No announcement yet.

SAM16 aka The Double D's

  • Filter
  • Time
  • Show
Clear All
new posts

  • SAM16 aka The Double D's

    I am presenting here a design I built a year ago called the SAM16, which stands for Synthetic Aperture Midrange, with 16 midranges. I used to work in the field of Synthetic Aperture Radar (SAR) a few decades ago, where a large radar antenna was synthesized by moving a small antenna along a path. This speaker synthesizes a large midrange from 16 small ones. It can also be thought of as a cross between Coax, MTM, and Line array topologies. I also call them the "Double D's". Not those double D's, but in honor of Mr. D'Appolito, since it is a form of MTM in both the vertical and horizonal directions.

    The center tweeter is an AMTPRO4, with foam wedges cut from an acoustic tile in an effort to limit it's horizontal dispersion. The mids are Peerless TC9FD18-08, for which I have to thank Zaph for pointing out their qualities. They have a vertical spacing of 3.7" c-c, which corresponds to 1/2 wavelength at 1830 Hz. The total Sd of the mids is 580 cm^2, which is larger than a typical 12" driver.

    When I designed this, I was planning on an XO of about 1600 Hz, but after listening I have settled on 2.0 kHz, LR4. The AMTPRO4 sounds better when crossed this high, IMHO. This mid-high box sits on top of a woofer box, which is currently a narrower box containing two Focal 7K415, which is a carry-over from my previous system. I have never gotten around to building a matching woofer base cabinet with larger woofers. The mid-high box is approx 21" W x 31" H, and made from baltic birch plywood. It has 3" radius corners, and is trapezoidal shape when viewed from the top. Internally there are 5 compartments, 4 for the mids and one for the AMT. I had intended to cover it with veneer but never got there (my wife's health has necessitated me spending much time caring for her in the last year). Lower XO is at 330 Hz, LR4. I use a digital processor for Crossover and EQ.

    The motivation is obviously to achieve a limited dispersion, especially in the vertical direction to minimize ceiling bounce over a wide bandwidth. The curvature of the lines allows the 4" wide tweeter to be placed between them, while keeping the average horizontal spacing less so that the horizontal 6dB beamwidth stays 'wide enough'. Unfortunately, with the higher XO frequency than originally planned, the horizontal beamwidth is less than I would have liked. If I ever build version 2, I will try to achieve tighter horizontal driver positioning. The mids that are placed beside the tweeter are to maintain the vertical line array behavior. Without them it would behave more like an MTM, with reduced uniformity of the vertical beamwidth over the bandwidth (see the MTM Disadvantages thread A secondary advantage of the curved lines is that it places the mids at varying distances from the edge of the baffle and the interior dividers, helping to smooth the response.

    This is getting long, so I will end here, I plan to post some polar plots later. Please give me your comments.
    "She don't love my speakers anymore..."

  • #2
    I'm very much looking forward to your results- it's an interesting array shape and if my mental model is working properly (it's morning, I've had coffee, so hopefully?) this could be an excellent setup. I've always liked the idea of arrayed drivers but haven't traditionally liked... traditional line arrays.


    • #3
      Surprisingly, I've seen a similar speaker, with Aura whispers surrounding a big AMT. I think it was a german guy on DiyAudio. I have a bunch of the tiny BMR drivers and was thinking I could use 6 or 4 of them surrounding a small tweeter to create a quasi coaxial speaker, and by elongating the array into an oval, create a midrange which matches the vertical directivity of an MTM woofer array.

      I'd love to see measurements on your work. Really interesting stuff - these driver arrays as a means of manipulating directivity are a very approachable technology for diyers.


      • #4
        As promised, I will try to post the polar maps as measured by my OmniMic. First is the vertical, followed by the horizontal. I think the vertical is pretty good down to 500 Hz, as the lower XO is 330 Hz to the two woofers. The horizontal is somewhat disappointing, as the AMT really broadens out above it's 2kHz XO. The foam waveguides are not as effective as I hoped at the lower end of it's band. This is definately a tight dispersion speaker in the mids and upper highs, but is very enjoyable to listen to in the sweet spot, with fantastic imaging and detail. I want to listen to the information on the recording, not the sound of my room!
        "She don't love my speakers anymore..."


        • #5
          Perhaps the tweeter isn't the right call for your design goal. This horn might suit your tight pattern better
          There are other horns with that pattern, I haven't used the PRV but the Selenium 17-25 is the same basic design and is a very fine sounding laserbeam horn.


          • #6
            I have been discussing horn loaded versus direct radiating designs on another thread, so I decided to update this thread too. I recently increased the upper xo frequency on this system to 2.7 kHz, from 2.0. I also moved it out a little further from the rear wall, then re-eq'd it with omnimic. I think it is improved this way, because of three effects. First, the widening of the dispersion between 2 and 4 kHz you seen on the polar maps here has been reduced. But I haven't had time to measure it again. Second, the harmonic distortion (and presumably IMD) of the AMTPRO4 really drops above 3 kHz. Lastly, the direct to reflected ratio at the listening position has been improved, smoothing the lower mids and upper bass. I am really happy with the sound. More recordings sound good, and good recordings sound great. More like headphones, without the 'in your head' effect, rather, with a great sound stage.
            "She don't love my speakers anymore..."