Clearly I have been looking at too many Fostex curves:
http://www.madisound.com/pdf/fostexdrivers/fe168ez.pdf
This one could cut your arm off and then club you with it....
http://www.madisound.com/pdf/fostexdrivers/fe168ez.pdf
This one could cut your arm off and then club you with it....
Hi all, back again.
You guys are "on your own" more than you might think. I expect that it'll two or three months until I can be serious about building or listening to anything at all, and there's no point in you guys waiting on my rehab process to reach completion until you build your own. Beside, tastes in drivers are very personal - we're getting lots of suggestions for PHL and Fertin, and far be it from me to prevent anyone from trying and falling in love with them.
If I had a blank sheet of paper I would want, in about this order:
1) Field-coil magnets for big wideranger (12") and HF driver. Neodymium for the HF if field-coils aren't buildable. Alnico sounds wonderful too, in a rather different way, but I like it a lot. I suspect the greatest audibility of magnets is in the 1kHz and lower range.
2) Really good response from the wideranger, whether coax or freestanding. This part is NOT optional. I can build complex EQ with notch filters, but I truly don't want to, since complex EQ has the potential of severely diminishing sound quality. Note about big-driver rolloff curves - if the dipsy-doodles are caused by nothing more than non-coincident arrival from different parts of the cone, that's benign. Breakup, though, is not. Pink-noise audition will tell the tale here - raspy sound is a no-no, and very difficult "fix in the mix". Pink noise should sound like falling water, not a tin can.
3) Hemp cone would be really nice, but if Hemp Acoustics or Tone Tubby are snared in patent wrangles, are very late or non-deliverable, or there's a refusal to build versions with acceptable response curves, well, that takes them off the table. The 12" Alnico TT sounds nice subjectively, but I'm concerned it might lose a lot of its sparkle and immediacy after equalization. The HA 8, 12, and 15 coaxes have unknown LF and HF response curves, and availability to the DIY community isn't known yet.
4A) The Tweeter! Argh! Doggone it! All this arguing about horn profiles! Not to mention that in my (limited) experience, horn loading does nothing to remove inherent diaphragm colorations - subjectively, it magnifies the inherent sound of the diaphragm material, a sort of under-the-microscope impression. Metal diaphragms sound - well, metallic - having a positive impact on the sound of brass instruments but not doing any kindnesses for violin and soprano tone. We have to be careful interpreting the measurements for HF drivers, which don't tell the whole story.
It's the beauty thing - some people don't care about this, but a speaker that sounds beautiful is extremely important to me. I don't see beauty as a "euphonic" artifact, but rather the ability to convey a subtle truth about the music. This is the single weakest area of high-efficiency speakers, in my opinion.
4B) And ribbons! Argh again! Now we get big problems with zippo vertical dispersion - no more than a few inches wide at the listening position - and zippo upper-midrange power-handling, with ugly requirements for fourth-order crossovers at fairly high crossover frequencies. There's also the nagging business of ribbon sag - I'd like to slant the OB about 5 to 8 degrees backwards so you don't have to listen to on-axis response of the mid driver (almost no driver sounds its best precisely on-axis), and I have a bad feeling what this will do to the ribbon vertical dispersion and linearity as it sags in the magnetic gap.
4C) Fullrangers. Yes, I know, people love these things, no electrical crossover, doncha know. But I don't like the sound of those miserable whizzer cones, which to me are pretty sorry excuses for tweeters. How many people would love them if they treated them as tweeters and only listened to the HF part through a normal crossover? I thought so.
4D) Believe it or not, I have a soft spot for paper-cone tweeters. Man, I'm really dating myself with that comment. There's something sweet, gentle and true-to-life from the best of these - but the efficiency - and power-handling - is about 10 dB too low for this project. Arrays? No.
5) Fortunately, we have a pretty free hand with everything from 500 Hz on down. That's whole principle of overlapping driver responses - wavelengths are starting to get nice and long - and I am attracted to using drivers with complementary tonal characteristics in this region. I remember hearing a comment a long time ago from one of the world's most famous recording engineers - he liked to combine dissimilar bass drivers and run them together, so they delivered a wider tonal palette than either driver could alone.
I agree with this. These days I'm concerned with subtractive colorations as much as additive ones. By "subtractive" I mean a type of coloration where the driver is incapable, despite any amount of EQ, of reproducing a certain type of tone-color. An additive coloration, by contrast, is a tone-color that is always present, like an unwanted color cast in a photograph. The photographic equivalent to a subtractive coloration would be an inability to render certain colors at all - which, by the way, is present in all photographic technologies.
OK, rant off, but these are things I think are important, well, at least to me. People are free to make their own set of subjective priorities, but any speaker I design will have certain subjective qualities.
