Beyond the Ariel

Driver edge diffraction

Given that pro-sound drivers will be used:

Any thoughts on addressing diffraction from the drivers(s) gasket(s) -- While there are exceptions, most prosound woofers have a pretty chunky face gasket. Can enABL really address this? Flush-mounting still doesn't address what happens where the surround hits the gasket.

I was especially curious about this when coaxes were still in the picture. 'Still seems relevant with the WR playing waaay into the midrange.

If this has been discussed elsewhere, kindly redirect me to wherever!

Thanks!
 
salas said:
Is timbral discontinuity from amp to amp something that you factor in, Lynn?

Not in the usual audiophile mix-n-match way - "this sounds good, let's try it with this" - but from a more analytical perspective.

Any amp that's worth considering is going to be quite flat from at least 15 Hz to 30 kHz, at a minimum. So the coloration we hear is a function of amplifier distortion, nothing else. These distortions can have quite subtle causes and may measure at extremely low levels, or even evade current measurement techniques. That doesn't mean they're not there, though.

On a subjective basis, amplifier colorations affect different parts of the spectrum. This is to be expected - HF distortions are correlated to nonlinear current delivery into capacitive loads (which may themselves be nonlinear), Class AB switching transitions, and current-gain variations as complementary and paralleled devices switch on and off.

LF colorations correlate with power-supply artifacts and storage mechanisms, especially time constants that are similar to the beat-rate of music - LF feedback instability can enter here as well. In solid-state, gradual shifts in operating temperatures at the silicon die (within the package) will also shift beta (gain) and HF rolloff characteristics - these are called memory effects.

Rather than have all the amplifiers be the same, it makes sense to use them in the frequency ranges where they have ample headroom and low distortion. Tube amps of any type are hard-pressed to deliver 350 watts to a subwoofer at 20 Hz. Similarly, a 200-watt Class AB transistor amplifier isn't the ideal choice for a compression driver.

My criteria is "intrinsic linearity". With no feedback and no "helper" circuits, how linear is the circuit, at what power level, in which frequency range? Is the distortion low or high-order? If the intrinsic distortion is 10%, it doesn't matter if the amplifier is a SET amp being operated at much too low of a frequency or Class AB solid-state amp at 1 milliwatt and no feedback to conceal the AB transition.

It's easy to scale the power requirements of a widerange dipole. The tweeter doesn't care about 1/f transitions, and needs only a few watts - of the highest quality possible, since IM artifacts fall into the most audible frequency range.

From 1 kHz on down, we're starting to need more power - the center of the power spectrum is between 200 and 500 Hz, sometimes called the "melody region". Choice of power amp in this region is going to be highly subjective. Some people like the Class AB high-feedback MOSFET or bipolar-transistor sound. Some people are into Class D. Others like traditional PP pentode amps.

The subwoofers need to be designed as an integrated system, with dedicated amplifiers that have superb DC and LF performance. The Rythmik is probably the head of the class here, with integral servo feedback to compensate for dynamic driver nonlinearities.

I see no problem with 350 watts of Class AB power for the subs, 30~60 watts of PP KT88 power for the midbass, and 300B DHT power for the mids and HF ranges. As a matter of fact, that describes the power lineup of some of the best systems I've heard around here.

But there are some hard-core SPL lovers who use 300-watt prosound amps with Klipschorns or Altec Voice of the Theaters - many of the denizens of Audio Asylum are big fans of this approach. If 130 dB is really important to you, of course, you're not going to waste your time with dipoles, but go straight for an all-horn system with serious PA credentials.

There's also a middle ground of audiophiles who like high-end transistor amps. Amps like this don't do anything for me, with one or two exceptions - the LNPA150 and Gary Pimm's low-power MOSFET amplifiers. I can't think of any others. Plenty of folks like exotic transistor amps, and if that's your taste, go for it!

But keep in mind this speaker is going to be pretty efficient - 97 dB/metre/watt at the minimum, and more likely 100 dB/metre/watt. That's 20X as efficient as most audiophile speakers. So whichever amplifier you choose, it should sound good at milliwatt and microwatt levels.
 
