Beyond the Ariel

That sounds right, so I must be reading the Klippel measurements wrong.. :eek:


I think it's the break-down of distortion products that I'm having problems with.

Ex. Scanspeak 18wu/4741t00

My worry is that you cannot trust the price or the quality for what these drivers seem to measure. They look good. They may sound good. But so do a lot of much cheaper drivers. I prefer some body to have done a decent technical review, measuring basic Thiele and once I have CSD pulse and step test I will do the rest. And these days too many drivers especially the so called top line can be priced on dubious measurement not performance. I see $220 dome tweeters with 35 gm neo magnets and lightweight alloy construction. How does this keep cool, how does it cost so much? and it has stored energy problem.

Designers nearly always keep too much back on their designs or even retrongineer leave something out, to pull out of the bag later or they may <reinvent the wheel> every 5 or 10 years and call it alloys or something else. If you are a virtually complete DIYer you will pass all this hype stuff. Even if you grab a driver and redesign the pole pieces. There are some that do all this tweeking. Is that a dying art. Take a real cheap driver and mod it. The good thing is parts made in an emerging economy have a time when good designs come at a low price in a time window of opportunity.Grab it,
 
Unintended endorsements aside, does B&C have a good low bass driver? I am a huge fan of their midwoofers.
Depends on what you consider "low bass".

The B&C21SW152 does very well in low distortion linear output to 25 Hz in a ported box, but with a relatively high Fs of 32 Hz, there are many loudspeakers capable of flatter, deeper response.
Nothing with more undistorted output above 25 Hz comes to mind, but in terms of lesser output levels, LF extension typical for home theater/music use there are more cost effective choices.

Josh Ricci's website gives excellent un-biased comparisons of some of the best drivers and LF alignments available:

Data-Bass
 
On the subject of carbon cones and parameters "shaping", here is an interesting tease from void audio:
Void introduces new carbon models

Nice teaser and could be really a great thing, but they're taking some liberties: a lighter stiffer cone doesn't imply lower vent velocity in my world, unless they mean a higher efficiency driver in a bigger box with a bigger vent but that's a pretty big stretch.
 
Last edited:
Me too. As a result of recent Skype/telephone conversations with Alexander of RAAL and Gary Dahl, it has occurred to me that band-splitting is a better way of reducing IM distortion than simply doubling-up drivers. In other words, rather than two drivers covering F3 to 700 Hz, better to have a F3 to 100~200 Hz range, and a 100~200 Hz to 700 Hz range, with separate amplification and equalization for the deep LF range. Maybe head-smackingly obvious, but sometimes several people have to tell me the same thing before I get it.

The requirements for the deep LF and LF/MF drivers are not the same. The LF/MF driver, in this case the GPA/Altec 515 Alnico, has a measured efficiency of 98.6 dB/meter/watt, a Qts between 0.19 and 0.2, and modest Xmax. It will be powered with a 20-watt triode amplifier, and there's no point in using it below 100 Hz. A small LF/MF closed box, or closed-box combined with highpass filter (resulting in a synthesized 3rd or 4th-order highpass), will restrict excursion to the linear region.

The 100~200 Hz crossover is not especially critical; it needs to keep excursion down for the LF/MF driver, and the overlap region (if any) will depend very much on prevailing room modes.

If the DIY'er has enough space, a LF/MF bass horn is a great idea, since the crossover frequency ends up being the same.

Excellent ! - seems like a good plan .
I've been down this route to some extent already, so like 3GGG I have a tapped horn ( Ok only one, currently ) to do the low bass, from 30-35Hz for one octave ( then ramping down ) . I also have a conventional sealed-box sub if I want to try that .
Where I differ from you is the next section up, for the upper bass and lower mids. I'm surprised you're sticking with a 15" driver there, but I suppose it might be headroom expectations and/or 'already having the drivers' ! ?
I've grappled with the 700Hz crossover area from the Azura for a while now , trying 12" Supravox in Onken ( too coloured, too much overhang ) , the Fane 8" on 120Hz tractrix ( somewhat coloured and not very articulate at normal listening levels ) . The 'problem' is that the GPA288/Azura is very quick and tonally strong in its lower reaches below 1.5k .
Only recently have I found something that clicks for me : I did a quick test involving a Jensen 12" alnico guitar speaker ( bought cheap ) on a decent-sized 25mm MDF baffle , and it blends really well - dynamically and tonally . Percussion, for instance, is excellent . Neither the baffle nor the driver are ideal, but these shortcomings can be fixed.
Next step is to get info ( from my friend James D ) on what I'll get in detail eg. at the low-end, from Tone Tubby 12" ( the 16 ohm one ) on a sensible-sized OB - he can run the calcs for different heights and baffles with shallow side-wings etc . These drivers ( thanks for the recommendation on your site ) look very promising and there already exists positive info on an OB design using them, on the Lowther America site, with a useful 70Hz approx roll-off at the low end. If it looks good I'll be ordering 20mm perspex for the baffles - which seems to have lower colouration than any wood I've tried for OBs - as proven in James's 'Quasar' design which has the best mids and upper bass of any speaker I've heard .