I still can't completely give up those blessed coaxes. I can see why the HF response is wrinkly if the horn profile is sub-optimal. I get that completely; horns are very very touchy about changes in profiles in mid-flight, and the transition from the machined center-piece to the cone/VC former and the rest of the cone isn't trivial. Fractions of a millimeter matter when it comes to horns.
What baffles me (no pun intended) is why the bass cone response is all crappy compared to the same bass cone without the HF assembly. Huh? OK, if the horn is overhanging and shadowing the cone, sure, that's going to really cause trouble with big reflections between the horn and cone, reflections that will be almost impossible to tame. Very similar to all the bad things that happen with whizzers, with the bonus of extra horn and crossover colorations too.
But with a Tannoy-style horn profile, there's no overhang, and better yet, no dustcap! Things should be just fine! But in practice we see some pretty dubious-looking curves compared to the straight bass cone. What's the deal here!?
You guys are "on your own" more than you might think. I expect that it'll two or three months until I can be serious about building or listening to anything at all, and there's no point in you guys waiting on my rehab process to reach completion until you build your own. Beside, tastes in drivers are very personal - we're getting lots of suggestions for PHL and Fertin, and far be it from me to prevent anyone from trying and falling in love with them.
If I had a blank sheet of paper I would want, in about this order:
1) Field-coil magnets for big wideranger (12") and HF driver. Neodymium for the HF if field-coils aren't buildable. Alnico sounds wonderful too, in a rather different way, but I like it a lot. I suspect the greatest audibility of magnets is in the 1kHz and lower range.
2) Really good response from the wideranger, whether coax or freestanding. This part is NOT optional. I can build complex EQ with notch filters, but I truly don't want to, since complex EQ has the potential of severely diminishing sound quality. Note about big-driver rolloff curves - if the dipsy-doodles are caused by nothing more than non-coincident arrival from different parts of the cone, that's benign. Breakup, though, is not. Pink-noise audition will tell the tale here - raspy sound is a no-no, and very difficult "fix in the mix". Pink noise should sound like falling water, not a tin can.
3) Hemp cone would be really nice, but if Hemp Acoustics or Tone Tubby are snared in patent wrangles, are very late or non-deliverable, or there's a refusal to build versions with acceptable response curves, well, that takes them off the table. The 12" Alnico TT sounds nice subjectively, but I'm concerned it might lose a lot of its sparkle and immediacy after equalization. The HA 8, 12, and 15 coaxes have unknown LF and HF response curves, and availability to the DIY community isn't known yet.
4A) The Tweeter! Argh! Doggone it! All this arguing about horn profiles! Not to mention that in my (limited) experience, horn loading does nothing to remove inherent diaphragm colorations - subjectively, it magnifies the inherent sound of the diaphragm material, a sort of under-the-microscope impression. Metal diaphragms sound - well, metallic - having a positive impact on the sound of brass instruments but not doing any kindnesses for violin and soprano tone. We have to be careful interpreting the measurements for HF drivers, which don't tell the whole story.
It's the beauty thing - some people don't care about this, but a speaker that sounds beautiful is extremely important to me. I don't see beauty as a "euphonic" artifact, but rather the ability to convey a subtle truth about the music. This is the single weakest area of high-efficiency speakers, in my opinion.
4B) And ribbons! Argh again! Now we get big problems with zippo vertical dispersion - no more than a few inches wide at the listening position - and zippo upper-midrange power-handling, with ugly requirements for fourth-order crossovers at fairly high crossover frequencies. There's also the nagging business of ribbon sag - I'd like to slant the OB about 5 to 8 degrees backwards so you don't have to listen to on-axis response of the mid driver (almost no driver sounds its best precisely on-axis), and I have a bad feeling what this will do to the ribbon vertical dispersion and linearity as it sags in the magnetic gap.
4C) Fullrangers. Yes, I know, people love these things, no electrical crossover, doncha know. But I don't like the sound of those miserable whizzer cones, which to me are pretty sorry excuses for tweeters. How many people would love them if they treated them as tweeters and only listened to the HF part through a normal crossover? I thought so.
4D) Believe it or not, I have a soft spot for paper-cone tweeters. Man, I'm really dating myself with that comment. There's something sweet, gentle and true-to-life from the best of these - but the efficiency - and power-handling - is about 10 dB too low for this project. Arrays? No.
5) Fortunately, we have a pretty free hand with everything from 500 Hz on down. That's whole principle of overlapping driver responses - wavelengths are starting to get nice and long - and I am attracted to using drivers with complementary tonal characteristics in this region. I remember hearing a comment a long time ago from one of the world's most famous recording engineers - he liked to combine dissimilar bass drivers and run them together, so they delivered a wider tonal palette than either driver could alone.
