Announcement

Collapse
No announcement yet.

6Db. crossovers????

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

  • 6Db. crossovers????

    Am I wrong or are they seldom used? What are the advantages and disadvantages? Specifically thinking about a nice woofer with a slow and low rolloff to a nice fullrange and maybe a tweeter crossed in really high. And is my thinking right that if a tweeter can be crossed safely at 2000Hz. 12 Db. then it would be protected enough crossed at 4000 6 Db.?

  • #2
    Re: 6Db. crossovers????

    You'd be surprised at how many commercial high end speakers use very simple 1st order xo's. The key is to find the right drivers (usually expensive) that will perform well under those conditions.
    Another key consideration is the power demands you plan to place on your design. If it's pretty constant 95+ dB levels, rock with lots of bass, then I'd suggest you shy away from 1st order.
    Live in Southern N.E.? check out the CT Audio Society web site.

    Comment


    • #3
      Re: 6Db. crossovers????

      in the world of car audio and two way ceiling speakers it is very common to run the "woofer" as full range and put a simple blocking cap on the tweeter. This more or less eliminates the crossover altogether since the cap can just be mounted directly to the tweeter. Usually in the 3 to 5uF range. The concern in doing so is that many midranges/woofer have a bad hump somewhere in the 3500-5500 range.

      Comment


      • #4
        Re: 6Db. crossovers????

        1st order crossovers can work well if the drivers are very well suited to them. But the vast majority are not. Protection issues aside the wide bandwidth overlap that is the usual result of 1st order filtering can lead to all sorts of phase and response issues. The higher 'purity' that results from the lower phase shift introduced by a 1st order filter often results in only trading one problem for another.
        in the world of car audio and two way ceiling speakers it is very common to run the "woofer" as full range and put a simple blocking cap on the tweeter.
        That's not just a car and ceiling speaker scenario, and the reason for so doing isn't design, it's cost.
        www.billfitzmaurice.com
        www.billfitzmaurice.info/forum

        Comment


        • #5
          Re: 6Db. crossovers????

          Originally posted by billfitzmaurice View Post
          The higher 'purity' that results from the lower phase shift introduced by a 1st order filter often results in only trading one problem for another.
          +1

          Where 1st order crossovers are used in high end "cost is no object" speakers the 1st order slope in the transition region is almost always augmented by an additional sharper cutoff somewhere an octave or two out in the stop band (often implemented with one or more notch filters), and the resulting "1st order crossovers" are a whole lot more complex (and expensive) than "a cap and an inductor" would suggest. In low end cost driven designs the tendency is to "augment" the slope by choosing drivers with high intrinsic losses in the stop band, often using soft cone drivers that decouple as frequency rises (a "polite" way of describing damped cone breakup) and the natural low frequency cuttoff of horn tweeters or the natural "box rolloff" of small chambered domes . . . both of which give the "1st order" crossover much steeper terminal slopes.
          "It suggests that there is something that is happening in the real system that is not quite captured in the models."

          Comment


          • #6
            Re: 6Db. crossovers????

            Originally posted by Dwight View Post
            Am I wrong or are they seldom used?
            They are used a good bit but not correctly. Most speakers claiming 1st order are generally closer to 2nd or 3rd order acoustic.

            What are the advantages and disadvantages?
            Advantages
            -more overlap could probably give better driver integration
            -phase response "looks" nicer at a spot in space, which is audible if you listen in anechoic chambers with your head in a vice.

            Disadvantages

            - VERY High crossover points are needed to give decent power handling, which ****s with off-axis response
            - shallow rolloff means crappy
            - more overlap means ever-crappy vertical polars
            - odd order crossovers create vertical off-axis peaking whereever -3db + -3db = +3db
            - drivers sum in phase quadrature, so if the desired transfer functions are off-the-mark, you might get phaseyness / sub-par imaging
            - shallow rolloff all but ensures non-pistonic behaviour in passband, unless you've got a ESL panel which have their own resonances of course.
            - drivers have their own acoustic rolloffs from things like mass, inductance, etc. that are normally 1st or 2nd order already.

            And is my thinking right that if a tweeter can be crossed safely at 2000Hz. 12 Db. then it would be protected enough crossed at 4000 6 Db.?
            Let's say tweeter X rolls off 24db/octave @ 2khz. That means it's probably 6db down at 2khz, and 30db down at 1hz.
            Let's say tweeter X rolls off 6db/octave @ 4khz. That means it's probably 3db down at 4khz, 9db down at 2khz, 15db down at 1khz, 21db down at 500hz.

            Let's say you need your speaker to produce 95db @ 3m in the midrange. That means just for starters your tweeter needs to be

            1) capable of almost ~87db @ ~750hz cleanly ( essentially > 0.3 cm^3 of displacement) - So you need to start with a very robust tweeter. That eliminates about 85% of options.
            2) virtually flat to below ~750hz (otherwise you can't call it 6db/oct for most of the passband, and still not all)

            ....Now in this scenario you've only dealt with power handling.

            How about power response?
            To get decent power response for a 4khz 1st order crossover you probably want a very small midrange IE 2"
            ...and guess what...now you need a 4-way.
            And with all the driver overlap in a 4-way with shallow slopes, you're going to have VERY random bandpass gain. Good luck getting flat response.
            :blues: Flat frequency response, a smooth sound power response free of resonance, careful driver-integration, and high dynamic range both upward and downward :blues:

            Comment


            • #7
              Re: 6Db. crossovers????

