Beyond the Ariel

I think that 1/4" in anything but a high frequency device is pushing it some as far as time alignment goes. You are going to have more problem with the vertical axis that is much greater distance than the front to rear differential at 1/4" which is about a 54Khz wavelength difference. There is no way that anyone who is using a distributed horn system is coming even close to this ideal, it isn't happening. I can find so much wrong with most horn implementations it isn't funny in the least. I want to see any coherent horn system used in a home that has a long bass horn, if you want to call them that in the fist place. Most horns in a home are going to have a cutoff much higher than 80hz in reality. 1/4 mouth size gets real big below 75hz and that is already going to be large. Paul klipsch did know what he was talking about on a theoretical basis but never once built a device that followed his own theories, his folded horns were all conic sections attached together through rough folds. I appreciated his theoretical papers and read most of them, but his implementation always left much to be desired.
 
There's a reason that frequency response graphs of the Klipschorn are so hard to come by, even though PWK measured them at the factory. The bass horn is nothing but up-and-down above 300 Hz, and is really looking zany by 500~600 Hz, where it almost-but-not-quite meets the mid horn.

The prototype thus far is very different than that. The AH425 & Radian 745Neo are very flat through their working range, have a smooth rolloff below 550 Hz, and enter the up-and-down region above 12 kHz. The GPA 416-Alnico is in piston range from 60 Hz through 1 kHz, and has a very broad first mode at 1.5 kHz, with very smooth rolloff above that. It's kind of the anti-Klipschorn in terms of frequency and impulse response, measuring more like a minimonitor.

The subjective impression was kind of surprising: somewhat like a really big electrostat, but with much more vivid tone color. About as transparent - and revealing - as Stax electrostat headphones, but with more physical bass. I imagine JMMLC's systems in France sound similar; very transparent, very dynamic, and vivid. Almost no horn coloration at all; every time I heard it sneaking in, it turned out to be a small misalignment or fault in the source equipment.

The crossover and time alignment turn out to be quite a bit more sensitive than lower-efficiency direct-radiators. My guess is that lower IM distortion, combined with greater headroom, reveals the setup (and source) faults more readily.

I've been thinking some more about the relative merits of the single-15, twin-15, and horn-loading. Most of us think of efficiency as the index for what kind of amplifier we can use, but that's not entirely true. I suspect the moving-mass of the woofer has more to do with whether the amp's power supply will be overwhelmed.

Case in point: Very few SET amplifiers can cope with the back-EMF's of a 15" driver; the big current impulses overwhelm the dinky supply usually found in SETs, and the bass gets flabby quickly. This is an area where Class A PP amplifiers sound better, and can shrug off the back-EMF currents from big, massive drivers.

Horn drivers are loaded by air, not cone mass, and do not present the same kind of back-EMF to the amplifier. Strictly a guess, but if you're using a 45 or 2A3 SET to cover the entire frequency range, horn bass is probably a better choice. 845 SETs are another matter, but building a good one is very, very difficult.

My amplifier is pretty unusual: a nominal 20~30 watt Class A PP amplifier with a 200-watt (or more) power supply, so it clamps on to the speaker and doesn't let go.

Hmmm ... well yes, I am designing this speaker for myself, but don't want to limit people to the really unusual amplifier I am using. It should still be usable with a 300B SET, for example, or any number of Class AB PP-pentode-with-feedback amplifiers.
 
Last edited:
The Altec is a bit more variable than the Radian, in terms of diaphragm centering. Connect the Radian, boom, you're done. It'll meet spec, and they sound very consistent from unit to unit. The 745Neo is a big step up from the 745PB, which has a ceramic magnet and somewhat different internal design.

The pin-location setup for the Altec diaphragm means some will be a bit off-center, which causes a dip in response around 3 kHz and a rise in distortion. My understanding is that GPA will custom-center the diaphragm if you ask them to, since audiophiles are not going to be blowing diaphragms and needing to replace them in the field, as professionals do. The user can do this too: the horn is removed, the compression driver is placed face-down on a clean, dust-free surface, the back taken off, and the compression driver is fed a low-level 200 Hz signal while the diaphragm is centered on the three pins. Never done it myself, that's what I've been told.

Don't want to get you guys all excited, but I've heard a rumor that the Radian 745 is going to offer a beryllium diaphragm version at some point in the future. If that happens, first diaphragm breakup moves up an octave to 15 kHz or higher, response is even flatter, and distortion goes down.
 
Last edited:
Have to get new ones for mine anyway, so I had intended to ask them to do so. Price wise, it is a wash for me, so it is about performance. HAve you seen a dip in 8.5K range that some have spoken of? I may have the exact number wrong, but I believe that it may have been suggested that this was due to alignment.
 
Well, hard to say which to choose. The GPA or Altec Alnico version sounds like, well, Alnico, which is really good. Not a fan of ceramic magnets from any vendor. But the neodymium magnet of the Radian is also really good, and I think the Radian diaphragm is a bit better, thanks to the Mylar surround damping, which gets rid of the low-level fizz of the Altec/GPA tangential surround.

