I'm neither fully in the objectivist or subjectivist camps - I try to keep a balance between measurement and subjective impressions when I develop or evaluate speakers.
I use measurements to guide me in the right direction and get me very close to the final design but the finishing touches like getting the attenuation between different drivers just right or getting the balance in the presence region or top treble just right I do by ear - at the end of the day if it doesn't sound right it isn't right and I won't enjoy them.
Interesting thing is in the 2 way system I recently finished I used as accurate as possible measurements and a lot of sim work and hand optimisation in Vituixcad to optimise for the flattest possible response, best phase tracking through the crossover etc, ordered all the components on faith and built the entire crossover, hooked it up and it was damn close to sounding right the first time, something I can't say on some previous attempts. 🙂 On previous speakers I've never had the same degree of measurement and simulation available to me to use and it sure makes the job a lot easier.
In the months that followed that I only made very small tweaks none more than about 0.5dB to get it to sound just right. Things like lowering the baffle step roll over point slightly (probably because I have the speakers somewhat close to the walls) lifting the presence region by only about 0.2dB, dropping the very high treble by maybe half a dB at the top - changes that are only just visible on the response graphs but made all the difference subjectively between "good" and "sounds right".
I don't know quite what gives ribbon tweeters their unique, clean sound either - but I suspect it's probably down to the almost complete lack of cone breakup (very clean waterfall plot) coupled with the high sensitivity and directivity control afforded by the waveguide.
I notice some of the same characteristics in other waveguide loaded tweeters that are not ribbons, so I think the directivity characteristics are definitely a good part of it. For example increased directivity can reduce cabinet diffraction problems greatly. I think everyone agrees that diffraction at high frequencies sounds bad, yet it is very hard to get rid of with traditional wide dispersion drivers and DIYable cabinets, with most typical cone + dome systems in rectangular boxes having severe diffraction issues in the treble even if they have round overs. Directivity really helps to solve this problem without elaborate and difficult to build cabinet designs, and this reduction in diffraction helps achieve a much flatter treble response in practice.
The current 2 ways use vintage Coral Flat 8 Mk2, (my avatar pic) and they work together very well with the ribbons, for the new 3 ways I've bought a pair of Tangband W8-1772 as a "modern equivalent" to do a similar job except not be covering bass as there will obviously be a separate woofer.
I've only done a quick rough measurement and full range listening test of the 1772 on a flat baffle so far (no time to do accurate measurements or start on cabinets yet) but I like what I see and hear - if anything I think they are smoother and superior drivers in many ways and look like they will need considerably less response shaping in the crossover than the Coral's did. (For example they don't have the pronounced 1Khz surround dip the Coral's have) So I'm quite optimistic about the results I'll get. It's just a matter of finding the time to work on the project.
Initially I'm going to aim for 4Khz 4th order linkwitz-reily but I always have the option to fall back to 3Khz as I know the tweeters can still manage that quite well.
I looked at the datasheet for the Scanspeak 10F and while the frequency response does look pretty smooth and workable it has nowhere near the needed sensitivity. The 12" woofers I have for the project are 92dB/W/M and will be mounted near the floor, (for deliberate floor gain) so midrange and tweeter will need to be at least 94dB/W/M to have any room for attenuation adjustment or shaping taking into account passive network losses. The AC G2 ribbons are 96dB/W/M and the Tangband is about 94-95dB/W/M depending on whose measurements you believe.
Last edited:
When I first got into audio, I was most concerned with bass. But as I have had more experience - my appreciation of tweeters has greatly increased.
The balance of extension and dispersion and power is much more difficult than it first appears.
My current favorite tweeter has not been mentioned yet, in this thread (at least that I saw), and that is Linaeum. These seem to break the rules. And there are two distinct versions - a 6" high version, and an 1.5" high version.
The taller version is in fact a midrange that covers the highs, too. It can be crossed over in the 400-500Hz range, and in larger arrays, it can go as low as 100Hz! It has very wide horizontal dispersion - because it effectively varies its width; depending on the frequency. The higher up it goes, the narrower the area it uses - it approaches a line source. Its weak point is vertical dispersion - though there are dipole versions that would tend to negate this, by having the backwave "fill in" the overall dispersion.
The shorter version covers a more conventional range - from about 2kHz and up. And it is usually a dipole, so it has the ability to have a huge soundstage. It also has the ability to move a lot of air - and this gives it an ease and clarity that almost never draws attention to itself.
There are circular tweeters that have some of the same characteristics - the dual ring radiator tweeter by Peerless is one example:
Peerless by Tymphany XT25TG30-04 1" Dual Ring Radiator Tweeter
The voice coil drives an inner and outer ring, so that it acts as a virtual POINT SOURCE - and so it maintains a much wider dispersion as the wavelengths approach the 1" diameter.