So anyway, Lynn, if you're definitely going with GPA 515 in that area, do please consider if OB is possible for that band as you should get a better blend into the Horn output . This recent test was definitely a clear class above anything else I've tried to do for the lower mids .
 
That Data-Bass link is great thanks for sharing that. How low should you go is an interesting thing to consider. Bruce Thigpin (all around nice gentleman) has his rotary subwoofer that probably gets about as low as anything, albeit not particularly cheap. Eminent Technology: home I also found the Precision Devices 21 inch subs from Humble HiFi pretty intriguing. If you read through his article he went from a no baffle to boxed version with some comments on what he found along the way. http://www.humblehomemadehifi.com/download/Humble Homemade Hifi_Serious Sub_copy.pdf
 
I'd probably go the motor-generator route before buying one of these ... at least a motor-generator has 100% isolation of transients, several seconds of "ride-through" due to rotating mass, and runs down slowly when the power fails completely. It's basically a little power station that's all your own, completely electrically isolated from the incoming AC power. They do make a whirring noise, which I why I would exile it to the garage in a small box where it could do its thing.

Hi Lynn,
I recommend the motor-generator
It is also used in large IT COLO facilities.
It is the only way to be sure of the sine wave.
Those UPSes which regenerate sine waves are very poor at getting the wave anywhere near a sine.
 
Member
Joined 2008
Paid Member
okay here's how I did mine (sub-bass)

I have been preaching this for years and years now. One has to be able to MOVE some air. My sub unit is just (one, mono) Infinite baffle utilizing 8 units of the Acoustic Elegance IB-15 back when John still produced them. They are indeed reaction-canceling.

Moving up the frequency scale, I have recently surmised that in order to keep up with the subs, I will also have to MOVE some air in the so-called mid-bass range. The plan is to use an Altec 515-8G, horn loaded, in addition to
probably a pair of something like 416-8B's for augmenting the lower spectrum of where a horn loaded 515 just won't produce, and I don't want to force it to. The entire mid-bass section is to be feed from the active crossover, 60-500Hz, using different amps for the 515's and the 416's. The section driving the 416's will be equalized. So, that wil be 3) 15's per side just for mid-bass.
Above 500, a horn loaded compression driver, of course, and a ribbon for the extreme highs as well. Okay, now call me crazy. Happy listening to all :)
 

Attachments

  • Crown K2  I.B. System 002.jpg
    Crown K2 I.B. System 002.jpg
    137.1 KB · Views: 591
  • total camera download 234.jpg
    total camera download 234.jpg
    105.2 KB · Views: 577
Lynn, I was going to ask you about using a combination of an 18" speaker for the extreme lows and a 15" above that for the higher frequencies. Now don't ask me to recommend an 18" speaker as I really don't know them intimately as to the fs values. The only reason I suggest the 18" is the cone area can move so much more air than a 15" with so little excursion. I do have a 24" speaker that I acquired in the past that I have thought about using as a subwoofer but there are no specifications and I have not run the Thiel Small values. From what I can see it looks like a copy of a JBL 18" driver with a larger frame and cone and the motor looks like a copy of a standard ceramic motor assembly.

Dr. Geddes can chime in here, but from what I've seen of 18" drivers, the very heavy cones perform badly above 100 Hz. Based on spec-reading alone, these devices appear optimized for DJ and theater use below 40 Hz, and are intended for use with extremely powerful (more than 500 watts) amplifiers. They seem very specialized to me, and are not general-purpose woofers at all. I have no idea which 18"-or-larger driver is appropriate for the purpose of audiophile LF (or VLF) reproduction. Servo feedback, maybe? It works best at the lowest frequencies.

I threw out the idea (borrowed from Alexander) of outward-facing and paired 15" drivers so reaction forces from the cones could be mechanically cancelled by rods that go through the enclosure and couple each of the mounting bolts to the complementary driver. Each 15" driver can be powered by a separate channel of a stereo power amplifier, and the L/R gains of the amplifier trimmed so the pair are an exact match (easily tested with pink-noise input, temporarily reversing phase of one driver, and looking for the deepest null in the 20~50 Hz region).

I'm looking for woofers that have excellent performance in the 20~200 Hz range, with emphasis on the lowest possible production of high-order harmonics. Unfortunately, Klippel doesn't reveal this, since it is a near-DC measurement, and I'm concerned about distortion above the driver resonance. Dr. Gedde's posts about high-order driver distortion are much appreciated. This information is not easy to find from driver manufacturers, who if they publish distortion information at all, only publish 2nd and 3rd-order curves, which are not that audible, particularly at low frequencies.

Back in the late Seventies, I tried the push-pull distortion-cancellation trick with one driver facing in and the other driver facing out, but the cancellation fades away when you get much above 100 Hz, so it's not as effective as it might seem.

This isn't surprising in light of the discussion above, since PP cancellation only cancels even-order terms, leaving the odd terms alone. It also relies on near-perfect phase and amplitude match; if these two parameters vary with frequency and level, then the resulting cancellation residue bounces up and down, leaving you worse off than with no cancellation at all. In other words, do PP right, or don't do it at all.

