|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.29.109.1
I keep thinking about this and wanted to see what others felt.Back in the 50's, 60's and 70's the dominant design philosophy for woofers and mids (tweeters, too; probably) was to use a cone material with high stiffness to mass ratio and good internal damping so that the driver rolled off smoothly above its pass band. The list of approaches is very long: Bozak with their rubber coated metal cones mids and tweeters and tapered paper and wool fiber woofers, Bextrene, the Leak Sandwich, Hartley's cone with the porous center and rigid surfaces, AR, KLH, Advent, EPI, JBL's Aqua-Plas coatings, etc. In turn, these cone types allowed little or no low pass filtering and fairly simple high pass filters of first or second order. I think it's safe to say that nearly all of us have enjoyed the benefits of this type of design, and many of us still do.
In the modern era, however, things changed. Cones are light and very rigid, using carbon fiber, metals (aluminum, magnesium, and even beryllium),Kevlar, glass fiber, ceramics and diamond. The cones are light and rigid, but are also largely undamped with severe breakup modes and resonances above the passband. It is not at all uncommon to see peaks 10 dB above the nominal output level in these breakup modes. While these cones are very smooth in their useful range, the peaks require steep filters (3rd or 4th order) and often additional notch filters to reduce the peaks to inaudibility.
The trade off in these two approaches is a more ideal cone material in the pass band on the one hand, and a much more complex set of crossover filters on the other. And while the more ideal cone material may offer some advantages in following the input, the energy storage and phase shift of the crossover would seem to negate it.
I'll offer an observation. If the crossover filter is steep enough, it may be that the input signal is attenuated enough so that it does not excite the resonances to an audible level. However, if a tone is input at a lower frequency and the speaker has some non-linearities, the speaker itself will generate harmonics of the input that can excite the resonances. If the input is fairly pure, it will not mask the resonances and the resonances will amplify the harmonics. This does not happen with a damped cone material. Note that these harmonics are generated after the crossover, so are not filtered by it.
Just some food for thought.
Follow Ups:
for us simple minds
a low distortion 10 inch driver puts out cleaner
bass than a hard working 6 in ported drivercarefull overlap of the multiple drivers with a simple
x-over yields better speaker performance dollar-for-dollar
(why Henry K usually wnet with 2 way designs)newer cone materials do not necessarily give better results
-sometime "different" result - not necessarily betterthe essence of the Vintage Asylum is that we recognize
"good sound and good design" = when we hear it age- not
withstandingI used to likethe older ADC 303 and 404 speakers- they were simple and sounded good - for their size and price / but not fancy design
I used to lunch with Peter Globa the head speaker engineer of
HH Scott- some of his designs still sound good to me
using off the shelf drivers from Rola-Jensen, CTS , and Utah
There seems to be a debate over which of the extremes you describe are better and most of today's DIY community seem to favor the low distortion drivers that need taming outside of the passband in addition to the main filter components in the crossover. Thus the need for good measurements and simulations to get the filter and drivers married correctly to make the total system. I have no problem with that if that's what one wants.For me, there's no debate about what I want. I'm well known for using low order series filters and a prerequisite for them to work is that the woofer have no breakup outside of the passband and the tweeter must be able to handle the shallow rolloff of the filter. Period. I like to keep my filtering as simple as possible so I use tamer woofers and more robust tweeters, even with the higher distortion numbers that go with them. I have no doubt that most of the distortions that are measured are likely not even audible in real life, with most drivers, in most systems, and in most listening rooms. Others will argue ANY distortion, no matter how measureable, is a no-no.
I would much rather have a "less than absolutley perfect" driver that sounds good with minimal filtering and "taming" than one I've got to hit nine ways to Sunday to get the nasties out. Plus, I'm not hung up on the Holy Grail of perfectly flat response and the absolute lowest distortion numbers, obviously. I just want to listen to music and not speakers.
There's a lot of naval gazing going on in that part of the hobby these days, IMO.
On the opposite end of the spectrum are the single driver people. Although some use a helper super tweeter and even a subwoofer the general theory is keep the crossover away from the frequncies where the ear is most sensitive.I noticed that Peerless now makes a fairly widrange 8" woofer model 830884 with shorting rings that looks quite good on paper (execpt for the 89db@1w/1m efficencey which is just OK). I've read a post on the speaker asylum that it can work with 6db crossovers with the right tweeter (buy you'd probably want a Zobel). What do you think? (See link below.)
Dave
It has cone breakup at 3 K, so with 6 dB electric XO should be crossed at 1 K or below. For 2 way, the right tweeter should go down to about 250 Hz.I believe 6.5" is the largest woofer that can be used in 2 way with low order XO. They generally have cone breakup at 4.5-5 K.
The link to the posts I read is below. It sems the user did have a problem with the 3K breakup and others suggested some solutions.Some of the inmates on the High Efficency Asulum use a notch filters but that adds complication and I never heard an impimented for this.
There are smaller versions but of course, you lose some efficency. I could always go to a higher sloped filter. It was the shorting rings that appealed to me.
Dave
- http://db.audioasylum.com/cgi/m.mpl?forum=speakers&n=230700&highlight=Peerless&r=&session= (Open in New Window)
Ken,Driver distortion is as much about the motor design as it is the cone. Using a 6.5" driver at high levels requires a very linear suspension and magnetic structure to compensate for the extra stroke it must have. A good 10" will have a lot less stroke at the same level, and lower distortion from the motor as a result.
