Beyond the Ariel

[swak]

Variac, I think you are rignt, he measured the 10" driver....

[catapult]

Speaking of coaxes, has anyone tried the new 8" 8CN252 from BMS? It looks smoother through the XO region than many coaxes. The impedance curve indicates a low-inductance (low distortion) midrange motor and no obvious cone breakups in the passband

http://www.bmspro.info/photos/bmspro_info/8CN252.pdf

It appears to be based on the new 8N215 midrange. At first glance, the distortion for that driver looks pretty high, up to 5% 2nd harmonic in spots between 300 and 3K, but that's actually pretty good for an 8" playing at over 115dB into 4pi space. Most 'hifi' drivers would puke or simply melt at that kind of SPL.

http://www.bmspro.info/photos/bmspro_info/8N215.pdf

[Lynn Olson]

Originally posted by swak
PaulW measurements are very helpful. Seems the published 18Sound curves have no smoothing, which is a good thing to pre-select drivers reading at them.

The 8" driver Lynn mentioned is really nice: if it gets the same lift on the lower end from the baffle Pawl W used one might be able to come with no crossover (but the LP filter at 150-200). It has a 10 Tm BL and 14.9gm mms. The 6" I mentioned before (6ND410) has a higher BL (11.6) and a much lower moving mass (8.2 gms), but the radiating surface is a bit more than half of that from the 8 inch driver, and Xmax is less than half (2mm p-p). A good thing is that both drivers have a moderate price, while the 10" Paul measured costs twice more.

Considering the gain from a similar baffle as PaulW´s, the 6"er should be roughly flat to 500 Hz (due to its rising response) while the bigger 8"would be flat down to 300Hz.

Wouldn`t the 8" incher still have a small rise above 1Khz up to the crossover point?

Lynn, how important for you is the BL/ MMS ratio?

General trends are easy to correct with a minor crossover adjustment, like coming in earlier with the inductor, shifting the value of RC Zobel compensator, or doing both together. Much of the fine-tuning of the Ariel was just this sort of adjustment, which alters the phase at the crossover as much as anything.

Hmm, good question about the BL/Mms ratio, which tells us the acceleration factor of the cone. Radiating area is important too, though. A Lowther/AER is blazing fast but really does need a front horn to come into its own - I've heard the same driver with and without a front horn, and trust me, it sounds much better with front-horn loading.

The radiating area on those things is just too small to give any sense of scale or dynamics with large-scale music. You can tell it's working too hard, even in the midrange. There's a sense of stress and slight hardness with small midrange drivers that just isn't there with larger midranges. It sounds to me like a power amp with an inadequate driver stage - when the music swings loud for a moment, the sound hardens, tone colors lose their realism and body, and natural quality is lost - just for a moment, then it returns. After a while you start to cringe just a little when the sound gets louder. Compression and blurring I can live with, but hardening - uh uh, don't like it.

That's why I keep nudging the discussion towards larger drivers, despite all the worries about polar patterns and this or that crossover objection. These can be addressed and resolved. Sound that hardens and loses its natural quality is a big deal, whether it's momentary or all the time. That's why I was so persistent about asking about compression drivers - brief periods of harshness is a big deal with these things.

The Ariel is pretty forgiving of loud transients - it mostly just gets all blurry, and some of the polypropylene "tupperware" plastic sound creeps in, just for a little while, then the sound opens up again. But the Ariel is obviously compressed to a horn enthusiast, even at 95 dB. This is plain old voice-coil heating, nothing I can do about it. The voice-coils in the little Vifas are pretty tiny, after all.

On the other side of the coin, what I'm trying to avoid is the crashing-into-a-brick-wall sound of classical big horn systems when they get too loud. This to me sounds like a dynamic expansion, not compression, but it is accompanied by a very rapid increase in harshness that exaggerates it even more. The old-school Altec A7 is the example here. This is the horn sound I'm trying to avoid.

When a traditional compression driver hits the wall, it sounds pretty awful - I suspect supersonic shock waves in the very small geometry of the phase plug, because the onset is very sudden and the sound is almost like wideband hash. I'm hoping the 2" exit drivers are better behaved in this respect.

