[Danny Richie]

A follow up question. What happens when a planar or ribbon type tweeter is used in multiples?

Steve

If you run them in a line with the playing surfaces within or less than an inch apart then you get about the same response as a single continuous driver of the same length.

[Danny Richie]

Just what I was going to ask, Steve. Amongst Eminent Technology LFT-8 owners, there is a set-up called "stacking" two pair of the speakers. It's not literal stacking, it's placing two of the speakers (which are mirror-imaged, tweeters on opposite sides of the left and right speakers) side-by-side, with the edges that the quasi-ribbon tweeters are mounted right up against placed together. This creates a sort of large co-ax driver, with the two quasi-ribbons surrounded by magnetic planar mid/woofers. The problem I saw with the arrangement is that, though the ribbons are next to each other, there is still a couple of inches separating them. Comb-filtering?

You will not only get comb filtering between the two lines of tweeters but within the planar magnetic sets of drivers as well. And the planar magnetic drivers already play up high enough that they cause cancellation effects with the tweeter in the off axis. Even on their website it clams an accurate phase relationship only to 20 degrees off axis. Beyond that there are some out of phase holes in the response caused by comb filtering. Stacking two side by side like that would greatly compound the problem and limit the sweet spot to about a foot wide spot in front of each speaker.

[Danny Richie]

The basic purpose of a waveguide is to control dispersion (popularly now also termed constant directivity), so that in the crossover frequency range, the dispersion of next lower frequency driver (woofer or midrange) matches that of the tweeter.  By limiting dispersion the waveguide also helps concentrate the sound into a smaller space thereby increasing ultimate output and efficiency. Amphion is just one company that successfully relies on waveguide design (they also use it to lower the crossover frequency).  Although looking similar waveguides are not considered horns which pressurize the sound wave.

By the way this whole effect we're discussing can happen with every speaker designed with a crossover if the size of the sound wave at the involved frequency is high enough compared to the distance between drivers being crossed over (which is one reason why midrange drivers are placed close to the tweeter in most designs).

So, as Tannoy knew decades ago and a few other vendors also apply nowadays, the best application of midrange/tweeter (or woofer/tweeter) is a coaxial driver design which takes care of the comb effect and provides the advantages of a waveguide.

Yep, that's all right on the money.

[harley52]

Hey Jackman,
 
 Here's what I was told about my speakers and from reading article's. To get a look at my speakers go to Vapor Sound to the Aurora. The White pr. are mine. The  waveguide of the tweet is rather large and the mid/bass has a much smaller wg. The wg on the tweeter makes it possible to lower the xover to 1000hz. How steep the xover is I don't know as I never asked, as I was very surprised with the xover point and had to know how that was possible!! Apparently, the wg  increases the output of the sound from the tweeter so the  xover can be lowered and another benefit is lowering of distortion as the tweeter isn't working as hard. Currently I have a sweetspot about 4' wide. Imaging suffers somewhat at the outer edges of the sweetspot but, there is still a credible soundstage. I look at imaging as an audiophile thing that has no place in music. It just isn't how we ear at a live event so it doesn't bother me.  That's the info as I understand it. I hope I haven't made it more complicated than it has to be in my explanation.

 Danny, do I have it right or I'm I just in the ballpark. I do look forward to the facts from Danny.

[bdp24]

You will not only get comb filtering between the two lines of tweeters but within the planar magnetic sets of drivers as well. And the planar magnetic drivers already play up high enough that they cause cancellation effects with the tweeter in the off axis. Even on their website it clams an accurate phase relationship only to 20 degrees off axis. Beyond that there are some out of phase holes in the response caused by comb filtering. Stacking two side by side like that would greatly compound the problem and limit the sweet spot to about a foot wide spot in front of each speaker.

Do Magneplanars also, then, suffer from comb filtering, as they have narrow quasi-ribbon or true-ribbon tweeters running vertically, side-by-side with wider quasi-ribbon or magnetic-planar midrange and bass drivers? Then there is Peter Gunn, who removes the Magnepan cross-over and replaces it with symmetrical first-order filters as part of his Maggie mod. Do the 6dB/octave slopes engender more comb-filtering than do higher-order slopes? Bruce Thigpen, the designer of the Eminent Technology speakers, credits the first-order cross-over in his LFT-8 for the excellent blend between the push-pull magnetic planar mid-range driver and 8" cone-woofer in that speaker that reviewers have noted.

