Possibly silly question: how do you determine the max crossover frequency for an MTM?

[fjhuerta]

I've read the max distance is the equivalent of the wavelength between the center of the woofer and the center fo the tweeter.

But I've also read that it's the equivalent of the wavelength between the woofers' centers.

Which one is correct? I'm tending to believe it's the later, but I'm not sure...

[augerpro]

I always thought it was woofer to woofer center must be less than one wavelength. But I've heard some say half a wavelength. I'm not sure why. I found this article interesting: http://www.birotechnology.com/articles/VSTWLA.html

[Geoff Gunnell]

Javier, I'm not one of the experts here, but around the crossover frequency where all three drivers are operating, the distance between the woofers, being greater than the distance from woofer to tweeter, is the limiting factor.

One way to reduce the effect is to go to a 2.5 way crossover as Roman does in the second design here:

http://www.rjbaudio.com/RS180MTM/rs180-rs28-mtm.html

Staying with 2 way crossovers, you can't go much lower than 1200 Hz, and at that point you have a small selection of tweeters and LR4 or Cauer elliptical filters are necessary. Jon's Modula MTM Center design here:

https://www.htguide.com/forum/showthread.php?t=11475

Takes those design concepts to the limit.

[fjhuerta]

I always thought it was woofer to woofer center must be less than one wavelength. But I've heard some say half a wavelength. I'm not sure why. I found this article interesting: http://www.birotechnology.com/articles/VSTWLA.html

Very interesting. The formula presented seems like an interesting way to decide on driver spacing / crossover point.

Although I'm positive this can also be checked with LSPCad's off-axis tools...

[rc white]

I think you will find that in the original D' Appolito design the high frequency driver was a cd horn, not a dome tweeter, the tweeter will produce side lobes that are not a problem with the cd horn.

[Dennis H]

Although I'm positive this can also be checked with LSPCad's off-axis tools...

Maybe try ARPE to get a quickie feel for the off axis response. LspCAD will do it but ARPE is more convenient. You need Excel with VBA and the Analysis Toolpack installed.

https://web.archive.org/web/20080501095434/http://www.pvconsultants.com/audio/radiation/arpe.htm

All the settings are a bit confusing so load this profile for a basic MTM. It has 4th order, Q=.5 crossovers. A true D'Appolito would use 3rd order Q=.7 crossovers.

[fjhuerta]

Thanks for the link! I didn't know about ARPE. I'll try it.

As for the original question... any hints? I've googled the question, with no luck.

[Geoff Gunnell]

Javier, if you are interested in two-source interference patterns you might try looking at Lesson 1 and Lesson 3 here:

https://web.archive.org/web/20080501182142/http://www.glenbrook.k12.il.us/GBSSCI/PHYS/Class/light/u12l1b.html

You can calculate the angles of the lines of constructive and destructive interference from the frequency and the distance between the sources.

I love the fact that these are physicists, yet even they know that a live room evens out many of many of these nodal patterns.

I wasn't blunt enough in my first post -- Javier, you can calculate this all day long, but in a two-way MTM with 'normal' drivers you can't get the woofers close enough and the crossover frequency low enough to avoid the problem, all you can do is reduce it by using a low crossover frequency and/or a small frame tweeter. Whatever drivers and spacing you try, it won't look good on paper in a non-reflective environment.