OB speaker Q

Lynn ... Sorry you are laid up. Recover quickly.

Bad news for you, but is great that you have taken time to get into my favorite "Pro speaker OB's". I am a P.Audio coax fan. Will take weeks to read it all. More than greatly appreciated.
Low Q drivers are a hard one to sell to my buddies as they are stuck in Qts = 1.0 land. I love the soggy sound on some.
Would you be kind enough to give the good and bad points of adding inline resistors to gain Q when it is necessary to bring the F3 lower?
If already mentioned will you link to post?

Thanks, Zene (Vancouver, WA ... and I do make trips to Jamac/Audionics).
 
Hello,


P.S. I received an e-mail from Lowther America recently, and they mentioned that there's a version of the Lowther with no whizzer cone - an obvious candidate for a ribbon tweeter to take over at 8 kHz. For those allergic to horns, the Lowther+ribbon would certainly be an alternative in the 1 kHz-on-up range. Now THAT would be an offbeat speaker - a trio or quartet of ToneTubby guitar speakers for the bass, MB, and low-midrange, and a whizzerless Lowther above that.


I wonder if the lowther could go lower than 1Khz, say 500-300Hz or something, without having significant distorsions. I ask this because I have no experience but I've noticed for example Mr.'s Martin King OB project using a lowther that down. Probably I'm alergic to crossovers in the 300Hz-5000Hz region or something... this would explain why I prefer a "fullrange" version.

The problem is that I think the lowther has a small Xmax for going that low. Mr.'s King simulations though reveal a small displacement for the lowther around 100Hz (well under 1mm, but of course with a cross point of 100Hz, 4th order HP filter).

I guess that a better option would be the Supravox 285-2000 exc. widerange. This 12", 102dB (!), whizzerless driver with plenty Xmax, would blend well with alnico TT, low-midrange driver. Over 5-8Khz a ribbon tweeter could jump in, or a good alnico supertweeter.

Lowther Open Baffle

Supravox 285
 
Re: OB speaker Q

Zene Gillette said:
Lynn ... Sorry you are laid up. Recover quickly.

Bad news for you, but is great that you have taken time to get into my favorite "Pro speaker OB's". I am a P.Audio coax fan. Will take weeks to read it all. More than greatly appreciated.
Low Q drivers are a hard one to sell to my buddies as they are stuck in Qts = 1.0 land. I love the soggy sound on some.
Would you be kind enough to give the good and bad points of adding inline resistors to gain Q when it is necessary to bring the F3 lower?
If already mentioned will you link to post?

Thanks, Zene (Vancouver, WA ... and I do make trips to Jamac/Audionics).


Which PAudio's have you used? Impressions? I thought many models seemed quite nice for the price. In fact I'm using the WN8S as a mid in a PA dipole here: http://www.diyaudio.com/forums/showthread.php?s=&threadid=102859
 
Sorry about the delay, the motherboard (and/or power supply) on my Mac Dual 2.7GHz G5 failed last night. This was the one and only overclocked model Apple ever made, complete with liquid cooling. I was wondering when it was going to fail, considering the peak temperatures on the processors were 70~73 deg C, and the idle temps were above 40 deg C.

After seeing those temps, I bought 3 years of AppleCare, and the tech comes to the house and does an on-the-spot repair, free of charge. He's going to earn his fee from Apple when he has to disassemble the complex liquid-cooling system to get at the motherboard. Fortunately, it didn't spring a leak - the coolant is a bright-green goo, a startling (and ominous) sight to see leaking out of any computer. I'll be using my TiBook laptop for the next few days.

-------

Wish the impulse response and CSD of the Supravox was better - those aren't very good measurements. FR is nice and flat, though, and people really seem to like the sound of the field-coil Supravox and Fertin fullrange speakers. The few times I've heard them at the ETF and RMAF, they didn't leave much of an impression - I really need to audition them more seriously this year at the RMAF show.