[Danny Richie]

Hey Jackman,
 
 Here's what I was told about my speakers and from reading article's. To get a look at my speakers go to Vapor Sound to the Aurora. The White pr. are mine. The  waveguide of the tweet is rather large and the mid/bass has a much smaller wg. The wg on the tweeter makes it possible to lower the xover to 1000hz. How steep the xover is I don't know as I never asked, as I was very surprised with the xover point and had to know how that was possible!! Apparently, the wg  increases the output of the sound from the tweeter so the  xover can be lowered and another benefit is lowering of distortion as the tweeter isn't working as hard. Currently I have a sweetspot about 4' wide. Imaging suffers somewhat at the outer edges of the sweetspot but, there is still a credible soundstage. I look at imaging as an audiophile thing that has no place in music. It just isn't how we ear at a live event so it doesn't bother me.  That's the info as I understand it. I hope I haven't made it more complicated than it has to be in my explanation.

 Danny, do I have it right or I'm I just in the ballpark. I do look forward to the facts from Danny.

Yep, the wave guide will increase the tweeters output in the lower ranges and allow a much lower crossover point. 

[Danny Richie]

Do Magneplanars also, then, suffer from comb filtering, as they have narrow quasi-ribbon or true-ribbon tweeters running vertically, side-by-side with wider quasi-ribbon or magnetic-planar midrange and bass drivers? Then there is Peter Gunn, who removes the Magnepan cross-over and replaces it with symmetrical first-order filters as part of his Maggie mod. Do the 6dB/octave slopes engender more comb-filtering than do higher-order slopes? Bruce Thigpen, the designer of the Eminent Technology speakers, credits the first-order cross-over in his LFT-8 for the excellent blend between the push-pull magnetic planar mid-range driver and 8" cone-woofer in that speaker that reviewers have noted.

The side by side placement of the drivers in the Maggie does open the door for out of phase cancellation in the horizontal off axis. How much or how bad depends on the slope and crossover point.

Shallow slopes do often allow smooth driver transitions. But because each driver plays further over into the range of the next it does open things up for more cancellation patterns in the off axis when one driver is delayed in time verses the other. And while at the crossover point a given off axis may not have enough time delay for complete cancellation, but one octave above that there is still plenty of output left to cause cancellation and a big dip in the response. So with first order crossovers dips in ranges above the crossover point are common.

Crossover point is critical in these alignments. Let's say the crossover point is 4kHz. At that frequency range the wavelengths are less then 3.5" long. So 15 to 20 degrees off axis is enough to create 180 degrees or more of phase rotation. But if the crossover is low like our LS-9 for instance, then everything changes. The LS-9 crosses over at 850Hz. In that range the wavelengths are over 15.5" long. So 15 to 20 degrees off axis might only cause 10 to 15 degrees of phase rotation. Even 90 degrees off axis is only about 80 degrees of phase rotation. 

[bdp24]

About how much phase rotation need be present for it to cause an audible problem, with cancellation?

[Danny Richie]

About how much phase rotation need be present for it to cause an audible problem, with cancellation?

If phase rotation at a given frequency is out by 90 degrees between two drivers then they neither couple or cancel each out out. The output is the same summed as they are individually.

If you start reducing that phase rotation back to less than 90 degrees then they start to couple and you get gain with a maximum amount of gain at 0 degrees of rotation.

If you start increasing the phase rotation past 90 degrees then you get less output than from either single driver. And at 180 degrees out they are completely out of phase and cause maximum cancellation.

[Danny Richie]

Here is a good illustration.

These are the measurements of two 3" full range TB drivers. They are in separate boxes but the drivers are near the ends of the box. They are on top of each other keeping the acoustic centers to within 5" pf each other.

They are measured at 1 meter and with one watt. The mic is dead center between them.

Red and green lines are each driver playing individually, and the blue line is the two drivers playing together. Note too that it was not easy to get the mic dead center. Just a fraction of an inch up or down and cancellation would occur in the top end.



Now this is what happens when the mic is moved up 4". Note how the shorter wavelengths in the upper range are now out of phase.



Here is what happens moving the mic up another 4". It is now 8" above center. Note now that cancellation points are in two places.



Okay, up 4 more inches making the pic 12" above center. Now there are three cancellation points.



The change in vertical off axis hardly effects each individual driver, but you can see the curves change some from the mic being off axis to them. But the cancellation effects are tremendous. Dual tweeters stacked one over the other has the same effect.

This causes an irregular floor and ceiling reflection that takes away some of the airy spacial cues that help create a layered sound stage.

[bdp24]

Thanks for the Acoustical Engineering 101 class Danny. It should be required for subjective reviewers!

