Announcement

Collapse
No announcement yet.

Uncharted Crossover Design Waters

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

  • Uncharted Crossover Design Waters

    As I continue getting my feet wet in the nuances of network design, I'm trying to acclimate myself to 3rd and 4th order passive designs. My question is this:

    In a two-way design, specifically for the tweeter side, do you bypass both series caps? And what about this cascading bypass method I've read about?

    TIA
    Dave H

  • #2
    Re: Uncharted Crossover Design Waters

    While you are just starting out the most important thing you can learn is that the audio community is overrun with idoliters.

    Capacitor bypassing in crossover filter circuits is just another pagan god.

    Bypassing has it's roots in power supply rails for equipment where low impedance is required through the radio frequency range or at some distance from the main capacitor bank where the inductance of the supply line necessitates additional local capacitance.

    At radio frequencies a large capacitor becomes an inductor or at least the supply line will if the large capacitor isn't right near by a critical point for an active circuit.

    A small capacitor in such an aplication will become the controlling element providing a low impedance point for your circuitry.

    At audio frequencies for the purpose of a speaker crossover all capacitors are always in their range and behaving like capacitors.

    Even the infamous non-polar electrolytic capacitor performs far beyond the capability of the best drivers.

    The difference between film capacitors and non-polar electrolytics is measurably dramatic but when you build them into the speaker the performance is dominated by the characteristics of the drivers and the capacitors become insignifigant.

    A bypass capacitor in the limited audio range simply contributes to the circuit in the same ratio in which it is implemented.

    That is if your bypass capacitance is 1% of the total for the main and the bypass then the bypass capacitors characteristics are contirbuting 1% to the circuit and the main capacitor is contributing 99% of it's characteristics.

    So if you use a non-polar electrolytic capacitor for your main capacitor and even though it's performance is beyond that of the best drivers in the world you decide to bypass it anyway with 1% of the total capactance coming from some type of film capacitor, then the difference will simply be a mere 1% of the way along the line between the performance of the non-polar electrolytic and the performance of the film (99% of the performance being determined by the non-polar electrolytic).

    Should you increase your bypass to 10% of the total then you simply have 90% of the capacitor performance being determined by the main capacitor.

    In a crossover filter a bypass capacitor always is a minority element while the main capacitor has near total control of the circuit.

    Another misunderstanding in audio is that crossover componets do not influence the circuit most at high frequencies.

    You've heard it said a million times that in the top octave is where you really hear the clarity or sweetness or what ever BS with a certain capacitor.

    Truth is that crossover componets have their greatest influence near the corner frequency of the filter created with them (crossover frequency) where their impedance is comparable to that of the circuit.

    As you get further from the a filter's corner frequency the impedance of crossover componets will go very high or very low and their influence over the circuit becomes less and less (the componets effectively are removing themselves from the circuit as you get further from the filters corner frequency).

    One of the biggest lie's in audio is the film and foil capacitor.

    Film and foil capacitors are claimed to be better because their ESR (equivilant series resistance) is lower.

    However the impedance curve for capacitors looks like a 'V'.

    It starts very high at low frequency and falls steadily until it reaches the capacitors self-resonance frequency (usually several hundered kilohertz depending upon the value of the capacitor) then the impedance climbs above the self-resonance frequency due to parasitic inductance (mostly due to lead length and the necessary length of those leads due to the size of the capacitor).

    ESR is only an issue in the vicinity of the self resonance frequency, for the rest of the spectrum the capacitors impedance is dominated by it's capacitance and parasitic inductance.

    Film capacitors typicly have very low ESR and as they approach their self-resonance frequency their impedance will usually deviate on the low side of what a perfect capacitor should be and then abruptly turn into an inductor.

    The problem with film an foil types is that their ESR is even lower and their deviation from perfect as they approach their self-resonance frequency is even greater than that of a metalized film capacitor.

    Not only that but film and foil capacitors are huge and because of their sheer size and the necessary lead length result in higher inductance and even more deviation from a perfect capacitor as compared to a metalized film type.

    Bypassing capacitors in speaker crossovers is never of any use save one.

    The capacitor bypassing issue is one of the fastest ways to find out that you are talking to or reading from a fraud to whom you never need pay attention to anything they say even though they could surely talk for hours about their cascading stupidity.
    Last edited by daryl; 09-21-2008, 03:40 PM.

    Comment


    • #3
      Re: Uncharted Crossover Design Waters

      Originally posted by Hackomatic View Post
      As I continue getting my feet wet in the nuances of network design, I'm trying to acclimate myself to 3rd and 4th order passive designs. My question is this:

      In a two-way design, specifically for the tweeter side, do you bypass both series caps? And what about this cascading bypass method I've read about?

      TIA
      You don't have to bypass caps with smaller values, so you can avoid them altogether. If you want to "trim" a small cap in parallel to meet your value-target, then so be it. Some people claim there is a sound improvement, but it could be in their heads, actually sound better, or be snake-oil. Only 3 options there. Personally, I don't think there is a reason not to do this if you want to.
      Later,
      Wolf
      "Wolf, you shall now be known as "King of the Zip ties." -Pete00t
      "Wolf and speakers equivalent to Picasso and 'Blue'" -dantheman
      "He is a true ambassador for this forum and speaker DIY in general." -Ed Froste
      "We're all in this together, so keep your stick on the ice!" - Red Green aka Steve Smith

      *InDIYana event website*

      Photobucket pages:
      https://app.photobucket.com/u/wolf_teeth_speaker

      My blog/writeups/thoughts here at PE:
      http://techtalk.parts-express.com/blog.php?u=4102

      Comment


      • #4
        Re: Uncharted Crossover Design Waters

        Wow!

