this construction method.  The result is these cars typically burn a quart of oil every 1000 miles.  Performance comes with a cost.  When the highest performance is desired, even the best materials have compromises.

I'm not sure that foam surround on such 12"woofer does better perform than rubber on frequencies less than 200-300hz.

This is VERY valid.  The other option is fabric.  Testing will be done.

The difference is that foam will weigh less and provide higher sensitivity.

Moving mass is very critical.  Saving 10grams of mass in a 12" woofer might net 1-2 db of sensitivity.  Commonly this is done with a thin paper cone, but for me this won't happen.  Such cones do flex.  This is not acceptable.

I DO apprecaite you input and commentary.  I AM "on the fence" regarding a foam surround.  I AM concerned about folks being able to obtain a quality replacement foam surround in 10 years (me included).    

As with last significant endeavor (the 1801), my intent is to build what I want.  If other folks want to buy it - okay.  It's all in God's will.  I feel blessed in this hobby.  

Personally, I often wonder if building speakers is the "right" thing for me to do with my spare time.  There are many thoughts, but these are primary...

1.  It's a very healthy hobby (and cheaper than golf)  

2.  God does have a plan  

Sincerely,

[hubert]

Dave,
Hubert said

when we look at the drivers market, foam is used almost on high sensitive midranges with high fs

IMO, the present "high end hifi" bass-midranges and woofers, at least from the wellknown european manufacturers, use at 90% natural or synthetic rubbers; foam is more often used on cheaper series of woofers or on real cones-midranges/uppermids(i.e. Audax PR series) or on a few tweeters(i.e. Focal inversed dome). But I admit not to know the entire drivers market.
I guess that a "standard designed surround"  builded with foam, rubber or whateverelse cannot damp the great energy at relative low frequencies transmitted by a woofers cone; IMO, to really damp the waves, once had to place a large amount of dampening material in woofers surround...but the moving mass would increase...; In fact, the real matter of surround is more on elastical proprietaries allowing to get a good phase relationship with the moving parts i.e. mecanical stiffness of the oscillant system, and other issues that I cannot translate in good english but that you know.
I am not a drivers engineer, I only know the physical laws of waves propagation and dampening and what I guess is only...what I guess in a theorical matter. If testings will proof the superiority of the foam surround on your woofer, I will understand the fact that you will use it in your 3 way.
 

[David Ellis]

I can see the market trends too.  Foam is visible on cheaper dirvers, but completely absent in hifi drive units currently produced.  The only place foam is visible is some higher end pro-sound units.  This is really quite strange, but I can totally understand the mindset.  Folks really do expect their $10000 speakers to last a lifetime without any maintenance.  

If a little foam makes things a little better, I'll use it.  I am not interested in building a suboptimal speaker.  I am interested in building the best speaker possible.  This will be expensive, and it might have long-term maintenance costs (i.e. foam replacement).  I fully realize my orientation in this matter might yield less than ideal marketing.  But, I am not interested in marketing.  I am interested in... the best set of compromises oriented toward sound quality.  

If (IF) I use foam, I suspect there will be a littany of questions regarding this choice.  In this regard, I have also continuously answered the following questions BEFORE folks purchase the 1801:

1.  Can a metal 7" driver really produce good midrange when used with a 3/4" tweeter.

2.  Can a metal 7" driver really produce enough bass for me?  

3.  Can a small 2-way speaker really push enough sound pressure for my room?

AFTER their 1801 project, nobody asks these questions     Very positive customer feedback is my reward for going slightly against the grain.

[hubert]

Dave,

OK, I understood your point of view about the surrounds problem and about the goal of your 3 way.

I am interested in building the best speaker possible

I learned that present permanent/standard manufactured magnets are worse than those made in past. They would lose partly their magnetic field (until 50%, even alnico!!!) relatively quickly (in a few years) and consequently the technical parameters of the driver are very affected.

May I know your thoughts/knowledges about this matter and how you will manage it in your future "best speaker possible"?

Thanks,
 

[David Ellis]

Neodymium has a very poor heat tolerance, but creates a very good and permanent magnetic field.

I haven't read "hear-say" about drivers loosing that much magnetic field, but do believe they lose some magnetic field.  I have some W18 drivers in my personal speakers made about 4 years ago.  They still measure the same.  Maybe there are different grades of ferrous magnets, but the companies I will use are reputable folks.   I believe they would use the good ferrous magnets.

I thought (?) Alnico was a very permanent magnet.

My summary on this matter is...

1.  There is a very solid technical basis for magnet degradation.

2.  I surely haven't measured any decline of my ferrous magnets.  Further, making the entire speaker from Neo magnet drivers is not possible at this time.  Each driver would have to be custom made.  The tweeter would be most expensive.  Oskar said a new design tweeter would cost @$30k+ depending on what I wanted.  SEAS makes a Neo midrange, but the sensitivity is too low.  The woofer cost increase would be minor.

3.  Since I cannot make the entire speaker from Neo drivers, it makes more sense to make the entire speaker from Ferrous magnet drivers.  This way, the slow decline in magnet power will be more uniform.  