Lowther, everyone's favorite love-em-or-hate-em speaker (like Decca cartridges) I do know, being friends with Tony Glynn back when I lived in Oregon in the Nineties. Lowthers keep amazing me - they measure so bad, yet they can sound so right, especially in a Big Fun rear-horn or Oris or Azurahorn front horn. Not what I'd call neutral or low-coloration, but really entertaining sound, with big thrills when the planets are all in alignment. It's the last part ...

As for excursion, they're kind of like paper-cone compression drivers, with huge magnets, and very light diaphragms. Horn-loading all the way, baby! So I'm kind of conflicted about the best way to use them - a 300 Hz crossover is obviously going to stop any kind of visible movement and keep the VC in the gap, so from an engineering viewpoint, it's the obvious thing to do.

But ... I've heard the same (both 6 and 8-inch versions) drivers in standard enclosures and horns, and the horn versions are just so much more dynamic, despite sometimes annoying colorations. The biggest difference was with the 6-inch Alnico driver, which sounds petite, very clear, and miniaturized on a standard enclosure. On a Azurahorn AH-204, they sound BIG and very impressive, nothing miniature about the sound at all, with Altec VOTT dynamics. The presentation is completely different, and I would say heavily favors horn loading.

This gets us back to headroom. Many many posts ago, I startled readers with a design target of 120 dB SPL. That doesn't mean I'm going to use a 100~200 watt amplifier; no, that's not happening except for the subwoofer. When I mentioned that figure (which is still a design target), that was headroom for the speaker, not what I was planning for the system as a whole.

My Karna amplifiers are good for 16 watts Class A1 at 0.1% distortion, and 32 watts Class A2 at 3% distortion - with very little high-order distortion components. But the 18Sound or ToneTubby 12-inch widerange drivers, along with the Radian compression driver, can handle far more. That's headroom. Even if you don't use it, it's there. You can hear that it's there - it's obvious, the speaker never sounds stressed or like it's even trying very hard. The immense peaks that come through make that obvious, like a big torquey V8 when you get on the freeway. Tap the accelerator, and you GO! The same for a speaker with 10 to 15 dB of reserve headroom - there's a sense of power in reserve.

I like headroom in electronics, too. John Atwood's Artemis phono preamp can put out 60 volts RMS at the output jacks. Think about that for a while. The driver stage of the Karna is a PP 45 DHT running in Class A with transformer coupling - it can cheerfully drive the grids of the 300B's 30 volts positive with no transition visible on the scope. That's headroom.

Moving on the diffraction/driver mounting question, and the choice of crossover. Well, I listened to the Summa's at the RMAF, and although I couldn't draw any conclusions about tonality thanks to the bottom-dollar Pioneer HT receiver and DVD player, it was obvious that system integration was outstanding, and smoothness and coherence of the off-axis measurements were borne out in listening. Earl Geddes chose to cross his 15" driver and 15" mouth horn around 1 kHz (one wavelength at 1 kHz) - and it sounded very well integrated to me.

So I don't think 1~1.2 kHz is a high crossover at all. I was very happy with the sound and tonality of the Bastani Apollo, and the Apollo 12" woofer free-runs with no crossover all the way up to 6 kHz - and I thought it sounded great, much better than the small-driver audiophile speakers in the other rooms.

As for mounting, my preference, as always, is for flush mounting of the front surface of the driver frame on the front baffle, so yes, you do have to make a precision-fit rebate cut with a router. It works very well for the Ariel with its 3.8 kHz crossover, and it should work even better at one-third the frequency.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Lynn Olson said:


Not in the usual audiophile mix-n-match way - "this sounds good, let's try it with this" - but from a more analytical perspective.


From 1 kHz on down, we're starting to need more power - the center of the power spectrum is between 200 and 500 Hz, sometimes called the "melody region". Choice of power amp in this region is going to be highly subjective. Some people like the Class AB high-feedback MOSFET or bipolar-transistor sound. Some people are into Class D. Others like traditional PP pentode amps.


You did factor in amp variations validly. Wanted to know your approach. Thanks Lynn.


In the School Of Sound Recording when I was 20 they told us that the 200-800Hz band is called the 'power range' and carries the body of music. Must be similar to the term 'melody region' it seems, only a bit wider.
 
