Announcement

Collapse
No announcement yet.

Another try at crossover; 2-way Nd105-4 with Peerless DX20BF00-04 3/4"

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

  • Another try at crossover; 2-way Nd105-4 with Peerless DX20BF00-04 3/4"

    So after my last attempt, which failed because of buying stuff before I asked for advice and then failed even more because parts where not delivered. I'd like to try again and check with the forum first. I already made sure all the parts are available.

    I know there are lot's of proven designs out there. But I'd like to make a pair of speakers for my office/man cave myself. These will need to fit next to my computer screen. I'd like to stay under 10Liters.

    That's why I chose the ND105-4. I know them from my Nano Neo's and am still very impressed with what they can do in such a small enclosure.

    I tried different tweeters (under 25,-) in the simulation, but only with the Peerless DX20BF00-04 3/4" I could keep the simulations frequency response between 87 and 83 db.

    Now my question, is my crossover simulation design going to work in the real world?


    crossover Impendance Frequency Simulation

  • #2
    Some observations and ?s

    Does the 1.5ohm impedance drop (near 2kHz) bother you, 'cause it bothers me.
    In a sim (using MY f/z files), my impedance profile matches yours, but not the FR profile, at ALL.

    Maybe you could attach your 4 F/Z files here?
    (I also realize that a lot of the world (you) use commas in numbers the way Americans use periods (for decimals).

    Hi pass on tweeter:
    The 0.082mH (82 micro-henry) coil in-line w/the tweeter rolls the top end down approx. -6dB. Why would you want that?
    Also, you show 1.7n(ohms) of series resistance next to the coil. I'm figuring a few ohms DCR for the small coil, plus a 1.5ohm series resistor in front of it.
    Is that your intent?
    The 1uF shunt cap does next to nothing.
    You don't show any DCR for the 1mH shunt coil (prob'ly around .7ohms or so?).

    Woofer LP filter:
    Your combo of a 0.27mH (270 micro-henries) series coil and a 33uF shunt cap pull your impedance very low, creating a huge response peak near 2k.
    Your sim doesn't reflect that at all (wondering if there's some DECIMAL/COMMA problem (wrong value magnitudes?) going on here).

    Oh... I also don't see any baffle-step compensation?

    Comment


    • #3
      The FRD and ZMA from the ND105-4 I downloaded from Dayton's website. The tweeter data is traced from the Spec PDF. I added it as an attachment.

      I also add screenshots of the FR, with the parts you mentioned removed. I was trying to address the peak at 23kHz. Not because I can hear it, but just to get the overall FR as flat as possible.

      About the impedance, I double checked, there is no DECIMAL/COMMA problem. It does bother me yes, I thought it looked weird. but at average it looked ok to me and my limited experience/knowledge.

      I have been fiddling for a few weeks with different values and components, and completely forgot about BSC... :face palm smiley:

      I'll just start from scratch. would you mind checking my ZMA & FRD file? I'm thinking there might be a problem with those which would explain the impendance.
      Attached Files

      Comment


      • #4
        Did you apply any modifications to the files, or did you model the xover straight from those?
        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:
        http://photobucket.com/Wolf-Speakers_and_more

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

        Comment


        • #5
          I know you used Dayton's F/Z data for the woofer, but please post those as well.
          Can you show me JUST the ND105-4 w/your 2-element low pass filter? (just a 0.27mH coil - 0.3ohm DCR, right? and a 33uF shunt coil).
          When I sim that, I get a (bad) hi-Q peak at 2000Hz, that's about +8dB higher than the unfiltered response.
          (This is causing your Z to drop so low, which YOUR sim shows, so it SEEMs like something's wrong w/your woofer.frd file.)