Of course many folks including myself find that MTM's with good sounding drivers are a cost effective way of powering a residential room, but that's just budget-limited anecdotal subjectivism

[rc white]

The original idea was that the tweeter should match the drivers horizontal directivity at the crossover frequency, using the approximation..

angle = (58*lambda)/d

a nominal 8inch driver reaches 90 degrees at around 1500 Hz. and the original intension was to use a cd horn for the upper frequencies. If we use a 40x90 cd horn the two woofers can be spaced by around 350mm. center to center to give a 40 degree vertical lobe, with very little side lobing, this is important because a dome tweeter having wide directivity at 1500Hz. will inevitably generate a five lobe pattern, a cd horn will have a main lobe with two vestigial side lobes.
In general when using a dome tweeter the vertical lobe should not be less than 60 degrees to lessen the magnitude of side lobes

with an even order crossover a good rule is

lobe angle = arcsin(lambda/2d)

lambda is the wavelength of the crossover frequency, d=the woofer tweeter spacing.
A third order crossover characteristic gives a slightly fatter center lobe

Another rule of thumb that is used is

d = 13768/fx

d=woofer center distance in inches, fx=crossover frequency

both of these give similar results for a main lobe of 60 degrees.
rcw

[fjhuerta]

Thanks for all your help! I now understand a bit better how an MTM should work. I suppose the fast answer is "no matter how close the drivers are, it ain't enough!".

I'll just stick to the lowest frequency I can cross over that while considering the tweeters' distortion.

Thank you all for your help!

[jdybnis]

Thanks rcw. That was useful info.

[augerpro]

The original idea was that the tweeter should match the drivers horizontal directivity at the crossover frequency, using the approximation..

angle = (58*lambda)/d

a nominal 8inch driver reaches 90 degrees at around 1500 Hz. and the original intension was to use a cd horn for the upper frequencies. If we use a 40x90 cd horn the two woofers can be spaced by around 350mm. center to center to give a 40 degree vertical lobe, with very little side lobing, this is important because a dome tweeter having wide directivity at 1500Hz. will inevitably generate a five lobe pattern, a cd horn will have a main lobe with two vestigial side lobes.
In general when using a dome tweeter the vertical lobe should not be less than 60 degrees to lessen the magnitude of side lobes

with an even order crossover a good rule is

lobe angle = arcsin(lambda/2d)

lambda is the wavelength of the crossover frequency, d=the woofer tweeter spacing.
A third order crossover characteristic gives a slightly fatter center lobe

Another rule of thumb that is used is

d = 13768/fx

d=woofer center distance in inches, fx=crossover frequency

both of these give similar results for a main lobe of 60 degrees.
rcw

Great info RC. I'll have to chew on that a bit and play with the math and see what the results are.

[Dennis H]

d = 13768/fx

In other words, one wavelength c-c for the woofers.

[Curt C]

In other words, one wavelength c-c for the woofers.

That is the rule of thumb that I generally suggest.

I'll add that for vertically orientated MTM designs you can deviate somewhat from this rule with 'some' audible penalty, as the normal listening window in the vertical axis is fairly narrow. The change in vertical angle between sitting and standing in a typical listening room likely will be 20+/- 5 degrees. Assuming a woofer spacing of 1 wavelength at the crossover frequency, the summed response at those angles (ignoring room reflections) will typically be around -3dB at the crossover frequency. Doubling the mid/woofer spacing to 2 wavelengths results in reductions in response of –10dB to –20 dB under similar conditions, which I would consider unacceptable for most applications.

There is some merit to considering vertical off axis nulls as advantageous to alleviate floor and ceiling reflections. Turn the MTM on its side, however, and those same off axis aberrations will become very significant, as a listening position much further off axis is possible. This is why typical MTM's designed for vertical placement rarely work well when used on their side as a center channel speaker.

The plots below show modeled vertical polar response for a vertically oriented MTM with mid/woofer driver spacing at 1 wavelength and 2 wavelengths. The crossover frequency is 2kHz.

Edit: I noticed I was remiss in that I only showed a 2nd order LR response. I’ve added a 3rd order BW and a 4th order LR to the plots below. Note the improvement in polar response with the 3rd order topology. I chose this topology for the Statements for this reason, and its inherently better power response.

C

[chrismercurio]

I think...

respect needs to be given to the polar response pattern provided by whatever crossover you choose....but Lynn Olsen's Ariel didn't and used a VERY high crossover to great effect. It was built and admired by many.

C