Interesting. I think that most so-called improvements in speakers are simply the result of people trying to find something new, anything at all, in order to differentiate their speakers from what has gone before because of the need to sell new ones. Speakers are a very mature technology afterall and speakers are VERY simple machines.It's like car starters. What's new since the last one I changed? Anything? They still spin, right?
where materials technology has changed drastically over the years. The only way Dayton-Wright was able to run a 12 kV bias was to seal the panels in SF2 gas. They also experienced various failures due to the relatively primitive coating technology of the 70s. Sound Labs, for example, runs their panels from 10 kV to 14 kV using modern polymer technology to make an ultra thin one micron diaphragm with a conductive coating that won't break down under those conditions. Their current panels are burned in at the factory at 800 watts. Try that with a pair of ESL 57s!While the magnetic drivers in your hybrid horns haven't changed much, some folks like Tom Danley have dealt with the coherence issues of using drivers with significantly different radiation patterns.
Back in the 60s and 70s, the popular stereo magazines regularly evaluated the bass output on a woofer on the basis of how low it would play the fundamental w/o eccessive "doubling", that is, producing a strong harmonic of the fundamental exactly one octave higher than the fundamental. Some speakers that simply doubled the bass notes below a certain frequency were actually pretty good, nearly convincing some listeners that they were hearing the music in all the octaves. THey used to say things like "THe lowest octave was mostly doubling, but this was not obvious without direct A-B comparison with a speaker capable of reproducing the lowest octave fundamentals."What's interesting is that I have almost never in recent years heard a reference to doubling by the DIY crowd. Given that most of these people seem happy with woofers 6 1/2 inches in diameter and smaller you would have thought that this would be a major topic of discussion for these guys along with phase coherence, baffle step compensation, driver interference etc. Could it be that modern woofers using modern materials are indeed less prone to doubling than an old 10- or 12-inch paper-cone woofer still is? Or is doubling still with us.
The Large Advent's (And the AR3's too)real claim to fame, of course, is that it could reproduce the lowest 1/3 octave (coming from the pedal tones of a pipe organ, for example) with minimal doubling.
On an only slightly different topic, I finally figured out what was wrong with that keyboard/pa amplifier speaker at church. Aside from the horn treble dispersion problems, basically its a big 12-inch woofer unhappily stuck in a vented box that is way too small giving a prominent mid-bass rise. Whenver the keyboardist hits a low piano note, it doubles an octave up, giving a warm but wooly, and decidedly artificial non-piano sound. The ear interprets the doubling as decidedly not sounding correct, and interestingly practically everyone, not just trained ears are sensitiveOf course, pianos themselves double, and a key reason to move from a 5 ft baby grand to a 9-foot concert grand is that the 9-footer will produce a stronger fundamental bass note with less doubling. In the case of the electronic keyboard, what started as decent sampled sounds ends up sounding arificial and "cheesy" aka wooly.
David,Your observations bring up a couple of points. There was a very interesting post by Ken Kantor on the MadBoard regarding Acoustic Suspension speakers, and by comparison, ported ones. That was in response to my son-in-law's asking why there are so few AS drivers around. I thought Ken's response very informative. He commented that it is much more difficult and expensive to build As woofers, because of their low resonance and floppy suspensions. He also commented that foam suspensions don't suffer from audible resonances like rubber does. And in fact, you do see a little bump and dip, indicative of a resonance, in the response plots on most rubber surround models. Usually up around 1 kHz. Ken also commented that since ported systems make use of a higher Q resonance to extend the woofer response, and once the forcing fuction is removed, they will ring at that resonant frequency. Ported systems do store energy.
A lot of the doubling isn't objectionable because of a psychoacoustic effect known as "synthetic bass" If you play a bass note with the harmonics in tact, but the fundamental removed, it is perceived as lower than a note one octave higher, even though the lowest reproduced tone in both cases is the same. (2nd harmonic of the first tone is one octave above the fundamental) The ear/brain processing fills in the fundamental from the interval of the harmonics. So if a speaker can't reproduce 32 Hz, but can reproduce the harmonics (64, 96, 128, 160, 192, etc) the brain will believe it's there. However, synthetic bass and real bass do not sound the same if you compare them.
Another effect is the ear's preference for certain frequencies. In recent years, two very expensive speakers tested by Stereophile had bass resonances at about 48 Hz, with rather large peaks (8-10 dB)at 55 Hz. In both cases the reviewers commented on how deep and powerful the bass was. I sincerely doubt that the designers of those speakers didn't peak the bass on purpose. If you put in a similar peak with an EQ, you'll see the effect; it's quite nice. My JBL's have a resonant frequency of 48 Hz and the bass seems satisfying and deep, that is, until I compare them to the Advents which have resonances down around 40 Hz. (those are measured figures) Then it's clear the JBL's don't go all that low, nor with the same authority.
I very much suspect that the 6.5" drivers in the ported boxes benfit from both the effects I described. In other words, if they double a bit (or a lot) and can produce a lot of energy between 50 and 60 Hz, then they will be perceived as having good bass. I think part of it is the measuring technique used by the DIY folks. It's all computer based and done in-room with a time gated system, so that response below 200 Hz is not measured. Thus, they reply on listening for the bass response. You can do close miked measurements with sweep tones or pink noise, but their systems don't do that in most cases, so they would need additional equipment besides learning a new technique. Plus, ported systems are a lot harder to do for the close miking technique and the results much less reliable. Even John Atkinson seems to have problems getting good results in his Stereophile measurements.
If you add up those three factors, I can see why a lot of modern systems don't have the bass of the older systems. I think doubling is still very much with us, it just isn't being talked about. It might be the elephant in the living room that no one wants to talk about.
The Greatful Deads' American Beauty album, Truckin' track, has low Bass that is simply a repetative one-note beat from some speakers, but becomes a beautiful up & down the scale Bass melody if the speaker is capable.
I have the Mobil Fidelity verson of Ameican Beauty and I love it!
This post is made possible by the generous support of people like you and our sponsors: