I've always been curious when I hear people say that drivers need to be within a wavelength of the crossover frequency to avoid lobing. This could make sense if you were way off axis, but generally the drivers are near ear height, so the actual distance from my ear to each driver is approximately the same, so shouldn't it not really matter how far apart they are as long as they're approximately the same distance to my ear? And if I'm actually critically listening then I'll be in the sweet spot which is where the drivers will be equidistant to my ears, and even if the drivers for most people setups aren't vertically level, a 1-wavelength distance between the drivers would result in pretty much a negligible difference in path lengths to the ear if the drivers are say 2 feet off of the listening axis vertically and 10 feet away. I haven't done the math, but I'd assume a 1 wavelength distance between the drivers would cause maybe a 1/4 wavelength difference in path lengths at most. If it doesn't make sense I can draw out an example of what I'm trying to describe. But I don't seem to think lobing is big of a problem as most people make it out to be and I don't really think a 1 wavelength or less separation is really necessary.
A good part of the reasoning is that with many crossover architectures, the drivers are not in phase with each other through the crossover. This makes the lobe(s) aim change through the crossover, sometimes into the floor, sometimes into the ceiling. See Linkwitz' article on the 4th order crossover SB1980-3way
That makes sense, but the added phase change from the difference in the driver distance would make a miniscule difference at best.
Lets assume for my drivers that the midrange is 1 ft above my ears vertically and 10 feet away horizontally. If for instance the tweeter is 6 inches above the midrange, then the distance to the midrange would be sqrt(120^2 + 12^2) which is 120.6 inches and the tweeter would be sqrt(120^2 + 18^2) which is 121.3 inches so a difference of .7 inches and at 5000hz the wavelength is 2.7 inches so it would be a phase shift of about 90 degrees which I can admit after the calculation that could potentially make a difference, but when a crossover already has 180 degrees phase shift for a second order. I just think it would be barely audible and I don't really get why people make such a big deal about it. I may be wrong about the 90 degrees making a difference but in the amount of phase difference already made by the crossover it just seems negligible to me
Um, no. Assume for a minute you're using first order crossovers. The direction of your lobes is governed by the phase difference, not by the driver setback. Please do read the article I linked. It's quite informative and interesting.
Understanding lobing is critical for successful crossover design. Here are a few simulations that illustrate the situation for a 2kHz LR4 crossover and 1cm dome drivers:
1. 10cm driver spacing
2. 20cm driver spacing showing lobe narrowing
3. 10cm spacing and -10cm acoustic center offset showing downward lobe tilt
4. 3rd-order Butterworth instead of LR4 showing upward lobe tilt
5. 6cm woofer showing (weak) lobe asymmetry.
1. 10cm driver spacing
2. 20cm driver spacing showing lobe narrowing
3. 10cm spacing and -10cm acoustic center offset showing downward lobe tilt
4. 3rd-order Butterworth instead of LR4 showing upward lobe tilt
5. 6cm woofer showing (weak) lobe asymmetry.
Attachments
Phase isn't the problem but it will affect the lobes. Lobes show lost power. Eg. a ceiling reflection may have a hole in its frequency response around the crossover. So you might overlap the drivers (in level) some more to compensate.
Lobing doesn't happen this way when the drivers already have controlled dispersion between the lobes.
It's one tool in a much larger unix-based analysis package I wrote. Its proprietary at the moment :-(
Last edited:
Hey Charlie. Sorry, that's the effective radius. In other words, cone diameter of about 4.7".
All speakers are a compromise. That's the first thing. You never get them perfect, because it's built into the physics. As soon as you split the signal in the crossover, it is detectable in the field.
What you CAN do is minimise effects. So, yes, getting the drivers close helps. It also helps if the drivers have vaguely similar dispersion. A tiny tweeter and a big bass do odd things to the sound.
This is the BW3 Sonics Anima by Joachim Gerhard. Lot to like there.
18dB/octave 90 degree phase BW3 is quite special. It has more consistent dispersion off-axis than 24dB/octave phase aligned LR4 which always has a notch above and below axis. In other words, BW3 doesn't change its sound so much with listening position.
Amusingly, a reviewer criticised Joachim's design for falling off above axis. He just flipped the polarity on the tweeter to fix that. The lobe then went the other way. 😀
What you CAN do is minimise effects. So, yes, getting the drivers close helps. It also helps if the drivers have vaguely similar dispersion. A tiny tweeter and a big bass do odd things to the sound.
An externally hosted image should be here but it was not working when we last tested it.
This is the BW3 Sonics Anima by Joachim Gerhard. Lot to like there.
18dB/octave 90 degree phase BW3 is quite special. It has more consistent dispersion off-axis than 24dB/octave phase aligned LR4 which always has a notch above and below axis. In other words, BW3 doesn't change its sound so much with listening position.
Amusingly, a reviewer criticised Joachim's design for falling off above axis. He just flipped the polarity on the tweeter to fix that. The lobe then went the other way. 😀
Attachments
And I completely understand that speakers a compromise and when available drivers should be as close as possible at a crossover that is high. But the reason I made this thread to begin with is I've seen thread after thread where someone has for instance a 6.5 inch driver and a 2 inch tweeter that would be about 5 inches center to center lets say. With the 1-wavelength "rule" this would with no actual calculation be somewhere around 2.5khz. I've seen many people say to never cross above that frequency even if it caused increased distortion in the tweeter. My point was that I think driver distance is a small effect compared to phase difference from the crossover and other things and shouldn't really effect crossover frequency choice. For example I have the rs100-4 and I may add a tweeter to it eventually and while I shouldn't cross above 4khz or so according to the rule, I think that's such a small effect compared to the fact that I want the full range to play as much of the range as possible.
TLDR I wasn't saying driver spacing has no effect on lobing, my point was just that it is a small concern that shouldn't effect decisions like crossover frequency much if at all.
Despite your incredulity, its good advice. Of course its healthy to question "common wisdom" in a field that is riddled with misconceptions, but the importance of minimizing driver spacing is not overstated. There is also a related concern with setting an upper-bound on the crossover frequency based on driver diameter. You want to ensure that the driver is not playing in the range where ka > 1 in order to avoid so-called beaming. The relevant metrics to measure these effects are the power response and size of the listening window. If you violate these rules then you (1) reduce the size of the spatial region where the FR is flat, and (2) increase the irregularity (decrease the smoothness) of the reflected spectrum of radiation.
Of course the crossover shape has a fundamental impact on the performance, but the character of the sound also depends on driver diameter and spacing. They cannot be ignored if you're serious about quality. Take a look at the woofer-tweeter spacing in the Infinity Primus series, for example. Obviously they are working hard to minimize the distance.
Its a small effect compared to what you "want"? By 5kHz, your RS100 will be about 10dB down at 60 degrees off axis. By 10kHz it will be 20dB down. You appear to be focused entirely on the axial response of your loudspeaker. You see, there is another bit of common wisdom that says "some speakers have low distortion and a flat on-axis response but sound mediocre, worse than some that don't measure perfectly flat on axis". Think about why that is.
But that's the same as saying the lobe size and shape is a non-issue, which is clearly inconsistent with the fact that the existence of a lobe implies a strong axial dependence of the frequency response. Either the smoothness of the frequency response doesn't matter, or the lobe geometry is important.
- Status
- Not open for further replies.