Lynn,

Wait about three weeks and then get back to Lowther America on drivers. They will have the first EnABL'ed EX series drivers by then. I processed a DX pair and sent them off to Limono just last week. Easily the most vivid drivers I have ever experienced. Absolutely not a sound dropped, period.

Can't speak to dynamic range but the treated DX drivers had completely quit shouting when driven hard. Just got louder and seemed quite comfortable about it. I was not so comfortable, I was measuring 107 dB peaks and 100 dB avg SPL on my trusty RS junk. Resolution was just like you get from expensive, narrow band tractix horns and excellent compression drivers, but with tremendous depth.

I am quite serious when I say I have never had this much information in my listening room, especially with this kind of dynamics and persuasive color.

We will have to see what Jon thinks and whether or not he will loan them back out, once he gets his hands on these brand new EX models.

Bud
 
P.Audio

Auqerpro ...
Have limited listening with P.Audio BM12CX38 12" coax on temporary OB's. Thought it had promise. Only heard with 2 T-amps and was way underpowered.
Getting new/better panel ready for it now. That's why I'm here to get ideas. Hafler 120 amp for the cone and a Super T (coming soon for the tweeter) with Ashly active crossover to play with frequencies.
One or two per side (yet to get) P.Audio E15-200S woofers for OB bass with the coax 12's.

Also have P150/2226 (JBL clone) to use in all horn bass bin and a pair of E12-200S woofers I have no idea what I got them for.

Both systems ready by 2016 at my present rate.

Zene
 
Lynn,

Your design a few pages back looks to have a lot of merit. I just finished reading your latest clarisonus blog entry. I might mention that SL probably didn't report your cascade solution as it was in conflict with one of his constraints that the system be relatively small form factor. The drawing you show is clearly not that.

The trapezoidal shape has a lot of diffraction benefits, I just wished that there was a more aesthetically pleasing way to design such a speaker. I am in the brainstorm part of my OB speakers as well and that is where I am stuck; designing speakers that are meant for maximum SQ but that my wife will allow in the LR.
:xeye:
 
EnABL & Lowthers

Bud was sent a pair of PM6A's, not EX drivers. These will be at RMAF in the fall, in an OB. FYI these have silver voils and Ticonal magnets, making them a cheap (by Lowther standards) but very nice driver.

With regards to dynamics and compression, I will state that if placed in a TQWT or bass reflex the dynamics do get restricted, but that is not a problem on an OB if you have a crossover network. I use 180 hz. on an 18" x roughly 48" baffle. Compression is caused as the driver exceeds x-max (1mm) which is easy to do when used full range without a BLH design.

The 5" drivers (which Lynn calls 6") do have less dynamics than the 8" unless horn loaded. They are also less efficient (94 dB instead of 99dB). On the other hand they have more extended highs and are cleaner sounding. It will be interesting to compare the 8" EnABL drivers to the 5" units in this regard.

I am usually very free in loaning things out, and after RMAF will be happy to let others try these drivers. Before the show there will be a bit of a learning curve and work to make them sound there best, so unless you are coming to Chicago to hear them you will have to wait.

With regards to whizzerless units: the 8" drivers typically have a mechanical crossover to the whizzer at 2.5 kHz. When poorly controlled this was the cause of the Lowther shout, but recent design changes do a much better job of control. I do not know of any measurements of the whizzerless cone. My educated guess is that use of the RAAL ribbon down to 2.5 kHz would seem to fit this driver well. The PM6A is only 96 dB efficient, fitting well with the RAAL 95 dB rating. Past experience with ribbons indicates that they lack the same dynamics as Lowthers, but the RAAL is supposed to be a lot better than what I have had.

Having played with the BassZilla design OB for awhile the sound is nicely scalable, it sounds right at low volumes and high volumes. For reasons I do not understand this is more true with the Alnico drivers (A series) than the neodymium drivers (DX and EX series). The Eminence woofer mates pretty well, albeit not perfectly, with a slight lower mid bass heaviness. This is more likely due to my crossover (a variant of what Dick Olsher recommends due to my use of the PM6A rather than his DX4) than anything else.

