Crossover filter - Effects on the vertical radiation

[Tom Danley]

I think that the one-half wavelength criterion would be more than adequate were it not for the reflections from the ceiling and the floor. The little polar diagrams above you chart clearly suggest that this criterion should be adequate regardless. But I'm not sure those little polar diagrams are correct. Maybe it is for a specific crossover, but I'm not sure. Ordinarily, if the vertical spacing between the drivers is 1/2 wavelength, there will be a notch in the polar at +/- 90 degrees (for a phase coherent crossover where the lobe pattern is symmetrical above and below the horizontal). In the little polar diagram for lambda = 2D, there is no evidence at all of the notch that I think should be there.

" Ordinarily, if the vertical spacing between the drivers is 1/2 wavelength, there will be a notch in the polar at +/- 90 degrees"
Correct, I posted a ripple tank and source model links which show this is the way it is.

I was under the impression this was a DIY design thread and thought if it wasn't in stone yet, that they might model or investigate a non-traditional approach. What i don't see though are measurements of the drivers themselves as one is actually dealing with the magnitude and phase of the drivers and your adding the crossover mag and phase to them and for a passive crossover, you need the impedance's in your model too.

Best Regards
Tom

 

[MrPeabody]

What I really, really wanted to say is that I appreciate everything that ctrl does on ASR. You'd think that someone must be paying him a salary to do everything he's done!

This has helped to get me off the fence with respect to phase-coherent crossovers vs. odd-order crossovers. I've been bothered for some time by the fact that with the phase-coherent variety, the power response has that -3 dB dip at the crossover frequency. You can see the evidence of this dip in the polar response plots. With the phase-coherent crossovers, you get two secondary lobes, one above the horizontal and one below. With both of them, the peak SPL at the middle of the lobe is weak compared to the main lobe. Similarly, all of the lobes with the phase-coherent crossovers are rounded, indicating greater attenuation approaching the full null, as compared to the lobes in the off-order crossovers. These effects are surely attributable to (or can be explained by) the dip in the power response. Compare this to the odd-order crossovers, you see two lobes that are nearly equal in vertical breadth and forward extent, neither of them more prominent than the other. The only downside to the phase-coherent LR crossovers that has ever occurred to me is the dip in the power response. But as I look at these polar plots, for the first time I think I have a good sense of the true consequence of those dips in the power response, and it just doesn't matter. If anything, it is desirable because the energy radiated far off-axis vertically (up and down) is less compared to the odd-order crossovers that do not have this dip. For my money, the way to go is with phase-coherent LR crossovers and with the least possible vertical spacing between the two drivers. This does however mean that instead of using the fore-aft offset of two drivers mounted on the same flat baffle, to correct for the tilting of the lobe, you're stuck with this and have to deal with it some other way. But if you build a 3-way speaker, the phase offset that is due to the fore-aft offset of the diaphragms for two of the drivers is likely insignificant in comparison to the wavelength at the crossover point.

 

[MrPeabody]

" Ordinarily, if the vertical spacing between the drivers is 1/2 wavelength, there will be a notch in the polar at +/- 90 degrees"
Correct, I posted a ripple tank and source model links which show this is the way it is.

I was under the impression this was a DIY design thread and thought if it wasn't in stone yet, that they might model or investigate a non-traditional approach. What i don't see though are measurements of the drivers themselves as one is actually dealing with the magnitude and phase of the drivers and your adding the crossover mag and phase to them and for a passive crossover, you need the impedance's in your model too.

Best Regards
Tom

I'm slowly trying to improve my understanding of everything that goes into good crossover design. It is coming together slowly in bits and pieces. One part of it that I still haven't tackled is the phase change associated with the driver's natural rolloff. There is an article on this subject in an issue of Audio Express or Voice Coil, that has been in the back of my mind for a while, that I want to read. All I need is some more of those round tuits. Also I think that after five or six more readings of several articles on the use of all-pass filters to implement phase shift, I may also possibly have that part of it figured out!

 

[fluid]

.
Additionally, I cannot get my head around the idea that there could be any possible advantage of increasing the vertical separation of the two drivers. Much less the idea that there would be one particular value (1.2 x wavelength) that would be optimal in some way. This only determines the vertical polar angles above and below the horizontal where the notches occur. When the vertical separation of the two drivers is increased, the angular spacing between the two notches decreases. The main forward lobe becomes squished, flatter.

That is understandable as it goes against a lot of recommended practices. Kimmo Saunisto is the author of VituixCAD (the software ctrl is using for crossovers and visualization) in case that is not common knowledge. He has been promoting the correct measurement and simulation of speakers using time locked full polar measurements for a really long time. The advice may not be applicable for every situation but it is important to look at all of the points I quoted before and not just the crossover spacing. He is recommending a systems approach where the enclosure, the spacing and the phase match or mismatch is controlled to get the desired mix of on and off axis responses. They are often looked at in isolation but for an optimized design they need to work together.

As to the audibility of dips in the power response there was some old research done by Lipshitz and Vanderkooy that is quite relevant

https://www.aes.org/e-lib/browse.cfm?elib=11454

Conclusion for those without access

 

[MrPeabody]

That is understandable as it goes against a lot of recommended practices. Kimmo Saunisto is the author of VituixCAD (the software ctrl is using for crossovers and visualization) in case that is not common knowledge. He has been promoting the correct measurement and simulation of speakers using time locked full polar measurements for a really long time. The advice may not be applicable for every situation but it is important to look at all of the points I quoted before and not just the crossover spacing. He is recommending a systems approach where the enclosure, the spacing and the phase match or mismatch is controlled to get the desired mix of on and off axis responses. They are often looked at in isolation but for an optimized design they need to work together.