That was a "wait ... what?" moment for me - every driver gets directional as the frequency wavelength approaches the size of the the driver's moving element. So, for dispersion - a POINT SOURCE would be ideal.
But a point is by definition - very small, and therefore can't move much air. So, how a tweeter works comes down to extension and dispersion and power.
The balance of extension and dispersion and power is much more difficult than it first appears.
My current favorite tweeter has not been mentioned yet, in this thread (at least that I saw), and that is Linaeum. These seem to break the rules. And there are two distinct versions - a 6" high version, and an 1.5" high version.
The taller version is in fact a midrange that covers the highs, too. It can be crossed over in the 400-500Hz range, and in larger arrays, it can go as low as 100Hz! It has very wide horizontal dispersion - because it effectively varies its width; depending on the frequency. The higher up it goes, the narrower the area it uses - it approaches a line source. Its weak point is vertical dispersion - though there are dipole versions that would tend to negate this, by having the backwave "fill in" the overall dispersion.
The shorter version covers a more conventional range - from about 2kHz and up. And it is usually a dipole, so it has the ability to have a huge soundstage. It also has the ability to move a lot of air - and this gives it an ease and clarity that almost never draws attention to itself.
There are circular tweeters that have some of the same characteristics - the dual ring radiator tweeter by Peerless is one example:
Peerless by Tymphany XT25TG30-04 1" Dual Ring Radiator Tweeter
The voice coil drives an inner and outer ring, so that it acts as a virtual POINT SOURCE - and so it maintains a much wider dispersion as the wavelengths approach the 1" diameter.
That was a "wait ... what?" moment for me - every driver gets directional as the frequency wavelength approaches the size of the the driver's moving element. So, for dispersion - a POINT SOURCE would be ideal.
But a point is by definition - very small, and therefore can't move much air. So, how a tweeter works comes down to extension and dispersion and power.
Is that dual ring anything like what Stephens offered 60 yrs ago? (Except the Stephens,like a lot of their stuff was gorgeous looking too.)
Had a look at the specs at PE. At first glance the on-axis output was truly excellent, but the 30- and 60-degree seemed poor. And then I noticed the plot ran to 40kHz. OK, if Peerless wants to show off to 40kHz, they will confuse some of us. To normal hearing, the plots are quite excellent, even if a bit deficient beyond 20kHz.
Temping to find a quality tweeter that goes down to maybe 500 Hz. But I've always thought it wise to have a single driver cover the whole of the main range, say 200 to 2000 or more if possible.
B.
Had a look at the specs at PE. At first glance the on-axis output was truly excellent, but the 30- and 60-degree seemed poor. And then I noticed the plot ran to 40kHz. OK, if Peerless wants to show off to 40kHz, they will confuse some of us. To normal hearing, the plots are quite excellent, even if a bit deficient beyond 20kHz.
Temping to find a quality tweeter that goes down to maybe 500 Hz. But I've always thought it wise to have a single driver cover the whole of the main range, say 200 to 2000 or more if possible.
B.
Last edited:
Hi Ben,
I am not familiar with the Stephens driver - do you have a link for it?
I think Roy Allison's tweeter worked in a similar manner; though it gets there in a different way. It sounded phenomenal; but I am told its limitation is/was power handling. The Allison tweeter had no suspension, as such, and the voice coil induces the diaphragm to flex, which causes concentric ripples inward, and they propagate up to the center of the little peaked dome (more or less).
It is the concentric rings that keep the dispersion much wider, than with a (rigid) dome; since it comes much closer to being a point source. So, I think the Peerless dual ring tweeter is a (possibly) more robust version of what Roy Allison designed (but do not patent, I think?) - and the Peerless has a much lower resonance point - 436Hz! Which puts in closer to the Linaeum, for that matter.
The Linaeum is unique in many ways - and I can't emphasize enough, that having 5+ octaves from one driver is an amazing thing to hear.
I am not familiar with the Stephens driver - do you have a link for it?
I think Roy Allison's tweeter worked in a similar manner; though it gets there in a different way. It sounded phenomenal; but I am told its limitation is/was power handling. The Allison tweeter had no suspension, as such, and the voice coil induces the diaphragm to flex, which causes concentric ripples inward, and they propagate up to the center of the little peaked dome (more or less).
It is the concentric rings that keep the dispersion much wider, than with a (rigid) dome; since it comes much closer to being a point source. So, I think the Peerless dual ring tweeter is a (possibly) more robust version of what Roy Allison designed (but do not patent, I think?) - and the Peerless has a much lower resonance point - 436Hz! Which puts in closer to the Linaeum, for that matter.
The Linaeum is unique in many ways - and I can't emphasize enough, that having 5+ octaves from one driver is an amazing thing to hear.
Usually you see pictures of the Stephens coax tweeter in shiny chrome housing. It was also sold separately, I believe. This eBay link below may be showing the insides. The engineering smarts for Stephens (so sad they are long gone) derive from the engineers at Lansing, or Altec, or some history at Bell Labs....
Vintage Stephens Tru-Sonic 122AX 12-ohm Coaxial Speaker 12" Culver City, CA | eBay
Vintage Stephens Tru-Sonic 122AX 12-ohm Coaxial Speaker 12" Culver City, CA | eBay
I wanted to throw in some important observations
1. Ribbons souind very good at low levels, but they have a limited excursion and require a high Xover frequency.
2. They are super delicate, some may break when you only blow against. so have a replacement diaphagm at hand. Usually available, because you were not the first
So if you are listening to Classic or Jazz from vinyl you are the right guy to take care of them.
If you like trance and party with friends better get an air motion transformer
1. Ribbons souind very good at low levels, but they have a limited excursion and require a high Xover frequency.
2. They are super delicate, some may break when you only blow against. so have a replacement diaphagm at hand. Usually available, because you were not the first
So if you are listening to Classic or Jazz from vinyl you are the right guy to take care of them.
If you like trance and party with friends better get an air motion transformer
While I can appreciate the thoughts on diffraction, I would only add my own personal observations in that the best way for a speaker to disappear from its surrounding enclosure is a very narrow baffle......less surface to ripple and when the sound waves hit the 4pi space to the sides, the speaker as a system literally disappears into its surroundings. Now this all being said, the design has to be sensible to reach listening levels that most here can appreciate......and this is where a discussion about power compression and subwoofers would need to happen. While the 10F isn’t sensitive or efficient, it IS capable of clean playback before power compression when used properly. With these observations in mind, I would confidently say an effort of an narrow baffle satellite MTM with a pair of 10F’s and a ribbon would literally disappear into 4 pi space and image like the best systems of reference.
The only flaw with an MTM with a ribbon as the T, is the C to C distance and XO frequency, which by virtue of the size of the ribbon, would be unlikely to fall in the wavelength of 4kHz.
I have two pairs of small midwoofs I was planning to use with my CD3.5H.
They are Fountek 4" with good response up to at least 6kHz. The ribbon is almost as large as the woofer frame, and for that reason I plan to either:
* go MMT instead
* use something smaller like Tangband W2-800SL which is a 2" inverted dome tweeter so that the Woofs can be as close as possible
Ultimately, I plan to use the ribbons with Monacor 4" midrange, with flat response to 10kHz (simply the best paper cone midrange I have ever heard). The woofer I am struggling to choose, maybe Visaton GF200, a DVC woofer, which I will use one coil, for maximum bass extension in a closed box, for a compact 3 way
I have two pairs of small midwoofs I was planning to use with my CD3.5H.
They are Fountek 4" with good response up to at least 6kHz. The ribbon is almost as large as the woofer frame, and for that reason I plan to either:
* go MMT instead
* use something smaller like Tangband W2-800SL which is a 2" inverted dome tweeter so that the Woofs can be as close as possible
Ultimately, I plan to use the ribbons with Monacor 4" midrange, with flat response to 10kHz (simply the best paper cone midrange I have ever heard). The woofer I am struggling to choose, maybe Visaton GF200, a DVC woofer, which I will use one coil, for maximum bass extension in a closed box, for a compact 3 way
Last edited:
A problem with almost every MTM since the target C-C is a quarter wavelength at the XO. Half and full wavelengths are tossed into the discussion as reachable goals but are still not close enuff.
Every speaker is a compromise. The designer has to make his own.
dave
Best for what. By your title you are looking for justification to your preference, not an objective question.
Personally, I have not been thrilled with the ribbons I have heard. Apogee was not bad. Many that can be bought with only money have quite high distortion and are very fragile. Most have horrendous response curves that must be dealt with. Most DO stop on a dime. I have hear OF the Decca, but not heard it. Fountek and Raal have not impressed me.
When making these judgements, what dynamic driver are you comparing it to? Some even quite expensive drivers are actually pretty poor and even the so-so ribbons may sound better. Are you going to compare a well executed $1000 ribbon with a well executed $1000 diamond dome? Or a $29 soft dome. For comparisons, crossovers must have identical implemented curves so you are not mistaking midrange for treble. EQ and level must be identical. Good luck doing that! You can cross some ribbons @ 1000, but not a dome tweeter, so by definition, the playing field is not even. Maybe with active crossover and DSP correction from 3000 up, you might get close to a reasonable comparison.
Outside my budget and I have not exhausted what I know about domes, so I don't worry about it. Same category as AMTs. If I can afford it, I don't want it.
Personally, I have not been thrilled with the ribbons I have heard. Apogee was not bad. Many that can be bought with only money have quite high distortion and are very fragile. Most have horrendous response curves that must be dealt with. Most DO stop on a dime. I have hear OF the Decca, but not heard it. Fountek and Raal have not impressed me.
When making these judgements, what dynamic driver are you comparing it to? Some even quite expensive drivers are actually pretty poor and even the so-so ribbons may sound better. Are you going to compare a well executed $1000 ribbon with a well executed $1000 diamond dome? Or a $29 soft dome. For comparisons, crossovers must have identical implemented curves so you are not mistaking midrange for treble. EQ and level must be identical. Good luck doing that! You can cross some ribbons @ 1000, but not a dome tweeter, so by definition, the playing field is not even. Maybe with active crossover and DSP correction from 3000 up, you might get close to a reasonable comparison.
Outside my budget and I have not exhausted what I know about domes, so I don't worry about it. Same category as AMTs. If I can afford it, I don't want it.
I cant speak for the OP but...
I speak of a £35 Fountek ribbon, verses a £60 Visaton dome (not one driver I use, has costed more than £100, including the very good Visaton AL130)
The funniest thing about that, is the Fountek is very smooth, within +/-1dB until 10kHz, with a gentle rise of 3dB up to and beyond 20kHz, a very very good CSD, and low distortion.
But I guess, data means nothing?
If your opinion is so narrow, that you believe a £30 tweeter cant outrank a £100 one, then that's pure snobbery. (The cost of a driver in a loudspeaker is less than 5% the retail cost, y'know...)
Theres simply no comparison, for me.
Perhaps it is different, when one spends £1000, but mostly likely 99% of that is marketing hype.... and expectation bias
I can hear soft dome breakup a country mile away, sadly it has ruined my enjoyment of my speakers using them, no matter how good the graphs appear to be.
I cant speak for ring radiators, or, for example the SB donut domes, as I have not heard them. Perhaps they would sound better. I like the T27, despite it's obvious flaws, bit tweets like that, are extinct. Every CD I've heard has sounded like (exaggeration) a sewer pipe resonator. Id not consider spending £xxx when I can get something that sounds spooky real for an order of magnitude less cash.
Most importantly, in my opinion, the lack of ferrofluid, makes a very tangible improvement (or rather a tweeter designed without FF sounds better, for longer)
I speak of a £35 Fountek ribbon, verses a £60 Visaton dome (not one driver I use, has costed more than £100, including the very good Visaton AL130)
The funniest thing about that, is the Fountek is very smooth, within +/-1dB until 10kHz, with a gentle rise of 3dB up to and beyond 20kHz, a very very good CSD, and low distortion.
But I guess, data means nothing?
If your opinion is so narrow, that you believe a £30 tweeter cant outrank a £100 one, then that's pure snobbery. (The cost of a driver in a loudspeaker is less than 5% the retail cost, y'know...)
Theres simply no comparison, for me.
Perhaps it is different, when one spends £1000, but mostly likely 99% of that is marketing hype.... and expectation bias
I can hear soft dome breakup a country mile away, sadly it has ruined my enjoyment of my speakers using them, no matter how good the graphs appear to be.
I cant speak for ring radiators, or, for example the SB donut domes, as I have not heard them. Perhaps they would sound better. I like the T27, despite it's obvious flaws, bit tweets like that, are extinct. Every CD I've heard has sounded like (exaggeration) a sewer pipe resonator. Id not consider spending £xxx when I can get something that sounds spooky real for an order of magnitude less cash.
Most importantly, in my opinion, the lack of ferrofluid, makes a very tangible improvement (or rather a tweeter designed without FF sounds better, for longer)
Last edited:
IMHO a true MTM (midrange/midbass no more than 1/2 wavelength apart) is not feasible with a ribbon tweeter due to the high crossover frequency demanded by the ribbon and the relatively tall physical profile of the tweeter. You just can't get the drivers close enough at a crossover frequency the ribbon will be happy operating at. So I wouldn't even attempt it.
The demand for very low crossover frequencies and very close spacing of very small drivers is what turns me off MTM in general though so it wasn't really on my radar even for non-ribbons. In the attempt to achieve a symmetrical vertical lobe MTM throws the baby out with the bathwater.
Last edited:
It's also important to compare the response on a real, finite cabinet. A driver might have a very smooth response flush mounted on an infinite baffle but due to excessively wide dispersion suffer severely from cabinet edge diffraction that completely ruins the response.
Smooth published infinite baffle responses don't predict how a driver will measure on a real, finite baffle.
This is where a waveguide loaded ribbon tweeter (or any waveguide tweeter for that matter) is at a major advantage - put it onto a regular finite baffle and the response will be much less affected by cabinet edge diffraction - you will get pretty close to the infinite baffle response in practice even on a real cabinet.
So it might start off slightly less flat on a perfect infinite baffle but actually measure flatter on a real baffle thanks to the directivity. This effect can't be underestimated. Nearly all dome tweeters mounted on traditional small flat baffles have severe cabinet edge diffraction anomalies in the frequency response which can't be solved with a 1/2" round over.
Anyway, I don't know what ribbon tweeters tvrgeek has measured that have "horrendous response curves", but I wouldn't put mine into that category.
With a small amount of gradual EQ in the high pass crossover near the crossover point and at the very top they are +/- 1dB and smooth with no resonances all the way from the crossover at 3Khz to about 15Khz. (Above that my microphone and hearing are both not up to the task...)
This is the Vituixcad sim for the crossovers based on real in cabinet measurements of both drivers:
Tweeter looks pretty smooth to me, considering this measurement includes cabinet edge diffraction from a shoebox cabinet without even round overs...
They sound clean and smooth even at high SPL where the full range driver is starting to break up a bit with bass / midrange intermod...
Attachments
Last edited:
That's exactly what I was saying 😀
In Boxsim simulations the dome tweeter diffraction from baffle edge, is not even significantly improved, with a 4" round over nevermind a 1/2 inch round over.
Leaving 2 options:
Waveguiding, or limiting off axis radiation at 90 degrees
Use of felt baffle absorbent, to do the same
Last edited:
Yes, I was agreeing with you and elaborating on your point.
Unless you have the cabinetry skills to build something shaped like a KEF blade (and I certainly don't! 😀 ) the only reasonable way to achieve a low diffraction signature is to intelligently use directivity controlled drivers so that they're not radiating significant energy at 90 degrees along the baffle at high frequencies in the first place. ("High" starting somewhere around 1-2Khz where diffraction really starts to harm the imaging, frequency and polar response and also where it starts to become easier to achieve useful directivity with moderate size drivers)
Much easier for DIY'ers to build, (also arguably the superior approach, ultimately) and I wish more people would consider this line of development rather than sticking with very wide dispersion drivers on regular flat baffles, which just isn't a good approach.
Either go all out elaborately curved cabinet like the KEF blade or use directivity controlled drivers. Of course there's no reason you can't use elaborately curved cabinets and directivity controlled drivers together - but in that scenario the fancy cabinets add a lot more effort and cost that perhaps aren't justified when you already have directivity control and end up being more about looks than a necessity for performance.
Last edited:
In its day it was pretty much the only ribbon. The London is loaded by a horn that goes as low as 1000 Hz (i modded/heard a set of Decca prototypes with a London and an 8” bass driver. They were stunning.
I had a couple customers with Londons & 4 KEF B200.
dave
A proper MTM is certainly doable. One just has to ue a suitable midTweeter and a low enuff XO (ribbons that go low enuff are non-existent, i have seen a few that can reach 1kHz). These examples are stunning. Less than a quarter wavelength centre-to-centre at the XO.
dave
And the crossover frequency on the pictured speaker is what - 1Khz ? Way, way too low for my liking.
That's not even a dome tweeter or ribbon in that design which only serves to prove my point! 😉
An MTM forces very low "tweeter" crossover frequencies, much lower than I'm willing to work with.
If you want to use a driver that's not really a tweeter but rather a very small cone full range driver then that's a different story but has its own compromises and trade offs.
But it says nothing about designing an MTM using a ribbon tweeter or even a dome for that matter. I stick to my original assertion that MTM's using ribbon tweeters are a flawed concept because the necessary crossover frequency to function as a single lobe MTM is incompatible with the physical height and poor low frequency capability of a ribbon.
If that's how some people have experienced ribbon tweeters I can see why they might have been disapointed.
Last edited:
The XO is 250 Hz.
The midTweeter also does a pretty good job of not needing a dome or ribbon. There is more than one way to skin a cat. One has to choose their compromises.
And my example reinforces that.
dave
The midTweeter also does a pretty good job of not needing a dome or ribbon. There is more than one way to skin a cat. One has to choose their compromises.
And my example reinforces that.
dave