I agree with this. These days I'm concerned with subtractive colorations as much as additive ones. By "subtractive" I mean a type of coloration where the driver is incapable, despite any amount of EQ, of reproducing a certain type of tone-color. An additive coloration, by contrast, is a tone-color that is always present, like an unwanted color cast in a photograph. The photographic equivalent to a subtractive coloration would be an inability to render certain colors at all - which, by the way, is present in all photographic technologies.
OK, rant off, but these are things I think are important, well, at least to me. People are free to make their own set of subjective priorities, but any speaker I design will have certain subjective qualities.
I still can't completely give up those blessed coaxes. I can see why the HF response is wrinkly if the horn profile is sub-optimal. I get that completely; horns are very very touchy about changes in profiles in mid-flight, and the transition from the machined center-piece to the cone/VC former and the rest of the cone isn't trivial. Fractions of a millimeter matter when it comes to horns.
What baffles me (no pun intended) is why the bass cone response is all crappy compared to the same bass cone without the HF assembly. Huh? OK, if the horn is overhanging and shadowing the cone, sure, that's going to really cause trouble with big reflections between the horn and cone, reflections that will be almost impossible to tame. Very similar to all the bad things that happen with whizzers, with the bonus of extra horn and crossover colorations too.
But with a Tannoy-style horn profile, there's no overhang, and better yet, no dustcap! Things should be just fine! But in practice we see some pretty dubious-looking curves compared to the straight bass cone. What's the deal here!?
Variac said:
Measure them yourself on a flat baffle, you see the actual response minus the fairly obvious 1/3 octave smoothing of the factory measurements. Imagine razor-sharp sawtooth peaks where you see soft-looking lumps on the factory curves. Why these things are so popular I have no idea.
At least Lowther colorations are gratifyingly euphonic in the right horn, like the Big Fun, Oris, or Azura. Dunno how Lowther gets away with it - it's the best-sounding bad-measuring driver out there, and has been for fifty years now.
I still think the 18Sound 12NDA520 is in the running, with a standard inductance-correcting Zobel network (these have low subjective coloration) and a series inductor that creates a 1st-order rolloff at 1.5 kHz.
With a first-order lowpass at 1.5 kHz, the 12NDA520 crosses itself at 3.5 kHz. It could be "speeded up" with an additional 2nd-order lowpass network at the same or slightly lower (3 kHz) frequency. This would be a subjective decision, crossover coloration vs quite directional HF driver coloration. The suggested 5 to 8-degree tilt-back of the OB would make a real difference here, probably making the 1st/3rd order dual-slope filter unnecessary.
Considering the size of the driver, the comb-filtering in the off-axis response is to be expected, and likely benign. The rise in the on-axis response is a natural consequence of directivity, and not an unwanted artifact (this would appear in a theoretically perfect driver with no self-inductance).
The 18Sound XD125 elliptical spherical waveguide horn/driver looks like an interesting complement with a 12, 18, or 24 dB/oct highpass filter somewhere between 3.5 and 4.2 kHz. (A mild shelving network would probably also be necessary.) The diaphragm is polyester, not metal, and the horn profile is the (claimed) result of finite-element analysis, not the obsolete Webster equations.
Hmm, this is all looking too simple. Sure am curious about these drivers. Subjective reports, anyone?
With a first-order lowpass at 1.5 kHz, the 12NDA520 crosses itself at 3.5 kHz. It could be "speeded up" with an additional 2nd-order lowpass network at the same or slightly lower (3 kHz) frequency. This would be a subjective decision, crossover coloration vs quite directional HF driver coloration. The suggested 5 to 8-degree tilt-back of the OB would make a real difference here, probably making the 1st/3rd order dual-slope filter unnecessary.
Considering the size of the driver, the comb-filtering in the off-axis response is to be expected, and likely benign. The rise in the on-axis response is a natural consequence of directivity, and not an unwanted artifact (this would appear in a theoretically perfect driver with no self-inductance).
The 18Sound XD125 elliptical spherical waveguide horn/driver looks like an interesting complement with a 12, 18, or 24 dB/oct highpass filter somewhere between 3.5 and 4.2 kHz. (A mild shelving network would probably also be necessary.) The diaphragm is polyester, not metal, and the horn profile is the (claimed) result of finite-element analysis, not the obsolete Webster equations.
Hmm, this is all looking too simple. Sure am curious about these drivers. Subjective reports, anyone?
Lynn Olson said:
My experience has not yet encompassed anything that efficient (well i did have PM6As in 1975, but we were driving them with dyna 40W SS amp of the day, so you can guess what it sounded like :^)
Closest to that that i;ve had thru here would be Visaton B200 (with phase plugs)
dave
planet10 said:
seems that we will have opportunity to compare B200 with Sonido 8" in few months....