Reaction-force cancellation is simpler, and keeps the enclosure from bouncing around at low frequencies. There's no effect on driver distortion, but since we're mostly interested in high-order terms, that's up to driver design and manufacturing techniques.

P.S. Good taste in amps, ScottL. Crown is the way to go if you want quality watts. I was surprised just how good the Crown Macro Reference sounded on the Ariels.
 
Last edited:
Hi Lynn

I've never used an 18" so I can't comment.

I will say that I just don't see the need for massive speakers going very low. In a small room it's all about EQ and the Fs of the driver makes almost no difference. So as long as the speaker can handle the power and excursion required by the EQ its fine. In his regard I can recommend B&C because it is very hard to burn out one of their bigger speakers (12s and 15s). I prefer several small closed box 12" speakers spread around the room to one massive array. Admittedly it gets tough setting up a system like this (I have a computer program to do it), but I get 20 Hz, at whatever volume level I have ever wanted with just three 12" subs coupled with my pair of 15" Summas. Lots of EQ to do this, but these days with MiniDSP there is no excuse for not doing it. There is a powerpoint of my sub setup concepts on my web site. (Use the link on the "Loudspeakers" page as the main page link seems broken.)
 
Last edited:
Excellent ! - seems like a good plan .
I've been down this route to some extent already, so like 3GGG I have a tapped horn ( Ok only one, currently ) to do the low bass, from 30-35Hz for one octave ( then ramping down ) . I also have a conventional sealed-box sub if I want to try that .
Where I differ from you is the next section up, for the upper bass and lower mids. I'm surprised you're sticking with a 15" driver there, but I suppose it might be headroom expectations and/or 'already having the drivers' ! ?
I've grappled with the 700Hz crossover area from the Azura for a while now , trying 12" Supravox in Onken ( too coloured, too much overhang ) , the Fane 8" on 120Hz tractrix ( somewhat coloured and not very articulate at normal listening levels ) . The 'problem' is that the GPA288/Azura is very quick and tonally strong in its lower reaches below 1.5k .
Only recently have I found something that clicks for me : I did a quick test involving a Jensen 12" alnico guitar speaker ( bought cheap ) on a decent-sized 25mm MDF baffle , and it blends really well - dynamically and tonally . Percussion, for instance, is excellent . Neither the baffle nor the driver are ideal, but these shortcomings can be fixed.
Next step is to get info ( from my friend James D ) on what I'll get in detail eg. at the low-end, from Tone Tubby 12" ( the 16 ohm one ) on a sensible-sized OB - he can run the calcs for different heights and baffles with shallow side-wings etc . These drivers ( thanks for the recommendation on your site ) look very promising and there already exists positive info on an OB design using them, on the Lowther America site, with a useful 70Hz approx roll-off at the low end. If it looks good I'll be ordering 20mm perspex for the baffles - which seems to have lower colouration than any wood I've tried for OBs - as proven in James's 'Quasar' design which has the best mids and upper bass of any speaker I've heard .

So anyway, Lynn, if you're definitely going with GPA 515 in that area, do please consider if OB is possible for that band as you should get a better blend into the Horn output . This recent test was definitely a clear class above anything else I've tried to do for the lower mids .

i was also considering the tone tubby (or two ?)for the range from 70-700 but never got to trying them . i would be very interested to see what you find out .
here is a guy using them with a horn on top also , some measurements on that page too.

ToneTubby 12" AlNiCo Drivers On An Open Baffle Plus Beyma CP380M With AH!1000 Conical Horn
 
Dr. Geddes can chime in here, but from what I've seen of 18" drivers, the very heavy cones perform badly above 100 Hz. Based on spec-reading alone, these devices appear optimized for DJ and theater use below 40 Hz, and are intended for use with extremely powerful (more than 500 watts) amplifiers. They seem very specialized to me, and are not general-purpose woofers at all. I have no idea which 18"-or-larger driver is appropriate for the purpose of audiophile LF (or VLF) reproduction.
I'm looking for woofers that have excellent performance in the 20~200 Hz range, with emphasis on the lowest possible production of high-order harmonics. Unfortunately, Klippel doesn't reveal this, since it is a near-DC measurement, and I'm concerned about distortion above the driver resonance.
The B&C 18" are very capable drivers, and although can take well beyond 500 watts, by no means require that much power. Very smooth, and very precise.
The screen shot below is a B&C 18SW115-4 in an approximately 9 cubic foot Fb 28 box measured outdoors at two meters, no smoothing at all. Larger box would flatten the response to 28 Hz.

The Eminence Lab 12 has excellent performance in the 20-200 Hz range, very low distortion too.
Like Earl, I'd be inclined to use multiple smaller drivers rather than one big driver- more options are available.

If you would like to see the higher order distortion response of either of these drivers I can provide them.
The distortion response of the B&C 18SW115-4 previously posted demonstrating high orders of distortion was at around 1500 watts, at home level distortion is hard to see above the noise threshold.

Art
 

Attachments

  • B&C 18SW115-4.png
    B&C 18SW115-4.png
    173.5 KB · Views: 538
Last edited: