Announcement

Collapse
No announcement yet.

Phase Question

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Phase Question

    Hi again everyone,

    I'm new to speaker design and still working on my first project. I have yet another question for this knowledgeable forum, this time around the subject of "phase".

    I really have no understanding of the concept of phase with regard to speakers, so I wonder if someone could explain it to me or point me to an appropriate learning resource?

    I've been working on a potential crossover design and it seems that I've managed to invert (?) the phase, based on what I can see from the resultant graphs from VirtuixCAD. Does these look wrong/bad to you?

    Click image for larger version

Name:	Phase.png
Views:	209
Size:	362.1 KB
ID:	1491254

    Thanks in advance, I look forward to hearing your feedback!

  • #2
    Phase is only an issue when the combined phase responses of the drivers is at or near 180 degrees apart, which will cause a cancellation sourced response dip. Each order of filtering results in 45 degrees of phase shift, so 2nd order/2nd order filtering results in 180 degrees total. When that's the case one cure is to reverse the polarity of one driver, usually the tweeter. The preferred method to eliminate the problem is to use 2nd order low pass/3rd order high pass. The total phase shift is 225 degrees, so there's no cancellation dip, with minimal expense. It also gives better protection and lower THD from the tweeter by reducing below pass band excursion.
    www.billfitzmaurice.com
    www.billfitzmaurice.info/forum

    Comment


    • #3
      Good news! You have NOT "inverted the phase" (based on your sim pics). What you're seeing in your 2nd plot is just a plot of how phase slowly "rotates" (at a given ear/mic distance) as freq. increases.

      If you play some bass notes through a (parallel-connected) pair of woofers (with them side-by-side) and then flop (reverse) the connections of ONE speaker, you should hear the bass "go away". That's because while one woofer is pushing forward, the other will be sucking in, and there will be very little "net" bass generated.

      The same effect can occur near your crossover freq (on a pair of drivers, like a tweeter and midbass), causing a cancellation at the Fc. In YOUR sim, reverse the connections (phase) of just the tweeter, and the "summation" line (the black one) will then show a "null" (big dip) at 2000Hz, your Fc. Not only that, but you can judge (off your sim plot) that your phase at the Fc is pretty good, 'cause you're summing almost +6dB (the diff. between where your drivers "hand off" - looking to be around 83-1/2dB?, and the black plot above that point, which is better than 89dB?) which only happens when your phase in near alignment.

      Comment


      • #4
        Drivers are in phase at the crossover point in your screenshot, so phase is fine in that regard, however it's everything else that looks wrong/bad. Can I go out on a whim and assume you provided some traced manufacturer data directly to the crossover design?

        Anyway, on the topic of phase, your understanding can be easily improved by using VituixCAD. Just load in the default drivers with flat response and some active high pass and low pass filter to provide some "text book" crossover. Adjust z axis of one driver to adjust delay of the driver (or use active gain/delay block), which adds "excess phase" to its result. Observe the interaction of frequency response and phase response. You should see, that when both drivers are in phase, their response adds together, and when 180 degrees out of phase, their response is cancelled out, and anywhere in between is something in between.

        Here's a little video illustration:
        https://drive.google.com/file/d/1uNT...ew?usp=sharing

        I'm not deaf, I'm just not listening!

        Comment


        • LewisH
          LewisH commented
          Editing a comment
          I have traced the manufacturer's, that's correct. My plan is to do this to buy the right drivers, then test them with a mic and measure the ts parameters myself, then sim that and buy crossover parts based on my measurements to try to replicate my original sim.

          Please explain what's wrong. This is my first build and i really appreciate the help because it's not a small amount of cash to blow!

        • dcibel
          dcibel commented
          Editing a comment
          Check your PM.

      • #5
        Excellent, that explains it perfectly, thank you to everyone.

        So, follow up question: is it just frequencies at cross over that can effect the sound with regard to phase? I.e. an out of phase 50hz can't interfere with a 3khz frequency.

        if ive understood correctly, i may have to look into a zorbel network or driver positioning more seriously to account for the crossover frequency interference.

        Comment


        • billfitzmaurice
          billfitzmaurice commented
          Editing a comment
          Phase only matters where the outputs of two or more drivers overlap at and near the crossover frequency. A Zobel is only used when you don't have impedance data. Driver positioning can be important when the HF driver acoustic center is behind the woofer acoustic center, which can result in a subjectively dull sound.

      • #6
        Phase is relating to time and distance, on it's own it's mostly irrelevant until you have multiple sources of audio playing, like multiple drivers in a system, or multiple speakers within a room. You can see in the little video I posted that the phase of a driver was adjusted, however it only affected the region where two drivers are playing.
        I'm not deaf, I'm just not listening!

        Comment

        Working...
        X