Dave

[hubert]

Hello Dave,

some testings with an Accuton midrange will follow if the C95 prooves better than the W18... not sure it will work, it may not be sensitive enough.

The killer   seems to be here:
http://www.clofis.nl/nl/thiel/C90-T6.pdf

However, it is very...expensive:
In Europe:
C95-T6: 259€
C90-T6: 459€

 

[David Ellis]

I probably should have extrapolated further. ..

I am not sure if 88.5db will be sensitive enough.  Many of the Accuton drivers are about 88.5db sensivite.  The Accuton C79 is what I was thinking of for a midrange.

But... yes, the new Accuton would appear to be "the deal", but VERY expensive.

Dave

[HChi]

Dave,
I have been following closely with your 3-way speaker design.  I visited Dennis Murphy once to audition 1801 bookshelf, prototype 1801 MTM floor stand, and saw a somewhat trapzoid-like cabinet with a large woofer a couple years ago. I like 1801 much, but would love to have a full range floor stand for my next speaker upgrade.  After listening to 1801, I am obviously much interested in your next invention!

Regarding your 3-way, is there a upper cap you would like to keep?  What interests me the most are the diamond tweeters from Accuton and they seems to be quite popular among a few high-end speaker makers.  Forget about price point for a moment; have you had a chance to audition the diamond tweeters?

Being a Class A high current amp user, 88.5db seems reasonable to me.  I can see that you like make the new speaker be more sensitive (, likely 90+ db), but how about the impedence curve that you have in mind?  Some good speakers I have heard have a difficult 2 ohm load at some frequency ranges, which may be a difficult drive for many amps.  I am just wondering what your are trying to shoot for.  If a high sensitivity and easy 8 ohms are the design goal, perhaps low watt amps are of better choice for its 'unique' magic!  Just curious!!

[David Ellis]

Regarding your 3-way, is there a upper cap you would like to keep?

... Not really but it'll be something that I am willing to spend my own $$ for.  I heard the Accuton Diamond tweeter in 2002 ( I don't remember the speaker).  It sounded good, but... not that good.  I really didn't think it sounded better than the other great tweeters.  This is my opinion, and certainly others are welcome to disagree.  I would not spend $ on this tweeter for my own living room, and this tweeter is not a possbility from my workshop.  There are many other great tweeters.  The Accuton C12, C23, Dynaudio Esotar, Raven R1, Scanspeak Revelator come to mind immediately.  

but how about the impedence curve that you have in mind?

It will dip no lower than 6 ohms dcr and have @8 ohms nominal.  Another goal is to make the impedance VERY flat so the amplifier will have a less reflective load.  This can be accomplished in the crossover, and will probably cost @$150 per pair of speakers.

What you mention about impedance and sensitivity can be VERY deceiving.  This is because MOST speakers are rated based on volts (i.e. spl/2.83v) for their sensitivity ratings.  However, most amplifiers ratings are stated in watts.  This discussion can get very long, but I'll offer the following simple example:

... a small rant....

A 90db speaker at 2 ohms, a 87db speaker at 4 ohms and a 84db speaker at 8 ohms will all draw the same current from an amplifier.  This current draw from the amplifer (ANY AMPLIFIER) is the source of heat and thermal compression within that amplifier.  Some folks are quick to recite hear-say concerning tube amplfiers and their problems with current load, but the same is true with solid state amplifiers.

Some reading of the AKSA site and Rod Elliots projects, easily reveal that higher power supply rail votage can be used in these amplifiers but only if the load will be 8 ohms nominal (6+ ohms dcr minimum).  Using the higher voltage in their @60watt amplifiers will allow these amplifiers to push @100 watts.  The sad truth is these amplifiers are DESIGNED for a 4 ohm minimal load and higher current capacity.  This limits their output to 60 watts.  Obtaining 100 watts from those same general components in the same general configuration would require the designer to approximately double the amplifier component count.  

I believe it's common for folks to desire a very minimal speaker crossover, but sadly uncommon for folks to desire a very minimal amplifier design.  Sure, there is a good crowd of SET folks, but these folks seem quite isolated from the masses in hifi.

My belief is, "The most effective amplifier for a given speaker is an amplifier designed for the impedance load of that speaker".

If a high sensitivity and easy 8 ohms are the design goal, perhaps low watt amps are of better choice for its 'unique' magic! Just curious!!

Given my statement above, this should be fairly clear.  I am aiming for 90-93db at a very flat 8 ohms nominal impedance with a low dcr of 6 ohms.  The dcr will likely land around 6.2 ohms.

The speaker in Dennis's living room does indeed sound very good, but the impedance is @4 ohms due to the crossover.  Given a flat baffle and the desire for correct phase, this low impdeance appears unavoidable.  However, with a slanted baffle (backward leaning) there is mechanical phase adjustment.  The electrical focus of the crossover can be oriented toward a higher impedance.