        Very comprehensive primer on caps and xo circuits . . Thanks guys.

        One is left to conclude then, that those companies touting bypass caps and cascading bypass caps, merely want to sell more caps!
        Dave H

        Comment


        • #5
          Re: Uncharted Crossover Design Waters

          Yeah but 'pagan god's', frauds, cascading stupidity....have I stumbled into some sort of middle-earth editorial?

          That being said (and it wasn't saying much) it was fascinating reading!

          Cheers!
          Marty

          Comment


          • #6
            Re: Uncharted Crossover Design Waters

            Originally posted by Hackomatic View Post
            As I continue getting my feet wet in the nuances of network design, I'm trying to acclimate myself to 3rd and 4th order passive designs. My question is this:

            In a two-way design, specifically for the tweeter side, do you bypass both series caps? And what about this cascading bypass method I've read about?

            TIA
            Before you get into the idea of bypassing caps, let me ask you this - do you know that a fourth order network is right for your tweeter crossover? If you are not working with driver response measurements that include phase and the tweeter's impedance then it would be essentially impossible to just pull a circuit out and know it will result in a correct crossover for a given driver combo. If it is not optimized then this will be much more noticable than capacitor type would ever be. For example, most tweeters require a second order crossover near 2khz in order to produce a 4th order in-phase acoustic crossover. If a fourth order circuit is employed then it would combine for a 6the order roll-off and the tweeter would need to have the polarity reversed to avoid a null at the crossover. Of course, none of this means that it would sum correctly. Just something to think about. Now, if you are using driver measurements and decent simulation software then you will be able to work these things out.
            Click here for Jeff Bagby's Loudspeaker Design Software

            Comment


            • #7
              Re: Uncharted Crossover Design Waters

              Originally posted by Jeff B. View Post
              Before you get into the idea of bypassing caps, let me ask you this - do you know that a fourth order network is right for your tweeter crossover? If you are not working with driver response measurements that include phase and the tweeter's impedance then it would be essentially impossible to just pull a circuit out and know it will result in a correct crossover for a given driver combo. If it is not optimized then this will be much more noticable than capacitor type would ever be. For example, most tweeters require a second order crossover near 2khz in order to produce a 4th order in-phase acoustic crossover. If a fourth order circuit is employed then it would combine for a 6the order roll-off and the tweeter would need to have the polarity reversed to avoid a null at the crossover. Of course, none of this means that it would sum correctly. Just something to think about. Now, if you are using driver measurements and decent simulation software then you will be able to work these things out.
              Dave H

              Comment


              • #8
                Re: Uncharted Crossover Design Waters

                Because they can and people buy them without educating themselves enough about what they are buying first. Just like people buying B*#e speakers and thinking they are the best.

                Sorry its probably a stupid answer!

                Comment


                • #9
                  Re: Uncharted Crossover Design Waters

                  Lord knows......

                  As for that last question, I think you'll answer that yourself once you start experimenting a little.
                  Click here for Jeff Bagby's Loudspeaker Design Software

                  Comment


                  • #10
                    Re: Uncharted Crossover Design Waters

                    Originally posted by Hackomatic View Post
                    If every network must be custom built and tweaked according to a driver's frequency curve, impedance, phase, etc., it begs the question . . how can Dayton, Eminence, and so many others sell all those prefab, garden variety crossovers?
                    Because garden-variety crossovers aren't THAT bad, comapred to cheap budget consumer garbage with just a 10uF NPE on the "midrange", a 3.3uF on the "tweeter" and nothing else. For instance, if you used a 2 KHz texbook butterworth X/O with a garden variety soft cone woofer and dome tweet with 1k resonance, and bothered to match levels, the result would likely be listenable. And you'd probably end up finding that it sounds better with the drivers in phase, instead of reversed. Why? It might not be too far off a 4th order LR. Certainly within tweaking distance - and with careful choice of drivers you could actually hit that target. That technique does limit you to rather mediocre drivers and takes some fooling with - but where you might have only $30 per side invested it may very well be worth the exercise.

                    And if you're working with PA drivers or, the response errors in the drivers may be far worse than what the "crossover" is doing to the overall sound - and at least it's putting proper filtering (for safety) on the drivers. Those systems are usally beat into submission with EQ anyway.

                    Comment


                    • #11
                      Re: Uncharted Crossover Design Waters

                      Originally posted by daryl View Post
                      Even the infamous non-polar electrolytic capacitor performs far beyond the capability of the best drivers.

                      And if you bother to take ESR into account in the models, the effect on Q in the networks can be exactly modeled. You do it with coils, and it works the same way with caps. This is the biggest mistake people make when using electro's IMO. Of course it's going to sound better with a lower-ESR cap when the network was designed and optimized for a cap with zero ESR. That's also why you don't seem to "need" anthing other than an NPE for a zobel (or circuit which looks like one). There's already a 4-10 ohm resistor there, what's a few tenths more?

                      The biggest beef I have with NPEs is all the little pieces of paper inside the cab when it goes kabang. With DFs of 10 to 20%, the ESR can be high and they will get hot. For speakers that will be subject to 100 watts or more average power, it's a good idea to use caps designed for switch mode power supply filtering if you need a big electrolytic. Need to make an NP out of two back to back. Those caps have low DF and can take several amperes of ripple current at 100KHz or more without overheating.

                      Comment

                      Working...
                      X