Announcement

Collapse
No announcement yet.

Advice on designing a spherical speaker

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

  • Advice on designing a spherical speaker

    I'm toying with the idea of making a spherical satellite speaker with a TEBM35C10-4 driver and ND65-PR 2-1/2" passive radiator. I would 3D print the enclosure with double walls and then fill the cavity with resin or plaster. My intent was to take the tuning for this Tectonic reference design:
    http://www.tectonicaudiolabs.com/wp-...SG_12OCT18.pdf
    I can match the volume exactly, but the N65-PR 2-1/2" is definitely smaller than the radiator in Tectonic design. I've put together this draft CAD model.
    Any comments or advice on the choice of drivers and the enclosure design would be very welcome.


    Luke

  • #2
    You can't just mix and match passive radiators. You have to do the math. I have never done it, but some of the sims tell you how. Might make a nifty little desktop. I for sure would print feet on it of some type. Sounds like a fun project.

    Comment


    • #3
      My plan was to tune it to the same frequency by using added mass and measuring the impedance curve, or a near field SPL sweep on the radiator. I believe either of those will work, assuming the starting mass of the radiator is low enough. I chose this one because it claims 3.2g Mms, which is lower than many of the others of comparable size.

      Comment


      • #4
        It seems like you can get close enough using the 0.6L box (.02cf) and that PR.
        The box tuning is a little low (near 100Hz, instead of the 110-115 target) and the F3 is slightly higher (120-125, instead of 115-120).
        (You'd need a pair of those PRs to tune higher, so ... )

        Oddly enough, a 5/8"id x 3" port (w/a generous flair - about 15mm x 75mm) tunes that volume to 110Hz, yielding an F3 near 90.
        Increasing the volume by 50% (0.03cf / 0.9L) tunes the box to 90Hz (w/that same port), w/an F3 below 80 !

        Depending on the Xmax (2.5mm from PE's specs?, or 4mm off your ref.PDF?) it will be limited to 2wRMS or 5wRMS.
        The smaller box (0.6L) can take about twice that power before Xmax (but you only gain +1-2dB).

        Comment


        • #5
          Do you think a cavity, inherently strong by shape, really needs a double wall? I wonder, if you did the shell, then just sloshed a couple layers of resin, maybe even fiber glass reinforced on the inside it would give a little more volume. Quicker print time? A big foot?

          Comment


          • #6
            I agree it needs some kind of foot or supports. Three legs, or a poppet-valve sort of foot are what I'd consider.
            Francis

            Comment


            • #7
              Originally posted by Chris Roemer View Post
              It seems like you can get close enough using the 0.6L box (.02cf) and that PR.
              The box tuning is a little low (near 100Hz, instead of the 110-115 target) and the F3 is slightly higher (120-125, instead of 115-120).
              (You'd need a pair of those PRs to tune higher, so ... )

              Oddly enough, a 5/8"id x 3" port (w/a generous flair - about 15mm x 75mm) tunes that volume to 110Hz, yielding an F3 near 90.
              Increasing the volume by 50% (0.03cf / 0.9L) tunes the box to 90Hz (w/that same port), w/an F3 below 80 !

              Depending on the Xmax (2.5mm from PE's specs?, or 4mm off your ref.PDF?) it will be limited to 2wRMS or 5wRMS.
              The smaller box (0.6L) can take about twice that power before Xmax (but you only gain +1-2dB).
              Chris,
              Thanks for the calculations, that's really useful. A port is totally on the cards. For prototyping, I can just print one with an interface to screw in where the PR is, and try both.

              I'm planning to pair these with a Voxel, so I think I can get away with a cross over ~150Hz. It shouldn't be getting much input signal near the tuning frequency, is the xmax so much of a problem? Would I actually be better off reducing the enclosure volume? I don't know whether that would give me more power handling or less because I would be bringing the tuning frequency up closer to the crossover. These are the kind of things I have very little idea about.

              Out of interest, are you using WinISD, or something else. I am meaning to get into some modeling, but right now I wouldn't be able to make the right decisions based on the results. :-D

              edit: A larger volume is easily printable too. I don't think I'd want to go beyond about 1000cc to keep them small enough for desktop use. They are about 13cm outer diameter at the moment, inner diameter is 10.46cm


              Luke

              Comment


              • #8
                Originally posted by tvrgeek View Post
                Do you think a cavity, inherently strong by shape, really needs a double wall? I wonder, if you did the shell, then just sloshed a couple layers of resin, maybe even fiber glass reinforced on the inside it would give a little more volume. Quicker print time? A big foot?
                The walls are probably overkill, but the main housing is only a 10h print at 0.2mm layer height, and about $5 worth of plastic. Pour filling the double skin is so much easier than coating the inside, it seems like the obvious choice. I'd probably go to 0.1mm layer height for the finished unit, but it is just waiting time, not touch time.

                Luke


                Comment


                • #9
                  Originally posted by fpitas View Post
                  I agree it needs some kind of foot or supports. Three legs, or a poppet-valve sort of foot are what I'd consider.
                  I'm not sure if I prefer three splayed legs in a tripod configuration, or a single mounting point for a broad stand, which could also be used for wall mounting. These are simple tweaks to make, so I'll probably publish both.

                  Comment


                  • #10
                    Just for argument's sake, this is a 3" port, entirely outside on the back.

                    Comment


                    • #11
                      If you're gonna "Voxel" it, it's funny but the vented .02cf box can take more power (and down to about 100Hz) than the 0.02cf closed box.
                      WinISD is your friend (and can do PRs w/the data PE specs).

                      Comment


                      • #12
                        Originally posted by zx82net View Post
                        Just for argument's sake, this is a 3" port, entirely outside on the back.
                        High Bypass Turbofan design!
                        Francis

                        Comment


                        • #13
                          Originally posted by Chris Roemer View Post
                          If you're gonna "Voxel" it, it's funny but the vented .02cf box can take more power (and down to about 100Hz) than the 0.02cf closed box.
                          WinISD is your friend (and can do PRs w/the data PE specs).
                          Interesting. That's with your 110Hz tuning?

                          Comment


                          • #14
                            Originally posted by fpitas View Post

                            High Bypass Turbofan design!
                            Yeah, I quite like it. It wouldn't be so great for a wall mount, but its fun for a desktop.

                            I could try a concentric design and make it a whole load smoother than the prototype shown here:

                            Resource for designing and building concentric ports for subwoofers and speakers. Includes calculator for making composite tubes of any size

                            Comment


                            • #15
                              Originally posted by zx82net View Post

                              Yeah, I quite like it. It wouldn't be so great for a wall mount, but its fun for a desktop.

                              I could try a concentric design and make it a whole load smoother than the prototype shown here:
                              Looking like that, it must have fast bass!
                              Francis

                              Comment

                              Working...
                              X