No announcement yet.

Speaker and component differences, for kits and L-R matching

  • Filter
  • Time
  • Show
Clear All
new posts

  • Speaker and component differences, for kits and L-R matching

    Question for the hive...

    ​When you setup a design, how much difference would you expect between the L and R speaker due to tolerances in the parts? For you top to bottom designers, do you adjust values between L and R to get as close of a match as possible?

    ​This also applies to kits. For example, purchasing a C-Note kit gets you the drivers and off the shelf parts. How much variance would you expect between kits and the pairs themselves? I suspect there isn't a lot of hard data on that, since those that buy kits typically do so because they haven't invested in measurement gear and how to use it.
    Electronics engineer, woofer enthusiast, and musician.
    Wogg Music

  • #2

    All transducers have a tolerance, therefore, yes, there is variation between left and right. By the time you factor in tolerance of crossover components, slight difference sin baffle diffraction, room modes, listening position, all the "noise" in tolerancing is averaged out

    If you design your own speaker from the ground up, you could do a different crossover between left and right, but I am guessing one would find that differences in values of crossover components are larger than the frequency response variation. For example, the next largest inductor may be too large and you are better off with staying with the same value.
    Zach Tripp


    • #3
      I think it's more of a question of "how much are you willing to accept" as a system designer. You can always add extra smaller components to balance out tolerances to get L and R matching. Those wider tolerances might even add a stereo flavor to your system as well, enhancing the experience. I'd imagine at the end of the day it's more a part of designer philosophy than anything else.


      • #4
        Drivers today are usually very close as far as specs go. Some places will " match" drivers for you for a couple of dollars more. Same with caps, inductors, and resstors. Some places will hand match, or order extras and macth them yourself.

        " Voicing is often the term used for band aids to cover for initial design/planning errors " - Pallas


        • #5
          Thanks peeps. Sounds to me like large variations are not typical, but of course small variations are expected. That's what I figured. I'm trying to wrap my head around a difference between the measured L / R results in my last project. Since I can't run an automated Monte Carlo analysis in WinPCD to see if the results could actually be expected given a 5 or 10% tolerance on parts, I'm stuck just accepting them as they are.

          Perhaps I'm being too stuck on matching the squiggly lines.
          Electronics engineer, woofer enthusiast, and musician.
          Wogg Music


          • #6
            Remember that tolerance spec is a spec of the component limits over temperature and voltage. Nothing is truly linear, but if your components are only stressed to 10% of the voltage rating and operate at room temperature, don't expect much variance over what the low level signal meter tells you. That said, if you use multiple components in parallel, the component value will be more likely to be closer to what you intend.

            Generally components made in batches will be similar to each other, but dissimilar between batches. Meaning that if you buy a pair of tweeters, chances are they are pretty closely matched, but if you buy one today and another in a year, who knows. A difference in tweeter sensitivity would probably be my greatest concern, since a wideband change of 1dB is quite audible.

            How much a difference in component size affects your design comes down to where the component lies in the circuit. Some may be very sensitive to the component value, some may be very tolerant. It is easy to adjust the values in your simulation and see how much a change in value affects the operation. If you've designed using real measurements, what I would do is run an impedance sweep of the system to see how close the built product matches what I've designed. Any deviations can be related to component differences, some small adjustment to the simulation should make it match reality very closely.
            "I just use off the shelf textbook filters designed for a resistor of 8 ohms with
            exactly a Fc 3K for both drivers, anybody can do it." -Xmax