SVS subs - How does "phase" work?

Discussion in 'AV Receivers' started by John Meeks, Sep 12, 2006.

  1. John Meeks

    John Meeks Agent

    Joined:
    May 13, 2001
    Messages:
    32
    Likes Received:
    0
    Trophy Points:
    0
    I don't understand how the "phase" adjustment works.

    I know that a switch can reverse the polarity to the speaker.

    But, if I set the dial to, say, 90 degrees, how does it adjust the signal to that?

    If you go back to the switch analogy, at one setting, a current flowing from + to - would pull the speaker in one direction, and the other setting would pull it the other way. But, at 90 degrees, it would pull it in both directions equally, it wouldn't move, and you'd get no sound.

    I don't see how it's possible to change the phase besides some fairly complex processes that would introduce significant delay. The concept of phase doesn't even make sense with some signals (such as noise).

    Maybe the SVS guys can give me some insight.
     
  2. Kevin C Brown

    Kevin C Brown Producer

    Joined:
    Aug 3, 2000
    Messages:
    5,712
    Likes Received:
    0
    Trophy Points:
    0
    I can tell you how I think about it. [​IMG]

    You are correct in terms of polarity. Positive = 100% in phase (in theory) and negative = 100% (or 180 deg) out of phase (in theory).

    Let's say the crossover freq is 80 Hz. The idea is to adjust the phase between the sub so that the peaks and valleys in the sine waves from the main speakers, and from the sub all line up: peaks to peaks, and valleys to valleys. A phase knob allows you to do that. Essentially, it's an analog delay knob. It allows you to "slide" the sine waves of the sub over to match the sine waves from the speakers.

    There are a few ways to do it, but one easy way is to get a discrete test tone disc, play tones at and near the crossover freq, and adjust the phase knob until the you get maximum constructive interference between the sub and speakers ... or, the loudest SPL.

    One slightly better way of doing it (I think), but it's a different philosophy, is to map loudness vs freq, and adjust the phase knob for the flattest response across the crossover. If it's 80 Hz, look at 40 to 160 Hz, and try to get the flattest freq response.

    Make sure to use the correction values if you're using the Radio Shack meter.

    One reason why I think phase *switches* are completely unacceptable. You simply can't the resolution with a 0/180 deg switch that you can with a continuously variable knob.
     
  3. John Meeks

    John Meeks Agent

    Joined:
    May 13, 2001
    Messages:
    32
    Likes Received:
    0
    Trophy Points:
    0
    Thanks for your reply, but I was more asking how it works, than how to adjust it.

    If you have two overlapping tones, if it delayed the right amount for one of them, the other one wouldn't be at the correct phase. In order to do this, you'd need to somehow convert the signal some kind of frequency vs. volume graph, and then reconvert back to audio with the appropriate shifts. I bet this would sound awful.

    Since music is a mixture of so many different frequencies (sine waves), this seems like it would be extremely difficult.
     
  4. Geoff Gunnell

    Geoff Gunnell Stunt Coordinator

    Joined:
    May 20, 2006
    Messages:
    58
    Likes Received:
    0
    Trophy Points:
    0
    The best technical discussion is by Rod Elliott -- a link to his pages is in this thread from DIY forums:
    http://www.diyaudio.com/forums/showthread/t-77499.html

    Mr. Elliott takes a dim view of phase controls, although admitting they are 'useful'. If you read reviews of music subs such as the ACI Titan, you will find quite a different point of view. In my experience, unless you are going to be locating all your subs between your mains, continuously variable phase controls are a requirement for any sub.
     
  5. Kevin C Brown

    Kevin C Brown Producer

    Joined:
    Aug 3, 2000
    Messages:
    5,712
    Likes Received:
    0
    Trophy Points:
    0
    I wonder ...

    From the link: If you do it the in phase signal, out of phase signal way, and then shift between them, is it a constant shift? Or a shift independent of wavelength? In other words, 180 deg out of phase is with respect to all wavelengths. If you're simply sliding in between in phase and out of phase for all wavelengths, it's wavelength independent. So maybe less like an analog delay, which is more time dependent.
     

Share This Page