If a high sensitivity and easy 8 ohms are the design goal, perhaps low watt amps are of better choice for its 'unique' magic! Just curious!!

Yes... and a couple thoughts....

1.  One of the reasons for my orientation towards higher sensitivity is an experience I had a few years ago.  I think this is mentioned elsewhere, but I'll mention it again.

I listened to a Tannoy Churchill on a single 300b tube amp.  It sounded good - very good.  There was something present in this setup that I never heard previously.  Some distortion was distinctly absent.  The dynamics were very good.  Sure, there was some cone flex.  Sure, the highs weren't perfect.  Sure, the bass wasn't powerful.  But, the midrange dynamics were very nice.  I attribute this to very low current (and low thermal compression) in the amplifer and in the speaker.  

2.  Given the X-max potential of the drivers I'll use, this speaker could really pack a wallop with @100+ wpc amp.  However, it'll most likely sound the best with a lower power SET amp @15watts.  Maybe a dual 300b amp?  I'll certainly provide some input but believe my customers will provide the subjective dissertation on this matter.

Oh, I should also add that the limitation of sensitivty in this speaker (and any speaker) is the size of the cabinet.  This speaker will be BIG, but will not grow larger than 5 cubic feet.

Dave

[brj]

Hi Dave.

Your (very informative) post alluded to several things that I've been wondering about, and I'm hoping that you can provide some additional clarification...

Another goal is to make the impedance VERY flat so the amplifier will have a less reflective load.

Did you mean reflective, or reactive?  (Or are they interchangable?)

Ok, first a few statements to clarify my own mind...

1) Impedance is a combination of resistive (real component) and reactive (imaginary component) contributions

2) The reactance of a component varies with frequency

3) Resistive loads are easier for an amplifier to drive

4) Signal reflections occur when two components have mis-matched impedances (with an acoustical result generally described as "smearing" of the sound)

5) Variations in a speaker's impedance curve mean that mis-matches with the amp will almost certainly occur at particular frequencies - even if the amp is well matched to the nominal impedance of the speaker


Assuming that I have all of that correct, what I was trying to isolate from your statement above was whether a flat impedence curve necessarily means the speaker is primarily a resistive load and therefore easier to drive and a "bumpy" impedance curve means that a speaker is more reactive in nature and therefore harder to drive?  I realize that impedance peaks at particular frequencies indicate regions of increased reactance, but I wasn't sure what you could assume about the overall ratio of resistive to reactive.  In otherwords, could the impedence curve be perfectly flat, but still dominated by the reactive contribution?


You made several other comments that seemed to connect impedance, current draw, and a speaker's abilities to handle dynamics, but I don't think I quite connected all of the dots - can you clarify?  Is it as simple as reduced current demand = less thermal compression = better dynamics?  If so, I still need the connection to impedance...

Thanks for any pointers!

[AKSA]

Some reading of the AKSA site and Rod Elliots projects, easily reveal that higher power supply rail votage can be used in these amplifiers but only if the load will be 8 ohms nominal (6+ ohms dcr minimum). Using the higher voltage in their @60watt amplifiers will allow these amplifiers to push @100 watts. The sad truth is these amplifiers are DESIGNED for a 4 ohm minimal load and higher current capacity. This limits their output to 60 watts. Obtaining 100 watts from those same general components in the same general configuration would require the designer to approximately double the amplifier component count.

David,

Without a long and boring discussion of power amp design, this is indeed largely true.  The AKSA and Rod's amp were both designed for 4R loads, because this is the safest option commercially.  There are so many 4R speakers out there that you'd be risking everything if the amp performed only with 8R loads.  In that scenario, it would be risking instability or even outright destruction with 4R loads.

Designing for 100W should account also for 4R, in fact a little lower, perhaps 3R.  To avoid approaching the SOAR limits of the output devices, you'd need to double the output stage size to two pairs rather than one.  Larger speakers often use more drivers than small speakers, and this usually lowers their impedance still further.  I've been embarassed with one manufacturer deliberately connecting his large 2.5R speaker to one of my 55W AKSAs.  It didn't sound effortless, so he then demoed one of his own amps to the customer - who'd originally intended to build an AKSA - and won the sale.  This sort of dirty trick is common in high fi, in fact, life in general.

AKSAs have been built which are stable down to 1.62R loads.  They use four output pairs!!  But the call for such a powerful amplifier is very limited, and so I've taken the decision to offer only the 55W and the 100W versions.

Cheers,

Hugh

[David Ellis]

Thanks for the confirmation Hugh.

I've been embarassed with one manufacturer deliberately connecting his large 2.5R speaker to one of my 55W AKSAs.

Yes, this is a dirty trick.  Much of this is clouded for customers, and the "safe" bet for amplifier manufacturers is hedging towards amplifier construction that will accomodate lower impedance loads.  However, as you illustrate, this approach is sub-optimal.

Dave

[smithsonga]

David-

Will you be benchmarking (or probably already have) other DIY 3-way designs.  Rick Craig's designs come to mind.  Comments here?

thx
Jim