As to the audibility of dips in the power response there was some old research done by Lipshitz and Vanderkooy that is quite relevant

https://www.aes.org/e-lib/browse.cfm?elib=11454

Conclusion for those without access

View attachment 129052

The credentials of Kimmo Saunisto are reason enough to accept at face value that there has to be something to his idea. If you have access to a more complete explanation of the concept, I would find it interesting and will appreciate it.

I also appreciate the link to the article from way back in 1985. Seems so long ago. From the excerpt you provided I gather that they simulated a dipole radiator but controlled the frequency response of the rear-directed energy independently of the front-directed energy. And they found in essence that the dip at the crossover frequency, with a phase-coherent crossover, would be inaudible, although a peak of the same magnitude would audibly affect the perceived sound. The full article would no doubt be interesting.

 

[fluid]

Kimmo is guarded in the information he provides, he has stated his commercial partners are not overly happy with him providing VituixCAD to all and giving away the lessons learned through all his years of building and experimentation which they pay him for. Much of the information is spread about in different places and there is no unifying document or explanation.

You could read the conversation a few pages before and after this link to diyaudio where the information I copied above was posted.
https://www.diyaudio.com/forums/software-tools/307910-vituixcad-post6538511.html

There was an example at post 2781 in the above thread, image below for those who are not members

'Proof of concept' with Dave's project, c-c = 1.2 x wave length at 2 kHz XO = 206 mm

The measurement preparations and associated documents for VituixCAD are worth reading if you haven't already.

My own interpretation is that he is aiming for a smooth DI, a flat listening window with a down tilt beyond 7K, a balance between early reflections PIR and power response using as simple a crossover as possible to avoid corrections that only benefit the on axis response and not the off axis.

With the recipe for a two way being
"One or more from the following:
1) Tweeter (wave guide) much less directive at XO than woofer.
2) c-c distance ca. 1.2 x wave length at XO.
3) Box shape to decrease directivity at XO and increase directivity octave above XO (smoothly with diffraction without sharp edges).
4) Phase match octave above XO and possibly clear mismatch octave below XO."

As for the Lipshitz and Vanderkooy paper this how the experiment was conducted, send me a PM if you would like more

 

[dasdoing]

the consequences for a subwoofer corssover are all over the place afaiui? since we now have the "vertical" almost in a horizontal plane?

 

[BenB]

The power response and the directivity index are more or less two ways of looking at the same thing, one being inverted vs. the other. There is always a dip in the power response for any phase-coherent crossover where the individual responses are both -6 dB at the crossover point. This is necessarily true because power is a scalar quantity that sums in the simple manner of simple addition for two drivers, and -6 dB for each driver means that power from each driver (individually) is half what it is nominally, far from the crossover point. -6 db implies twice halving of power (halving of SPL and voltage), i.e., that power for each driver is 1/4 the nominal value, thus combined power will be 1/2 the nominal value, thus a -3 dB response anomaly at the crossover point. It won't go away so long as the crossover is phase-coherent for the two drivers and so long as both are -6 dB at the crossover point, such that pressure response will be flat across the crossover frequency.

For a coaxial (or more precisely, a concentric) design, there will be no dip in the power response for a phase-coherent crossover. The same is roughly true anytime the acoustic centers of the speakers are within 1/4 wavelength of each other. Your observations about voltage and power are misleading. A speaker's power consumption does not dictate its movement. Current is the driving force of a speaker, and it really should be no surprise that one can manage to achieve the same acoustic output with 1/2 current in one speaker, and 1/2 current in another speaker.

 

[BenB]

Personally, I have looked at the vertical dispersion characteristics like those shown here by ctrl
https://www.audiosciencereview.com/forum/index.php?attachments/1620568877232-png.128803/

and came across wondering not how to make the response in the crossover region match the response elsewhere. Instead, I have wondered how to make the response at all frequencies match the response at 2 kHz.

 

[KSTR]

W-T arrangement, 1/4 wavelength distance, 0 degree phase offset:

Note that this is for idealized spherical infinitesimally small point sources in free space (no baffle). In reality the pattern is much more benign, from driver directivity alone.

 

[dasdoing]

how to make the response at all frequencies match the response at 2 kHz

two woofers and a constant directivity horn

 

[BenB]

two woofers and a constant directivity horn

I guess I forgot to mention that I also want the response to be as wide as possible in the horizontal plane.

 

[dasdoing]

I guess I forgot to mention that I also want the response to be as wide as possible in the horizontal plane.

vertical array

e.g. https://www.audiosciencereview.com/...-70j-1-review-constant-beam-transducer.21469/

 

[fluid]

vertical array

Have a look at Ben's Avatar and read a little of his history with vertical arrays
https://www.diyaudio.com/community/...raight-cbt-with-passive-xos-and-no-eq.330031/

 

[TNT]

two woofers and a constant directivity horn

Most probably not. Any speaker with more than one source, and that has overlapping frequencies (possible solely in the x-o region only) - will interfere destructively in certain directions which result in a not flat power response. See D'appalito/MTM...

//

 

[dasdoing]

Most probably not. Any speaker with more than one source, and that has overlapping frequencies (possible solely in the x-o region only) - will interfere destructively in certain directions which result in a not flat power response. See D'appalito/MTM...

//

the idea is to create a line source. MTMs look caotic becaus you need a horn to restrict the vertical of the tweeter. also with a compression driver you generally can cross lower