Both have the characteristic Big Sound of large-format compression drivers, while the AH425 is the lowest-coloration horn I've heard to date. The only realistic alternative to the AH425 would be the Altec 1005 or larger multicell horn; although not as smooth as the AH425, they sound mighty good, and may have better LF extension thanks to the larger horn-mouth. More LF extension means you can nudge the crossover down to 600 Hz.
 
Last edited:
It comes down to the magic of Alnico versus the hassle of hand-centering the diaphragm ... but the neutrality and see-through of the Radian 745Neo are pretty special, too. Both are in a completely different league than any small-format driver.

One of these days I'd like to hear the large-format, long-discontinued Yamaha drivers that Martin is using on the AH425.
 
Last edited:
Brush Wellman is already making replacement beryllium diaphragms for some of the Radian drivers so you may want to look into that. Even if Radian makes one it will be from a Brush Wellman made dome. In the past Radian drivers have always had some issues in the upper frequency range. the 1" drivers have had problems around 16khz for many years. A friend of mine use to be one of the engineers at Radian some time ago. He is also telling me that some of the newer Radians have improved significantly over past models.
 
All of the NEO are rather of the newer designs. My friend was telling me about one of the drivers that has an extremely large diameter ceramic magnet system he seems to like. The problem with Radian is there has been a high turnover of designers and each one wants to change things around so there has been little consistency from year to year. The owner of the company stays the same but behind the scenes things keep changing.

ps. Some of the very early Radian drivers use very poor steel in the motor assembly and this lowered the flux density and caused many problems. It was not low carbon steel that was used. Things have just cropped up over the years that have just cause nothing but problems with a basically good design.
 
Last edited:
myhrrhleine,
When checking the acoustic output of a speaker in the freefield, how are you making a distinction between phase and time alignment? Both are a measure of time, phase just given in an angular measurement, but if the phase is lining up the time alignment must also. We are talking acoustical phase angle, not just the electrical angle here.

I cannot say for sure about acoustic phase. I know electrical is noticeable. My misunderstanding. Time alignment is very noticeable and was easily perceived when I tried mating FR driver to a woofer.
 
buzzforb,
The difference in the acoustic centers of a 1/4" are going to be much greater at high frequency than they will be at low frequency. So the match between let's say a mid-range and a bass speaker with a small shift will not cause the same time misalignment as you will have between the mid and a high frequency device.
 
Agreed. My understanding is that ideally you want to be within 1/4 wavelength of the crossover frequency, which naturally becomes more difficult as frequency increases and wavelength decreases. In playing with FR and woofer, I crossed in 150-200 range, so it was non issue. Thankfully the FAr was in seperate box because I had to move it back in vertical plane or the disparity was really obvious. We are talking 1/2" of movement backwards.
 
At the 2012 RMAF show, I spoke with the product manager of Materion (formerly Brush Wellman) about their 3" diaphragm for the Radian 745-series compression drivers. They are in development, but no firm date yet.

I've heard from one of the application engineers in Radian that the Neo version of the 745 is quite different than the still-in-production ceramic-magnet version, and took two years to develop. I can't verify that (I don't work for Radian), but the efficiency has gone up a rather surprising amount, and it's not a measurement artifact, either ... it really is in 113~115 dB/meter range.

The response of the Radian 745Neo the AH425 (unlike the published Radian measurements), is the flattest I've ever seen on any horn/compression driver combination, falling within +/- 1dB over the working range, and impulse response within 0.4 to 0.5 mSec, which is pretty remarkable for a compression driver. My best guess is the Radians are going into high-end studio monitors.

http://radianaudio.com//components/com_virtuemart/shop_image/prod_files/745NEO.pdf
 
Last edited:
Having listened to FR speakers a lot, I am very sensitive to both phase and time alignment. It is one of my biggest worries with this project.

Both the Ariel and the new loudspeaker are more "coherent" sounding than full-rangers with whizzer cones (Lowther, Zu, etc.), which have pretty noticeable group-delay problems in the 1~5 kHz region. Full-rangers with single paper cones (no whizzer) have about the same degree of coherency as the Ariel and the new loudspeaker.

My speakers have always had less than 5 to 10 degrees of phase divergence between drivers in the crossover region, which gives an impression of a single driver, even when you're right next to the loudspeaker. The acoustical crossover is also free of peaks and nulls, so there are no sudden phase deviations in the crossover region to steer the polar pattern up and down. (Wilson speakers are notorious for phase deviations in the crossover region ... I can usually hear it across the room whenever I go to hifi shows.)

The closest "coherency" equivalent would be an old-school 12" or 15" Tannoy with the paper cone; tonal vividness is about the same, but the sound is more modern in character, with a bit of Altec flavor in the bass. It does not sound like a JBL at all.

The parts in the crossover do matter, however. Solen caps are strongly discouraged, along with most flat-pack inductors. The Jupiter caps are my personal favorite, but many people have good luck with Janszen and Mundorf (neither are cheap, unfortunately). Ohmite and Mills wirewound resistors are recommended; avoid the cheap sand-cast variety.