Beyond the Ariel

[augerpro]

Originally posted by ttan98



DE10 may not be a direct replacement, have not checked the spec., however I have used DE10 it is very good, also recommended by Magnetar. min. X-over freq. is 2.5Khz. About $38 each, yes $38.

If you want to go upmarket, go for DE250, it is excellent, I heard from users no equalization is reqd. min. x-over freq. is 1.5Khz. I think with a little tweaking you can cross as low as 1Khz. About $130 each.

PE has the DE250 on sale for $105 right now. Assuming a 1.5K Hz LR4 crossover, how do you guys think it compares to the BMS 4540ND or 4550, Beyma CP380/M, or the cheaper 18Sound driver? I'd most likely use the DDS Eng 1-90 waveguide with them.

BTW Lynn, what made you decide to use teh LeCleach profile over a Geddes waveguide?

[Lynn Olson]

Originally posted by augerpro

BTW Lynn, what made you decide to use the Le Cleac'h profile over a Geddes waveguide?

The Geddes waveguide is patented, and requires a special grade of foam - Dr. Geddes indicated over in his thread that not just any open-cell foam would do, and I take him at his word. The existence of the patent, and the requirement for specialized foam, means the only (legal) source of the waveguide is Dr. Geddes himself, and I expect he has better things to do than deal with a bunch of high-maintenance DIY'ers. I sincerely hope his commercial ventures are a success, so more people can hear his designs.

I remember reading the Summa has a complex crossover - 22 parts, or something like that. That's more complex than I'd like to do, and I don't want to go down the digital-EQ route for the main part of the audio spectrum.

There's also a big element of personal taste here - I've heard Le Cleac'h horns sound very good with minimalist systems, and that's encouraging in terms of keeping overall complexity down. I've designed complex-crossover systems before (my first commercial speaker designed in 1976 used 46 parts in the crossover), and am not enthusiastic about going back down that route.

The Ariel has a much simpler crossover, and I found chasing out cap-coloration a nuisance with that speaker. With more revealing high-efficiency drivers, I expect the parts-coloration problems to be worse, not better, so that's a strong incentive to avoid crossover complexity. I usually draw the line at one notch filter per driver, and try to avoid them if at all possible. As mentioned earlier, this is a personal-taste matter, and there are plenty of others who have no problem at all with complex crossovers (multiple notch filters etc).

P.S. Nice little find today - Al Klappenberger makes a nifty little transformer tweeter attenuator. These gizmos are way better sounding than L-Pads, since the tweeter sees a low source impedance instead of the series resistance of the L-Pad. By putting a TTA between the compression driver and the crossover, and putting an 8-ohm resistor across the primary of the TTA, it is possible to buffer out almost all of the wild impedance variations of the compression driver from the crossover network. Highly recommended for all horn speakers with passive crossovers - I certainly plan to use a transformer or autoformer for just this purpose.

P.P.S. Al's discussion on elliptical filters, or extreme-slope crossover networks, is well worth reading. Horn/waveguide drivers are in a special position to benefit from these types of networks, since the increase in excursion below horn cutoff is so extreme for compression drivers.

You really want the electrical power to be very greatly reduced at frequencies at driver resonance or below, since the horn is outside its passband and no longer providing any diaphragm loading. That's why professionals use 24 dB/octave crossovers, for example - not just to protect the drivers, but keep distortion within reasonable bounds. Yes, there is a price to be paid in terms of impulse response, but the reduction of IM distortion is a worthwhile tradeoff.

[badman]

Nice attenuator and only 1/4 the price of the equivalent Fostex. Pretty sweet!

[Salas]

Originally posted by Lynn Olson

P.P.S. Al's discussion on elliptical filters, or extreme-slope crossover networks, is well worth reading. Horn/waveguide drivers are in a special position to benefit from these types of networks, since the increase in excursion below horn cutoff is so extreme for compression drivers.

Are we talking the bottomless notch thing? I thing Joseph Audio uses similar stuff, called 'infinite slope'.

[Patrick Bateman]

Originally posted by Lynn Olson
[B]


My favorite of all the speakers at the show was the AudioKinesis - and in a very difficult room, too. I very much enjoyed listening to my Mercury "Picture at an Exhibition" disc at Row 15 playback levels - this is a CD that is nearly unplayable on most hifi systems due to the extreme slewing requirements on DAC converters, amplifiers, and speakers. The contrast between the bottom-dollar Eminence woofers in the Emerald Physics and the pair of Alnico 10" TADs in the AudioKinesis was nothing less than stunning - yes, drivers matter!

You can't add Quality by turning a knob on the equalizer!

Duke LeJune has done a really good job of combining audiophile finesse with engineering sensibility. Every speaker he's made has been an improvement on the last. I wonder what he'll be creating five years from now?

It would be excellent to see his gear sold in more markets.

Could you imagine how much he'd sell if his products were available in a boutique in Santa Monica, New York or Bellevue?

[caninus80]

Hi all.
Its superb project...

