I have examined a number of speaker driver specifications and done some measurements on drivers in my possession and have learned that the phase response of individual drivers can vary considerably with frequency (already known to many of you, but news to me).
I know that some speaker designers will take the phase of individual drivers relative to each other into account when designing a speaker system and may use passive component filters in the crossover to match the phase of separate drivers at their crossover points, but what about a bi-amplified system in which each driver is powered by its own amplifier and the crossover is implemented electronically in the signal chain? The bare driver will still have its individual phase response. Assuming that the electrical filter is phase-coherent, should the separate drivers have passive components added to match their phase at crossover? Please note that I refer to an all-analog system, not to one with digital signal processing.
Can anyone offer informed comments, opinions, or suggestions for further research?
I know that some speaker designers will take the phase of individual drivers relative to each other into account when designing a speaker system and may use passive component filters in the crossover to match the phase of separate drivers at their crossover points, but what about a bi-amplified system in which each driver is powered by its own amplifier and the crossover is implemented electronically in the signal chain? The bare driver will still have its individual phase response. Assuming that the electrical filter is phase-coherent, should the separate drivers have passive components added to match their phase at crossover? Please note that I refer to an all-analog system, not to one with digital signal processing.
Can anyone offer informed comments, opinions, or suggestions for further research?
Yes that´s true. Recently I changed my woofer (JBL 2215H 15") to a dayton 385 15" and as I use active crossover I´m able to alter the phase of each drivers.
With the JBL I have to invert the phase of the midrange driver to have a smooth response in the crossover frequency (200Hz), with the Dayton I have to use the same phase of the midrange not the inverted one.
With the JBL I have to invert the phase of the midrange driver to have a smooth response in the crossover frequency (200Hz), with the Dayton I have to use the same phase of the midrange not the inverted one.
There is no fundamental difference between passive filter design and active biamped design, aside from impedence curve interactions.
I've never really believed in splitting the functions of crossover up into independent blocks, such as crossover section, bass shelving section, phase allignment section, etc. With smart design you can put them all together in a minimum number of optimized blocks.
Your basic understanding is correct that each driver will have a phase response from its bandpass and shape and also due to its physical location (effective depth). What really matters is the phase difference between the drivers. Adding crossover sections will tend to make the driver bandpass narrower. This will bend phase downwards for the woofer and upwards for the tweeter. Most of the time (really, all of the time) you can find the right combination of crossover slopes and driver polarity to get the sections of a multiway to have a combination of near-enough phase curves and good corner shapes to blend fairly seamlessly.
Read about Linkwitz Riley filter design but keep in mind that most writers of filter papers live in an ideal world of electrical filters with no messy driver responses or interunit physical delays.
David S.
I've never really believed in splitting the functions of crossover up into independent blocks, such as crossover section, bass shelving section, phase allignment section, etc. With smart design you can put them all together in a minimum number of optimized blocks.
Your basic understanding is correct that each driver will have a phase response from its bandpass and shape and also due to its physical location (effective depth). What really matters is the phase difference between the drivers. Adding crossover sections will tend to make the driver bandpass narrower. This will bend phase downwards for the woofer and upwards for the tweeter. Most of the time (really, all of the time) you can find the right combination of crossover slopes and driver polarity to get the sections of a multiway to have a combination of near-enough phase curves and good corner shapes to blend fairly seamlessly.
Read about Linkwitz Riley filter design but keep in mind that most writers of filter papers live in an ideal world of electrical filters with no messy driver responses or interunit physical delays.
David S.
Here's a superb write up by David that should help in understanding phase.
Crossover mods for the AR4x - Mods, Tweaks, and Upgrades to the Classics - The Classic Speaker Pages Discussion Forums
Crossover mods for the AR4x - Mods, Tweaks, and Upgrades to the Classics - The Classic Speaker Pages Discussion Forums
I would guess that your JBL is typical of JBLs in that it moves out when a positive voltage is applied to the black terminal. Ie reverse phase to that of every other driver.
Thanks to all for the replies. In measuring my drivers and experimenting with resistors and capacitors to flatten impedance and modify phase, it appears to be quite a challenge to arrive at the desired points for each element. It appears that I will have to settle for "good enough".
But, how do we define "good enough" in matching phase? Are there any rules of thumb that will help to guide us (for example, drivers no more than 45 degrees out of phase relative to each other, or each driver no more than 15 degrees out of phase relative to zero)? I am assuming drivers that are time aligned, with their "acoustic centers" located in the same vertical plane.
But, how do we define "good enough" in matching phase? Are there any rules of thumb that will help to guide us (for example, drivers no more than 45 degrees out of phase relative to each other, or each driver no more than 15 degrees out of phase relative to zero)? I am assuming drivers that are time aligned, with their "acoustic centers" located in the same vertical plane.
Many JBL Pro drivers are designed to move backwards when faced with a positive voltage unlike most other drivers.
Some Tannoys do that as well.
- Status
- Not open for further replies.