Thanks for the info TS!
The Altec drivers I've pulled 'frams out of were so tight on the pins I don't see how they could be shifted. Radians were not that tight. Haven't pulled any others.
When you say "bump the new diaphragm while feeding it signal and watching for distortion on a scope" what do you mean? I'd like to learn the technique. Distortion is hard to see on the scope, unless it's big. But there may be subtle clues to watch for, if you know what they are.
Impedance sweeps I can do. Where do the jaggies show up? Above Fs? At Fs?
The Altec drivers I've pulled 'frams out of were so tight on the pins I don't see how they could be shifted. Radians were not that tight. Haven't pulled any others.
When you say "bump the new diaphragm while feeding it signal and watching for distortion on a scope" what do you mean? I'd like to learn the technique. Distortion is hard to see on the scope, unless it's big. But there may be subtle clues to watch for, if you know what they are.
Impedance sweeps I can do. Where do the jaggies show up? Above Fs? At Fs?
I'm sceptical about TrueSounds method actually working on Altecs' large drivers / though I'll agree they all should be checked ( measured & readjusted if necessary ) .
- The top-plate of Altecs' 288 ( Alnico ) series is bolted down to the return pot . The top-plate of the ferrite version is glued in place . Neither should move.
- OTOH, the top-plate of the large JBL ( Alnico ) series are held in place by the shear force of the huge magnet . With these drivers ( 2480, 2482, 2440, 2441 ) ; malletting ( not that gently ) the top-plate sideways will adjust the gap geometry ( symmetry ) / perhaps "fixing" an ill-fitting diaphragm ( just like TrueSound suggests ) .
- Here's Altecs' pdf directions for adjusting the locating pins on their newer drivers ( 288G onwards ). This was Altecs' preferred method / one needs the shimming "spider" . I don't know if GPAs' 288H driver includes this feature ( for adjusting the locating pins ) .
- Here are 2 close-ups ( collages made from Brandons' plots ) of the two 288H(s) in question .
- I'm hoping that the difference in FR between these two GPA 288s is merely due to the first diaphragm needing realignment .
<> cheers
- The top-plate of Altecs' 288 ( Alnico ) series is bolted down to the return pot . The top-plate of the ferrite version is glued in place . Neither should move.
- OTOH, the top-plate of the large JBL ( Alnico ) series are held in place by the shear force of the huge magnet . With these drivers ( 2480, 2482, 2440, 2441 ) ; malletting ( not that gently ) the top-plate sideways will adjust the gap geometry ( symmetry ) / perhaps "fixing" an ill-fitting diaphragm ( just like TrueSound suggests ) .
- Here's Altecs' pdf directions for adjusting the locating pins on their newer drivers ( 288G onwards ). This was Altecs' preferred method / one needs the shimming "spider" . I don't know if GPAs' 288H driver includes this feature ( for adjusting the locating pins ) .
- Here are 2 close-ups ( collages made from Brandons' plots ) of the two 288H(s) in question .
- I'm hoping that the difference in FR between these two GPA 288s is merely due to the first diaphragm needing realignment .
<> cheers
Last edited:
LOl, the Altecs aren't worth even talking about any way if you are serious about great sound "earlk." Way over priced and mediocre. Call the current manufacture that copied this travesty if you are skeptical. Be sure to lube up good first
OK, you don't care for Altec or GPA. Which large-format compression drivers (in current production) meet your standards?
Are you in the JBL Everest camp (with beryllium diaphragms), the Classic Audio Reproduction rebuilt JBLs with field-coils, the Oswalds Mill RCA-based drivers, Meyersound studio monitors, another pro monitor, or - none of the above?
I'm genuinely curious about what you like in the high-efficiency field. The one thing I've learned from my adventure into the realm of high-efficiency speakers is that opinions are much more strongly held than other areas of audio - so it's always a good idea to find out what XYZ person considers a quality system.
Last edited:
You seem to favor compression drivers. But I am quite curious about the linearity of compression drivers because compression and expansion in the diaphragm vibration compression chamber is supposedly non-linear.
My experience with audiophile-class dome tweeters is they run out of gas above 95~100 dB SPL (at a 3 meter listening distance). Adding a waveguide/horn to this type of tweeter only buys a few dB more headroom - the main gain is directivity control. The other difficulty is that adding a waveguide/horn to a tweeter not designed for that mode of operation is non-trivial project to undertake - many physical prototypes, many measurements, and an uncertain outcome. Soongsc, I admire the remarkable persistence you've shown in your project designing a waveguide for a dome tweeter, but I'm not all that interested in directivity control - it's efficiency and headroom I'm after.