Vix have B200 (shouting,spitting still
) and Oly will have one pair of Sonidos ,and they're in neighborhood - in Belgradein fact-Sonidos will be in our dirty hands this weekend ( I hope) but I can promise first box for them so soon;
but-in case that I can convince lazy Oly to meet Vix........
while I'm at it...
poor man dream OB can be made with :
Eminence Beta15
Eminence Alpha 6 heavy tweaked
Eminence APT80 (or just diff horn -APT200)
rich man OB:
something good as 15" ALtec ??
something good as 12" Philips AD1256/Mx tweaked and little eq-ed
something good as smallest RAAL (kill me if I'm biased)
I know that this wasn't of much help.......
btw- I'm just listenin' 1256 on one side.......completely naked (I mean spk,not me)
Well I guess that the range of 12" widerange is very limited. I can think at 285-2000 , H30 or the coax version - the KM30, and the Fertin 30S (the link is for the 30S, the ferrite magnet, but there is a 30EX, electomagnet verison).
All these drivers have some advantages over the PA ones. There are designed specialy for the home audio. All of them have alnico or electomagnet motors and you don't see PA companies working with that (I can think only at TAD, which by the way should really be taken in consideration TAD )
It's clear that these drivers graphs are normalised but the fact is that they are addresing to home audio so I guess that actually they are more suitable than PA. All of them have SPL's around 97-100dB so I guess that's enough. If I could live with a lower cross point, almost sure I'd go with a 15" top of the line TAD driver such as TL-1601 (a, b, or c version).
I didn't mentioned that none of these drivers have whizzer cones. Probably Fertin or one of the distributors could provide a FR graph.
The more I think at this, I incline to believe that the TAD would be the better choise for one wanting a lower cross point. Of course, some mails with the staff are needed in order to provide the FR graps. Being a PA company they should provide a good reation with their customers.
The more I think at this, I incline to believe that the TAD would be the better choise for one wanting a lower cross point. Of course, some mails with the staff are needed in order to provide the FR graps. Being a PA company they should provide a good reation with their customers.
Zen Mod said:[you told him so already.....if I remember well.......or someone else certainly did
hehe-piece of broom stick,little sharpened .....can make wonders ........
yes i did... broomstick probably a bit fat... some 25mm dowell and some sand-paper would be a better start ... you can just leave them as a cylinder to start.
dave
Lynn Olson said:
No subjective reports, but it IS too simple, IMO. Have you considered the directivity of a 12 inch driver used above 1kHz? If you're at all concerned about HF waveguide directivity, I think this would be of equal concern. It is common practice to have equal directivity of HF horn (or waveguide, same thing IMO) and LF at the crossover point. You'd be far from this with a 10 or 12 inch driver crossed above 3 kHz.
Owen and Jeff Mai,
the dispersion mismatch would be large if a 12" driver is used, that was the problem I tried to point at.
I feel the 6" from 18sound looks really nice, no crossover, very high sensibility, low stored energy...
Now, since the idea has been here to use a larger driver for higher dynamics I have a question:
how bad would the combing effects be if a PAIR of 6" are to be used to 5kHz? The straight sides of the drivers make them placeable very close to each other, naturally in TMM config.
if this is bad, how about one of them with an appropriate low pass filter?
Owen,
The 6" measurements are taken with the driver on a 2 liter enclosure. It does not go very low, but I would not mind crossing my 15" drivers at that point, since they do very well up to 600-700Hz.
here a very pertinent quote from their website about things discussed in the beginning of this thread:
"A consistent heat transfer is guaranteed by the encapsulation of the magnetic structure in the interior of the basket, offering a large contact space between the back plate and the dissipating structure. The curvilinear cone-surround, has been created using computer aided vibrational modelling software to move all undesired bell modes out of the usable frequency range.
The 45 mm edge-wound voice coil assembly and terminals have been designed to minimise the moving mass while reinforcing force transmission."
the dispersion mismatch would be large if a 12" driver is used, that was the problem I tried to point at.
I feel the 6" from 18sound looks really nice, no crossover, very high sensibility, low stored energy...
Now, since the idea has been here to use a larger driver for higher dynamics I have a question:
how bad would the combing effects be if a PAIR of 6" are to be used to 5kHz? The straight sides of the drivers make them placeable very close to each other, naturally in TMM config.
if this is bad, how about one of them with an appropriate low pass filter?
Owen,
The 6" measurements are taken with the driver on a 2 liter enclosure. It does not go very low, but I would not mind crossing my 15" drivers at that point, since they do very well up to 600-700Hz.
here a very pertinent quote from their website about things discussed in the beginning of this thread:
"A consistent heat transfer is guaranteed by the encapsulation of the magnetic structure in the interior of the basket, offering a large contact space between the back plate and the dissipating structure. The curvilinear cone-surround, has been created using computer aided vibrational modelling software to move all undesired bell modes out of the usable frequency range.
The 45 mm edge-wound voice coil assembly and terminals have been designed to minimise the moving mass while reinforcing force transmission."