              Originally posted by Deward Hastings View Post
              +1

              Where 1st order crossovers are used in high end "cost is no object" speakers the 1st order slope in the transition region is almost always augmented by an additional sharper cutoff somewhere an octave or two out in the stop band (often implemented with one or more notch filters), and the resulting "1st order crossovers" are a whole lot more complex (and expensive) than "a cap and an inductor" would suggest. In low end cost driven designs the tendency is to "augment" the slope by choosing drivers with high intrinsic losses in the stop band, often using soft cone drivers that decouple as frequency rises (a "polite" way of describing damped cone breakup) and the natural low frequency cuttoff of horn tweeters or the natural "box rolloff" of small chambered domes . . . both of which give the "1st order" crossover much steeper terminal slopes.

              You mean there are "high end" designs using first order filters above 2Kz?
              why?

              the only thing that springs to my mind is car audio, computer speakers, and background music. Maybe if you wanted to use a fullrange without crossover and just wanted an easy way to throw in a little mylar dome for the ultra-high.

              IDK. I'm having a hard time seeing the reason for high frequency blocking caps in high end. Why would they do it?

              Comment


              • #8
                Re: 6Db. crossovers????

                Originally posted by moron#99 View Post
                You mean there are "high end" designs using first order filters above 2Kz?
                Yes.
                "It suggests that there is something that is happening in the real system that is not quite captured in the models."

                Comment


                • #9
                  Re: 6Db. crossovers????

                  Originally posted by Deward Hastings View Post
                  Yes.
                  I have never read any good treatment of phase distortion that acknowledges the possibility of differential perception. As such, the safest assumption in design is to assume that it does exist and does make a difference. I would say 6" of image displacement is acceptable in a living room speaker. The worst case scenario is 180 degrees of phase shift.

                  wavelength = speed / frequency
                  0.5ft = 1125 / F
                  F - 2250Hz

                  I just can't see any reasonable argument in favor of 6db blocking cap except cost, simplicity, and availability. From a quality of sound perspective - aka high end audiophile - i'm just not getting it. why would they do it? what is their sales pitch?

                  Comment


                  • #10
                    Re: 6Db. crossovers????

                    Originally posted by moron#99 View Post
                    why would they do it? what is their sales pitch?
                    The usual justification is some form of "transient perfect" . . .
                    "It suggests that there is something that is happening in the real system that is not quite captured in the models."

                    Comment


                    • #11
                      Re: 6Db. crossovers????

                      hmm, well that's either BS marketing or the engineers don't actually know how to set about correcting impulse response. you don't fix impulse response with a tweeter. that's like bailing out the titanic with a tablespoon.

                      Comment


                      • #12
                        Re: 6Db. crossovers????

                        I like them, but I haven't had enough experience with high order crossovers to give a fair comparison.

                        A good design and writeup are Markk's ER18DXT:

                        http://www.audioheuristics.org/proje...XT/ER18DXT.htm

                        Dynaudio designs often use 6 db (electrical) crossovers. Don't be confused by the caps and coils around the tweeter, they are an all-pass network (aka ladder delay). I built the focus and the crossover worked well. (I was not as happy with the woofer.)

                        http://www.gattiweb.com/dynaudio.html

                        Comment


                        • #13
                          Re: 6Db. crossovers????

                          Originally posted by moron#99 View Post
                          hmm, well that's either BS marketing or the engineers don't actually know how
                          Some very good speaker designers have produced some very good speakers using that approach . . . so "don't actually know how" might be a bit of a stretch. While I'd be inclined to agree that pictures of square waves are little more than marketing gimicks I'd also suggest that there is a lot to be learned from studying rather than simply dismising out of hand the better 1st order crossover designs. Many of the tricks necessary to make them work well are beneficial (and often ignored) when designing other crossover topologies.
                          "It suggests that there is something that is happening in the real system that is not quite captured in the models."

                          Comment


                          • #14
                            Re: 6Db. crossovers????

                            Originally posted by Deward Hastings View Post
                            Some very good speaker designers have produced some very good speakers using that approach . . . so "don't actually know how" might be a bit of a stretch. While I'd be inclined to agree that pictures of square waves are little more than marketing gimicks I'd also suggest that there is a lot to be learned from studying rather than simply dismising out of hand the better 1st order crossover designs. Many of the tricks necessary to make them work well are beneficial (and often ignored) when designing other crossover topologies.
                            every time you throw out a tablespoon of water the titanic is a little bit less sunken than before. I would suggest that you will never make the titanic float with a tablespoon.

                            A 90 degree phase error at 250Hz is 13 inches long and lasts for 1.0ms
                            A 90 degree phase shift at 2500Hz is 1.3inches long and last 0.1ms

                            IMO, until you fix the problem at 250Hz then the one at 2500 isn't all that important. If these designers are touting their work with a tweeter for improved impulse response then it means they have improved the lower frequencies by accident rather than engineering.

                            Comment


                            • #15
                              Re: 6Db. crossovers????

                              Originally posted by moron#99 View Post
                              IMO . . . accident rather than engineering.
                              Good for you that you have it all figured out and know all there is to know. It will make speaker design almost boringly simple . . .
                              "It suggests that there is something that is happening in the real system that is not quite captured in the models."

                              Comment

                              Working...
                              X