[Oscillate]

"Thanks for the Acoustical Engineering 101 class Danny."

+1

[bdp24]

The ET LFT-8 has a 10K Hz x/o from the midrange panels to the Quasi-Ribbon tweeter strip (not a typo.....10,000 Hz!), with 1st order slopes in both directions. Will stacking the tweeters next to each other (with, say, three inches separating them) result in comb-filtering, at that high a frequency? The lower x/o, from the panel to the woofer, is also 1st order, at 180 Hz. The Magnetic Planar driver has a 180 to 10,000 Hz bandwidth in the speaker! The cone woofer is 8", and it's in a 0.75 cu.ft. sealed cabinet. Anybody thinkin' what I'm thinkin'?

[Danny Richie]

The ET LFT-8 has a 10K Hz x/o from the midrange panels to the Quasi-Ribbon tweeter strip (not a typo.....10,000 Hz!), with 1st order slopes in both directions.

That will cause cancellation in the left and right off axis response as just a little movement in either direction will make them out of phase.

Will stacking the tweeters next to each other (with, say, three inches separating them) result in comb-filtering, at that high a frequency?

Sure. They would easily cancel each other out. The further you separate them the worse it gets.

The lower x/o, from the panel to the woofer, is also 1st order, at 180 Hz. The Magnetic Planar driver has a 180 to 10,000 Hz bandwidth in the speaker! The cone woofer is 8", and it's in a 0.75 cu.ft. sealed cabinet. Anybody thinkin' what I'm thinkin'?

Servo woofers baby! All open baffle!

[srb]

Danny, here is one other configuration that I've not seen anywhere else.  Although XTZ uses a single tweeter for all of their stereo speakers, their "Cinema Series" M6 (LCR) speaker has a "Quad-Tweeter array".

The explanations given are:

"M6 is separate Left and Right speakers to avoid mid mounted tweeter problem"

"The four 1" soft dome tweeters are arranged in a special way to ensure not only a perfect balanced dispersion between to cover the listening room whith same timbre and on the other side not to wide spreading to get negative influenced sound from the room .The controlled dispersion is possible because of the usage of four tweeter and this is also the reason for the clean sound even the level is very high ."

"3 of the tweeter works from 1.2kHz to 3kHz and they control the dispersion in this very sensitive frequency range.  This means no phase problem in the very, for the ear, sensitive range 2-6kHz there is no crossover point as in most speaker."

"The quattro tweeter also helps to control the spreading perfectly both in tweeter range and in the midrange. It gives a optimized vertical directivity to reduce floor and sealing reflections and a controlled dispersion in horizontal direction."

"Controlled directivity but still no “beaming” effect . Quatrro array give also advantage like avoid early reflections from sidewalls.The placement of the four tweeter improve the diffraction effect and the extra baffle improve it further.But by use take advantage of placement and using the baffelstep effect the energy is pointed into the listening room and less on the outside and this reduce the early reflections from sidewalls."

What is your take on this very unusual tweeter grouping?

Steve

 

[Danny Richie]

Steve,

If three of the tweeters are only playing up to 3kHz then there will likely be little cancellation between them even in the off axis. The 3kHz wavelength is about 4.5" long.

If they all played to the top then there would be a comb filtering effect that will limit the off axis to about the same as the small woofers used with them.

2kHz to 6kHz is not a critical range for the ear though. You are already getting out of fundamentals and into harmonics in those ranges.

And despite what they may think, it will NOT "help to control the spreading perfectly both in tweeter range and in the midrange". Most people don't realize that the mid-range is in the 300Hz to 500Hz range.

And I am not really sure what they are thinking on this one:

"Controlled directivity but still no “beaming” effect . Quatrro array give also advantage like avoid early reflections from sidewalls.The placement of the four tweeter improve the diffraction effect and the extra baffle improve it further.But by use take advantage of placement and using the baffelstep effect the energy is pointed into the listening room and less on the outside and this reduce the early reflections from sidewalls."

None of that is really true. And did they really use grammar errors in their ad? It reads like it was written by a Chinese fellow that really is not fluent in English.

[bdp24]

That will cause cancellation in the left and right off axis response as just a little movement in either direction will make them out of phase.

Sure. They would easily cancel each other out. The further you separate them the worse it gets.

Servo woofers baby! All open baffle!

Exactly! I'm using 12's, but three 8's would make for a narrower cabinet. And for he who, for whatever reason, doesn't want to go OB, one of the Servo 8's could be put in the sealed 0.75 cu.ft. box, replacing the ET woofer.