[JoshK]

Originally posted by Lynn Olson


The Geddes waveguide is patented, and requires a special grade of foam - Dr. Geddes indicated over in his thread that not just any open-cell foam would do, and I take him at his word. The existence of the patent, and the requirement for specialized foam, means the only (legal) source of the waveguide is Dr. Geddes himself, and I expect he has better things to do than deal with a bunch of high-maintenance DIY'ers. I sincerely hope his commercial ventures are a success, so more people can hear his designs.

I remember reading the Summa has a complex crossover - 22 parts, or something like that. That's more complex than I'd like to do, and I don't want to go down the digital-EQ route for the main part of the audio spectrum.

Just in response to the above and not to argue staunchly for the Geddes' profile, but I think there is a bit of misconception, at least the way I see it.

The Geddes' profile, from what I know isn't patented. The overall design might be but the profile is prior art according to what I've read what Geddes himself says. A few of us DIY'ers have built excel tools for graphing the Oblate Spheroid for different parameters and Geddes himself had a look at a tool I helped write.

In his book I beleive he argues that he OS profile has the minimum HOMs and so that is what he uses. The foam insert is to further reduce HOMs and could be used on the Le'Cleach just as easily as the OS. Or the corollary, the foam insert could be excluded from the OS just as it would with your Le'Cleach and theoretically still obtain better results.

The complexity of the crossover is, I think, a biproduct of Geddes optimizing the whole power response and not just the axial response. Conceivably you could approach the crossover with the OS profile just as you would the LC profile.

Basically, all I am saying is that I don't think your arguements rule out using the OS based on complexity, patents or foam plugs. I think the real issue is trying to find a manufacturer who can make a large enough OS waveguide to use as you would like to use the LC.

Josh

[Ed LaFontaine]

If we could narrow down the parameters of the desired LC or OS, maybe some intrepid soul would knock one off.

To my knowledge, Geddes' only reservation about the routine Josh refers to was that the throat entry shown was perpendicular to the axis. The driver Geddes used has an angle of 6 degrees. So, Geddes' point was that the entry angle of the throat needs to match the driver.

[Sheldon]

Originally posted by Lynn Olson


The Geddes waveguide is patented, and requires a special grade of foam - Dr. Geddes indicated over in his thread that not just any open-cell foam would do, and I take him at his word. The existence of the patent, and the requirement for specialized foam, means the only (legal) source of the waveguide is Dr. Geddes himself, and I expect he has better things to do than deal with a bunch of high-maintenance DIY'ers.

Maybe just a technicality, but as far as I can determine, the patent on the foam has not issued, but he does have an issued patent on at least one profile: http://patft.uspto.gov/netacgi/nph-P...geddes+earl%22

I say a technicality, as I don't suggest that people take advantage of Dr. Geddes ideas without compensation or permission. He sold me some foam. He might be willing to license individuals for a modest fee.

Originally posted by JoshK
The Geddes' profile, from what I know isn't patented. The overall design might be but the profile is prior art according to what I've read what Geddes himself says. A few of us DIY'ers have built excel tools for graphing the Oblate Spheroid for different parameters and Geddes himself had a look at a tool I helped write.

In his book I beleive he argues that he OS profile has the minimum HOMs and so that is what he uses. The foam insert is to further reduce HOMs and could be used on the Le'Cleach just as easily as the OS. Or the corollary, the foam insert could be excluded from the OS just as it would with your Le'Cleach and theoretically still obtain better results.

The complexity of the crossover is, I think, a biproduct of Geddes optimizing the whole power response and not just the axial response. Conceivably you could approach the crossover with the OS profile just as you would the LC profile.

Basically, all I am saying is that I don't think your arguements rule out using the OS based on complexity, patents or foam plugs. I think the real issue is trying to find a manufacturer who can make a large enough OS waveguide to use as you would like to use the LC.

Josh

I agree with Josh, no technical reason not to try the OS profile. I disagree in that it shouldn't be much harder than a Le'Cleach horn. The mouth size controls the lower frequency limit in either case, and that's the big part of the horn, therefore the hardest part to make. As far dealing with it being deeper, making it in two pieces, entry and mouth sections, wouldn't be difficult. The one issue with being deeper is the larger offset from throat to mouth.

Sheldon

[Jmmlc]

Hello,

I think that inside Earl Geddes's OS waveguide the foam ensures 2 functions:

1) reducing the residual HOMs
2) absorbing reflected waves resulting from a poorly terminated mouth.

For a Le Cléac'h horn (the name was first given by David Mc Bean when introducing that profile in Hornresp), with a mouth opening at more than 180 degrees there is very few reflected waves from the mouth to the throat and the second function seems irrelevant. About the first function, Lynn as others reported no horn colouration from both the OS waveguide and the Le Cléac'h horn so to use a foam infilling of the later horn seems seems not very useful, even it can be detrimental if the foam possess some variation of absorption coefficient with frequency (something that can be useful to linearize passively the OS waveguide frequency response which one is far from being constant without equalization).

Best regards from Paris, France

Jean-Michel Le Cléac'h

Originally posted by JoshK


In his book I beleive he argues that he OS profile has the minimum HOMs and so that is what he uses. The foam insert is to further reduce HOMs and could be used on the Le'Cleach just as easily as the OS. Or the corollary, the foam insert could be excluded from the OS just as it would with your Le'Cleach and theoretically still obtain better results.
Josh