Returning to the point made earlier, we have to consider the area of the voice coil as well as BL/Mms and overall cone area. The bigger the voice-coil area, the more resistant it will be to dynamic compression from voice-coil heating, which is a matter of simple physics of the temperature coefficient of copper. Since it takes a few seconds for air to carry away the heat (to the magnet), there's also a time-constant involved as well, slurring the dynamics like a mistimed compressor.

[Lynn Olson]

Originally posted by catapult
Speaking of coaxes, has anyone tried the new 8" 8CN252 from BMS? It looks smoother through the XO region than many coaxes. The impedance curve indicates a low-inductance (low distortion) midrange motor and no obvious cone breakups in the passband

http://www.bmspro.info/photos/bmspro_info/8CN252.pdf

It appears to be based on the new 8N215 midrange. At first glance, the distortion for that driver looks pretty high, up to 5% 2nd harmonic in spots between 300 and 3K, but that's actually pretty good for an 8" playing at over 115dB into 4pi space. Most 'hifi' drivers would puke or simply melt at that kind of SPL.

http://www.bmspro.info/photos/bmspro_info/8N215.pdf

Both look very good, no question about it. BMS and 18Sound put a lot of technology into their drivers, and it shows.

The coax has that weird dip in the 1~2 kHz region that almost all coaxes seem to have - this has to be an interaction between what's going on in the center of the driver and the cone. It almost looks like an anti-phase plug - an absence where there should be a phase plug or damping pad to smooth the transition from piston-band radiation to a narrower beamwidth.

The BMS 8N215 does look good, all it needs is a notch filter tuned to 4 kHz (the ripple in the impedance curve is a hint it is a pretty severe peak that shouldn't be ignored). The ripples below 1 kHz can be ignored, though - they are caused by the small test box, not the driver. Everything below 1 kHz should be the inherently flat piston-band region for a driver that small (unless there is a gross defect in the spider design).

[rj45]

The radiating area on those things is just too small to give any sense of scale or dynamics with large-scale music. You can tell it's working too hard, even in the midrange. There's a sense of stress and slight hardness with small midrange drivers that just isn't there with larger midranges. I

Long time lurker, first post on this thread.

Thanks for being a stickler on this, Lynn. I worked in a good recording studio at one time, and I really miss how the big studio monitors could pump out lots of very clean sound.

Listening to a metal TM right now, and man, there's just a bit of grit on the transients when I pump it.

Q: shouldn't a standard HD or IMD test find this stuff at some arbitrary power level?

-Don

[Lynn Olson]

Originally posted by rj45

Q: shouldn't a standard HD or IMD test find this stuff at some arbitrary power level?

-Don

You'd have to be a little sneaky and make the IM test quick: measure for no more than 50 milliseconds and get the hell out quickly, allowing the speaker to return to rest and cool off. This is much closer to playback with music that is only moderately compressed or not compressed at all.

I like the multitone tests. The big fault of all drivers - regardless of technology - is the massive presence of IM distortion. This is due to the constant-acceleration characteristic of direct-radiators, where the excursion increases at a rate of 12 dB/octave or an astounding 40 dB/decade! Think about that for a bit.

The implication is that lower frequencies are always intermodulating with higher ones, simply because the higher frequencies hardly move the cone at all. Worse, sounds that are below the passband of the crossover are still causing trouble with IM distortion, even though they are themselves inaudible thanks to the acoustic rolloff.

If I recall right, horns are constant-velocity, not constant-acceleration. This means that excursion increases at a rate of 6 dB/octave, or 20 dB/decade. Below cutoff, though, excursion goes way up, since the horn-loading effectively disappears. So you still need crossovers that work, to protect the CD from the region below cutoff.

Another consequence for the crossover is that if you want the drivers to be in-phase with each others, a direct-radiator and a horn need to have crossovers that have a 90-degree phase relation to each other. For example, with direct-radiators, you match an acoustic 4th-order with an acoustic 4th-order (the Ariels work this way).

When combining a horns with a direct-radiator, you match a 3rd-order with a 4th-order, or a 2nd-order with a 3rd-order crossover. Note these are acoustic crossover slopes, not the electrical slopes. The acoustic slope is the vector sum of the electrical and driver-response slopes.

The business of horns operating as constant-velocity devices is dredged from the furthest recesses of the memory banks, so I could be completely wrong. I certainly was about the 18Sound driver test, when I mentioned the 12" driver when the test clearly referred to the 10" driver. Oh well, keep on truckin'.