BTW, a thoroughly enjoyable thread, with much interesting information. It is always a treat to look at the same problems from a different angle, and this thread has it in spades.
 
JoshK said:
Lynn,

Your design a few pages back looks to have a lot of merit. I just finished reading your latest clarisonus blog entry. I might mention that SL probably didn't report your cascade solution as it was in conflict with one of his constraints that the system be relatively small form factor. The drawing you show is clearly not that.

The trapezoidal shape has a lot of diffraction benefits, I just wished that there was a more aesthetically pleasing way to design such a speaker. I am in the brainstorm part of my OB speakers as well and that is where I am stuck; designing speakers that are meant for maximum SQ but that my wife will allow in the LR.
:xeye:

Josh and Jon, thanks for the timely and informative comments. Much appreciated.

On the size issue, one of the things that appeared on this thread (or another DIYAudio OB thread) was the discovery by the European researchers that the single-point model of driver radiation, as used by HF Olson and many others, was not in fact accurate. Driver size matters. Large, diffuse sources have substantially less diffraction than small, single-point sources, and moreover, go deeper in the bass than predicted by the outdated single-point models. (This is why the Edge simulator behaves differently for different-sized drivers.)

This is a very significant finding that is still overlooked by some of the OB theoreticians. The implication is that baffles need to have a large percentage of their surface covered by drivers - I also see a corollary that drivers need to be grouped together as closely as possible so they radiate as a unit. A vertical line may not be the optimum shape, at least once you get into the 1/f rolloff region.

That's why I'm "stacking up" the drivers that are closest to the floor, creating a group of 6 drivers (including the mirror images). You could go even further and have a closely-spaced group of 8 drivers (a quartet of physical drivers) - like a Marshall stack on stage, with the WR driver above it.

A straight-line vertical stack loses the close coupling with the floor reflection, and throws away a lot of potential 1/f gain. It has to be remembered that although the floor reflection is a nuisance from a measurement perspective, it is completely natural-sounding, something we expect to hear.

It's the absence of a floor reflection that sounds unnatural - and don't expect a strong floor reflection in modern multimiked recordings, where the mike is quite close to the instrument or singer. Remember, in nature, the only time you don't hear a 2~3 mSec ground reflection is when you're sitting high up in a tree or are right at the edge of a deep canyon.
 
Question for you horn guys

This is a question for all the horn enthusiasts perusing this thread:

With a good LeCleac'h profile 550 Hz horn - like, say the Azurahorn AH-550 - and a crossover around 1 to 1.4 kHz, what differences will I hear between 1", 1.4", and 2" exit compression drivers of similar high-quality design?

Looking at the Radian product line, for example, the drivers would be the 465, the 745, and the 950. The 465 has a 1.75" diaphragm, and a 1" exit. The 745 has a 3" diaphragm and a 1.4" exit. The 950 has a 4" diaphragm and a 2" exit. All of the drivers have a roughly 2:1 compression ratio, or maybe the ratio depends on the relative areas of the diaphragm and the exit.

If the horn size and crossover frequency are the same, how would the 3 drivers sound different? This is a serious question, and I'm curious about the answer. Yes, I know, the bigger drivers get into trouble above 10 kHz. I can see that from the curves. But what about the rest of the range, from 1 to 10 Khz?
 

CV

Member
Joined 2002
Paid Member
Hi Lynn,
This is somewhat tangential...
Tom Danley explained how all direct radiators produce a 90 degree acoustic phase shift (when wavelength > driver size), whereas proper horns (or very large ESLs) don't.

It struck me that it might be possible that LF units on a large baffle couple to each other in such a way that the "effective" driver area (from a radiation patterb perspective) is much larger than Sd x n. May well be that a large LF array has temporal characteristics nearer to horn bass as a result.

As for compression drivers, I'd think that the higher the x/o, the smaller the throat you'd want. Just my gut feeling from the consideration of having the throat diameter << a wavelength - I'd feel weird having a source lobing away right at the throat.

This may well be an oversimplification given the action of the phase plug, but still, one wouldn't stick a Lowther in a VOTT midbass horn.