There are also efficient ribbons these days - but trying to use them below 2 kHz is asking for trouble.
There are not many choices if you want significant headroom in the 700 Hz to 10 kHz range. The only choices I see at present are horn-loaded (pro-monitor) midrange cones crossed over to small-format compression drivers & horns, or large-format compression drivers & horns.
There are also efficient ribbons these days - but trying to use them below 2 kHz is asking for trouble.
There are not many choices if you want significant headroom in the 700 Hz to 10 kHz range. The only choices I see at present are horn-loaded (pro-monitor) midrange cones crossed over to small-format compression drivers & horns, or large-format compression drivers & horns.
Last edited:
Re midrange - I like to crossover the midrange below 300 Hz for several reasons. the phenolic midrange Altecs with a good one inch driver above it crossed 300 and 3K can sound good. The 288's struggle that low and don't sound as good as the phenolic brother in the midrange. Good one inch drivers sound much better then the 288's in the treble. I prefer a cone though crossed around 200 to 2K. A horn loaded cone is better then the Altec phenolics and big TAD, JBL, ect drivers in the midrange but struggle a bit in the top of the band and typically force a lower crossover point then a decade (bad!)- or even better is several good cones (small diameter, low mms, sensitive and smooth to begin with) in a tight circular array around a one inch driver and horn. Done right it offers the efficiency (and can be crossed loer - 200 Hz or lower depending on several factors) of a large horn as well as controlled directivity without having a large midrange horn spaced far away from the treble horn. (It requires a big horn to load the mid below 300 Hz in free space) With the right spacing between the mid drivers and treble horn and the right crossover point to the concentric treble horn you can get great results with hair trigger dynamics, pure tone and less overall coloration and fuss as using a large midhorn. I'm sure there are plenty here that will say this won't work 
but of course it does.
but of course it does.
Interesting concept! What size midrange would you use in your "ring of fire"?
Got pix? Sounds like fun, anyway.
Agreed that the 288 will struggle below 500Hz, even on a 300Hz horn. That's what the phenolic frams were for. But there are those who do not like the sound of the phenolic. I can vouch for the symbiotic frams sounding worse than the all aluminum. But I've not used the phenolic, so can't say from personal experience.
Got pix? Sounds like fun, anyway.
Agreed that the 288 will struggle below 500Hz, even on a 300Hz horn. That's what the phenolic frams were for. But there are those who do not like the sound of the phenolic. I can vouch for the symbiotic frams sounding worse than the all aluminum. But I've not used the phenolic, so can't say from personal experience.
four to six inch work for me. Eight 5.25 with an 8" round treble horn is what I listen to now for midrange and treble. It's cheap to try, even cheap blow out drivers with good response can get you an idea of the what the array can do. Of course they beam with a lobe all around them but that is a good thing. I think of them as a super coax. The low mid can be outstanding and 'fast' compared to say a 15, 12, or ten inch coax. the upper crossover point is very critical though due to spacing between the cones and cones and cones and horn driver. Once it's right it's seamless. I'm not here to "sell" this idea, I just wanted to suggest what may be wrong with these GPA drivers.
Last edited:
I think you may be right. Currently I'm not aiming for such high levels. But I was curious because how audible the non-linearity of the driver is.
Just recently, we did some listeing tests in a mixing studio of a well damped room. The speaker was a full range driver. This was the first time I actually hear some effects of non-linear BL in this particular driver. Although Mr. Klippel had demonstrated this effect and explained it very clearly during some presentations using a model of more significant non-linearity, I was not totally convinced at the time until I actually hear it on my system. So I'm starting to wonder how audible linearity of other aspects effect sound.
One thing I have found is that the cleaner the driver is, the more power you can drive it with till it starts to be annoying. It also seems that with horned drivers, it's hard to get the decay to happen equally throughout it's used frequency.
Last edited:
Cool! We'd love to see pix. Never seen this type of rig before. Reminds me of the original bass reflex patent with a ring of vents around the woofer.
I've done something similar, but I used six, 6" woofers around a 6" horn. Think of a MTM with two 6" drivers on each side, to keep the width to a minimum. I liked it, but I then tried using four of the woofers to make a copy of Danley's Synergy Horns. I like it as well, but I couldn't implement either one to where I thought they were outstanding. They were just good.
No matter what the design, implementation is the key. A well designed tm by someone like Lynn or Zaph is going to sound better than any mediocre design by myself. However, I really enjoy trying something different, rather than the typical designs you see.