[jeff mai]

Originally posted by Lynn Olson
That's why I was so persistent about asking about compression drivers - brief periods of harshness is a big deal with these things.

I'm hoping the 2" exit drivers are better behaved in this respect.

Hi Lynn,

I wonder if a part of the problem is the fact that so many people are tolerant of gross frequency response issues. I've read many times about horn enthusiasts trying this or that EQ and then claiming the EQ destroyed some of the qualities they liked in the horns.

The so called "Newman Criteria" applies to some degree to all but the smallest horns even if they aren't constant directivity:

http://www.studio-systems.com/audiof...lifford/98.htm

Anyone ignoring this is living with a peak or bump at around 2-5kHz. Some people like it, but this is a major source of harshness when things get loud.

It's amazing what a compression driver / horn system can sound like with proper attention to this. I'd wager that the large format Altec system you heard with multicells was using a Hiraga crossover which uses a broad notch filter to address this issue. See your copy of SP #11 for details. I'm using a nearly identical notch filter on my large 204 Hz Azurahorns.

The other issue is the driver / horn combo itself. Once you choose a good horn geometry you'll find that some drivers sound much better than others on the same horn. Unfortunately it comes down to the "suck it and see" method unless you can borrow a unit.

[Lynn Olson]

Originally posted by jeff mai

It's amazing what a compression driver / horn system can sound like with proper attention to this. I'd wager that the large format Altec system you heard with multicells was using a Hiraga crossover which uses a broad notch filter to address this issue. See your copy of SP #11 for details. I'm using a nearly identical notch filter on my large 204 Hz Azurahorns.

The other issue is the driver / horn combo itself. Once you choose a good horn geometry you'll find that some drivers sound much better than others on the same horn. Unfortunately it comes down to the "suck it and see" method unless you can borrow a unit.

Hmm, what compression driver are you using with the AH-204? I presume a 2" Altec.

The Azurahorn I've been listening to in the local Denver area (it's not mine) is made for the 5" format Lowther, and has impressively low horn coloration, an encouraging sign for the smaller-format Azurahorns.

[jeff mai]

Originally posted by Lynn Olson
Hmm, what compression driver are you using with the AH-204? I presume a 2" Altec.

The Azurahorn I've been listening to in the local Denver area (it's not mine) is made for the 5" format Lowther, and has impressively low horn coloration, an encouraging sign for the smaller-format Azurahorns.

I'm using a Nippon Gakki (Yamaha) compression driver that is no longer made and difficult to find - though Bear was selling them on the Joe-list a while back. Martin from Azurahorn uses these drivers too, although I think he prefers 1" drivers on his smaller horns.

They are the same size as the large Altecs which are 1.4" throats with an approximately 3" diaphragm. They are very good drivers, but they don't really have enough HF extension for me to recommend them. My pair takes a steep dive at 12kHz in addition to the normal 6db/octave rolloff above 5kHz. The Altecs are not quite as smooth from memory, but have more extension.

If you liked the Altecs, I think a pair of 288C drivers would be a good starting point, but there is less choice in the 1.4" throat format. There are more 2" format drivers around to try. Keep in mind I once saw a compelling argument that a 4" diaphragm was too big from the point of view of break-up modes and many of the 2" formats use these.

[swak]

There are very nice things about the 6ND610 I keep thinking about. First the 102dB sensitivity (on baffle on the use-range, ie. down to 500Hz). While the 8" driver suffers a little less from compression at a certain power, the smaller driver is about 7dB (!) more sensitive. Second, the BL/Mms ratio is much higher on the smaller driver. Thirdly the voice coil diameter of the smaller driver is only 5mm smaller than that of the bigger one. And last but not least, the smaller driver goes up to 5Khz, while the bigger one only up to 2-2.5khz (as per datasheet recommendation), which 1) allows for a ribbon on its best range 2) puts the crossover point further away from the ears most sensitive range.

For the ease of large-scale music reproduction, taking 2 enABLed 6" drivers per side could be a *very* nice solution: much less compression as a single 8", 15% more radiating surface as a single 8", increased sensibility (105dB!), higher crossover point, higher BL/Mms ratio, simple crossover. I would avoid MTM anyway, so TMM.

The only downside I see is the possibility of combing effects and radiation anomalities. The high sensibility would narrow the tweeter selection for people using one amplifier for the mids and highs. Pricewise it would only mean 100 EUR more (total).