          (your tweeter files look fine)

          Comment


          • #6
            I didn't modify any files. Attached are the ND105-4 files.
            Attached Files

            Comment


            • #7
              Chris, I tried your suggestion. When I just sim the ND105, it has a peak at 2000Hz. I guess I'll start with a completely new sim.
              Attached Files

              Comment


              • #8
                If you didn't modify the stock files to adapt to your box/baffle design, then you missed a step. You cannot just model a xover with IEC/infinite baffle measurements. You will be way off.

                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:
                http://photobucket.com/Wolf-Speakers_and_more

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

                Comment


                • #9
                  https://sites.google.com/site/undefi...d-measurements

                  Paul has a great step by step outline to incorporate the box design, baffle step losses, and diffraction into the .frd files so that you can really get the simulation closer to real world application.

                  Good luck!

                  Comment


                  • #10
                    I don't even have to look at your files, 'cause I saw that exact peak at 2k. There's virtually NO way (in your orig. sim) that THOSE files could have given you a summed response (both drivers) that looked as flat as yours did. MAYbe your schem. showed that coil and cap value, but something else actually got plugged into the simulation??

                    When using Dayton's files (which I THINK have "measured phase" - not sure), and the one's you traced (which all have a phase value of ZERO most likely), you MUST generate "minimum phase" files AND plug in 3-D driver offsets from each other. the DEPTH ("Z") axis dimension will mess with the phase angle the most. You also SHOULD add in baffle-step and diffraction effects (as Wolf said). It would be GREAT if we could just plug in some files from PE, and you CAN, but it won't be very accurate. You should be able to get a good 2-way using these drivers.

                    ALSO, don't forget to enter the series resistance (DCR) of coils you use. This (extra) resistance WILL affect a driver's output (diminishing it) in all series positions in your filters.

                    Comment


                    • #11
                      As it so happens I'm currently also building a 2-way speaker with the ND105-4, and had some trouble with the crossover in the beginning as well. I eventually came up with something that seemed good enhough for me, however I have not tested it yet and it is with a different tweeter. The tweeter I am using is the Dayton Audio ND28F-6

                      For the ND105 I used a second order filter with a notch to dampen the peak that the second order filter creates. It made the crossover a little more expansive than I wanted but I hope it will be worth it.
                      The frequency response of the entire speaker seems to not take baffle step into account but I will mostly be listening to the speaker at an angle which reduces the higher frequency quite a bit with this tweeter.

                      hope this helps you in some way
                      Attached Files

                      Comment


                      • #12
                        Thank you all for your responses! I've got some homework to do ...

                        Originally posted by scottvalentin View Post
                        https://sites.google.com/site/undefi...d-measurements

                        Paul has a great step by step outline to incorporate the box design, baffle step losses, and diffraction into the .frd files so that you can really get the simulation closer to real world application.

                        Good luck!
                        So today I had a go at this, but the file keeps crashing everytime I try to "auto extract phase from FRD file"

                        I'll try again later this week on another computer with a different version of excel.

                        Comment


                        • #13
                          Reiney, your filter shouldn't CAUSE a peak - it it does, you should be able to hit the same Fc using different cap/coil values. Wouldn't recommend a "notch" to fix a filter issue.
                          Post it if you'd like me to take a look at it.

                          Comment


                          • #14
                            I tried lot's of things, even freshly installed a new windows and office, but I cannot get this Phase extraction to work. I'll guess I just pick one of the many other proven designs that are out here and save myself from another headache or two.

                            Comment


                            • #15
                              Phase can take care of itself, if you can measure, and tweak. Within reason, if you match target slopes, you can get close. This sim was done using the active section of PCD to simulate a baffle step. It's roughly in the ballpark. I turned the active filter back off to post this response. With the active filter on, the response at 100hz drops down to match the target.Try these values, and layout with your program, and see if you get similar response.

                              Oh.. just remembered. I used the 15 degree frd. The on axis frd has issues.

                              Click image for larger version  Name:	nd105-4 filter.jpg Views:	1 Size:	393.2 KB ID:	1392396

                              Comment

                              Working...
                              X