From a practical perspective, myl experience is limited to drivers of 1.4 or 1.5", all of which need tweeters. I've run S2 and RCA field coils on the same 1.5" 340Hz horn (I commissioned that and the 550 from Martin) and they sound completely different.

Speaking of phase plugs, that's what governs the compression ratio - I assume you were using the term more loosely than that as shorthand for exit/dia area...

Best and hope your recovery is proceeding well,
cv
 
CV said:
Hi Lynn,
This is somewhat tangential...
Tom Danley explained how all direct radiators produce a 90 degree acoustic phase shift (when wavelength > driver size), whereas proper horns (or very large ESLs) don't.
cv

But when I do a near field measurement on the direct radiator, they are square wave-like. There is no so-called phase shift. The key is flat frequency response. Maybe I never done it right :)

Personally I think the strength of horn speakers is their dynamics (which is a result of low memory effect and low surround/spider hysteresis). A horn speaker has a back EMF much larger than direct radiator. That means the ratio of back EMF to the impedance is much higher too, which in turn means better cone control. It can be mathematically shown that the hysteresis in those horns speakers are lower.
 

CV

Member
Joined 2002
Paid Member
Brian,
Very interesting! What frequency did you use?

I wonder if nearfield vs farfield is an issue here as well?

As for surrounds/spiders, I've long felt that these do more harm than they are given the blame for. I suspect adherents of true compression driver bass are hearing the absence of driver hysteresis as you imply...

best,
cv
 
rythmikaudio said:


But when I do a near field measurement on the direct radiator, they are square wave-like. There is no so-called phase shift. The key is flat frequency response. Maybe I never done it right :)

Personally I think the strength of horn speakers is their dynamics (which is a result of low memory effect and low surround/spider hysteresis). A horn speaker has a back EMF much larger than direct radiator. That means the ratio of back EMF to the impedance is much higher too, which in turn means better cone control. It can be mathematically shown that the hysteresis in those horns speakers are lower.

Reading between the lines here, that would imply as the bass array grows larger - and has more intimate coupling with the floor reflection - efficiency goes up, the ratio of diaphragm mass to air-load becomes more favorable, and by inference, back-EMF's increase as well, gradually approaching the horn condition.

This is the opposite of a ribbon or magnetic-planar, with very low BL products, which produce the characteristic flat impedance curve. As efficiency of a bass array goes up, all those little ripples in the Z curve should become more prominent, if back-EMF's are increasing (relative to DC resistance of the voice coil).

P.S. Thanks for the thoughts about the sound of different compression drivers, which I guess comes down to "listen and find out for yourself". I am told the exit pipe of a CD sometimes has a certain rate of expansion, which may not be the same as the inlet to a horn, thus causing a discontinuity, depending on the degree of mismatch. This would cause some horns to work with some CD's, but not with others. Something I need to find out more about.
 
Inline resistors

Lynn ... You may have missed my question when your computer went nuts. May I ask again?

"Would you be kind enough to give the good and bad points of adding inline resistors to gain Q when it is seems necessary to bring the F3 lower?
If already mentioned will you link to post?"
Zene
 
Administrator
Joined 2004
Paid Member
Re: Question for you horn guys

Lynn Olson said:
With a good LeCleac'h profile 550 Hz horn - like, say the Azurahorn AH-550 - and a crossover around 1 to 1.4 kHz, what differences will I hear between 1", 1.4" and 2" exit compression drivers of similar high-quality design?

A good question that I hoped would get a reply. Maybe it still will. It may be hard to find someone who has actually done the tests.

The 1.4" and 2" drivers I've heard, Altec, Westrex, TAD, some JBL and Renkus Heinz, all sounded very smooth to me. Very clean. They tended to be crossed lower than 1400Hz, though.

I can't say the same for 1" drivers, but they may not have been of the same quality, so that would skew results. Would be interesting to know the differences.

Meyer sound uses a 4" diaphragm compression driver coupled to an 80° x 40° horn in their CQ1. A box I used for years. Sounds very nice. Don't know the exit diameter. And of course being Meyer, it's got heavy processing in front. I always wanred to drag a pair of CQ1 home for a "Hi-Fi Listen." Never did.