John
Last edited:
Paul has realized a quite similar setup *without a horn* in the middle in his great system
http://www.diyaudio.com/forums/multi-way/90804-large-midrange-ob-scott-g-9.html#post1955350
Putting a horn in the middle seem to be *really* challenging to get the blend over right
Michael
Thanks, Truesound, for describing in detail what you're doing. It's always illuminating to read an in-depth description of an enthusiast's system - it tells you the key design priorities, and the path that was taken to get from here to there. I admire the original solution you've come up for a high-efficiency loudspeaker - I've never seen a commercial equivalent to what you're describing.
My reservation about arrays is the difficulty with time arrivals when slightly off-axis. Nit-picky, I know, and kind of pedantic when looked at closely. But even theoretically perfect transducers will behave this way - the impulse is perfect only at one point (directly ahead), and breaks into discrete packets at all other points. Low-passing the transducers mitigates the problem by blurring the edges of the transient, but doesn't change the fact the transients are offset from each other at all emission angles except directly head-on to the drivers.
Some designers consider this misalignment of arrival times a useful method for restricting dispersion - and in the frequency domain, it kind of does that. But in the time domain - where I mostly work - the misalignment is more evident, and cannot be equalized or corrected, since it varies with listening angle.
Crossovers do all these evil things, but at least the harm is restricted to the crossover region. With a broadband driver array (of any shape), the time-arrival misalignment occurs over the entire working range of the drivers, and gets progressively worse the further off-axis you get.
I'm not letting horns off the hook here. Many horns, particularly old-school horns, have gross degradation of pulse response off-axis, and aren't too good on-axis, either. Diffraction in the throat and mouth exact a severe penalty in the time domain, and are not readily equalized away.
Notwithstanding all the theoretical worries mentioned above, a coaxial driver array mounted around a small-format horn is a really neat idea. It certainly is an elegant alternative to a conventional coaxial driver, with all the problems they have with small horns and reflections between the horn and the cone of the bass driver. There's an interesting continuum of design between a conventional coaxial driver, the Unity horn with its array of drivers within the horn, and your circular array of drivers right outside the horn.
My reservation about arrays is the difficulty with time arrivals when slightly off-axis. Nit-picky, I know, and kind of pedantic when looked at closely. But even theoretically perfect transducers will behave this way - the impulse is perfect only at one point (directly ahead), and breaks into discrete packets at all other points. Low-passing the transducers mitigates the problem by blurring the edges of the transient, but doesn't change the fact the transients are offset from each other at all emission angles except directly head-on to the drivers.
Some designers consider this misalignment of arrival times a useful method for restricting dispersion - and in the frequency domain, it kind of does that. But in the time domain - where I mostly work - the misalignment is more evident, and cannot be equalized or corrected, since it varies with listening angle.
Crossovers do all these evil things, but at least the harm is restricted to the crossover region. With a broadband driver array (of any shape), the time-arrival misalignment occurs over the entire working range of the drivers, and gets progressively worse the further off-axis you get.
I'm not letting horns off the hook here. Many horns, particularly old-school horns, have gross degradation of pulse response off-axis, and aren't too good on-axis, either. Diffraction in the throat and mouth exact a severe penalty in the time domain, and are not readily equalized away.
Notwithstanding all the theoretical worries mentioned above, a coaxial driver array mounted around a small-format horn is a really neat idea. It certainly is an elegant alternative to a conventional coaxial driver, with all the problems they have with small horns and reflections between the horn and the cone of the bass driver. There's an interesting continuum of design between a conventional coaxial driver, the Unity horn with its array of drivers within the horn, and your circular array of drivers right outside the horn.
Last edited:
directivity control *is* the same as controlling impulse response (coherence in time of arrival) over radiation angle
So - you have to decide
I have summarised that in short here:
http://www.diyaudio.com/forums/mult...erman-diy-show-gelsenkirchen.html#post1993650
To give an practical example what I mean - below you see impulse response measurements of my latest NEO3 quasi min phase dipole horn at 0 10 20 30 deg
The according FR you can see over in the Manzanita thread
http://www.diyaudio.com/forums/multi-way/110583-fast-fun-inexpensive-ob-project-4.html#post2005141
You can expand that concept to more than one driver easily (meaning - in theory its easy) if you think about the Quad 63 for example...
Michael
Last edited:
I was wondering when CSD data was going to be posted on that design. Didn't they promise to post some data?
Additionally, if the impulse is like what you post here, the response cannot be like the one in this link.
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-334.html#post1930266
Oops, the vertical scale of your impulses are different... Non-minimum phase.
Last edited: