May 2014

Quick and Dirty DIY Dipole Surround Speakers
Article By Jeff Poth

  It has been a few months, dear readers, as my career has taken a much appreciated uptick plus family life is demanding and excellent. While this doesn't stop me from building some things, I'm very limited in my time and writing for Enjoy the Music.com has paid the price. I have a lot of projects to write up, but these have a time sensitive component -- they're based upon a driver/waveguide combination that's on clearance and has limited supply at this link.

Yep, the cheapest drivers I've ever worked with. They are actually rather cool- better frames than typical for TV speakers, a copper cap on the pole helping with both power handling and distortion, and a solid composition dustcap. The guide is also rather well-designed. The little triangular protrusions to the sides are damped acoustic chambers, and the slot is divided with braces that help prevent standing waves and other internal issues. The slot actually has a quite broadband output from about 80 Hz to 10 kHz, with some lower but still meaningful output all the way to 20 kHz.

I had some binding posts and glue to order from Parts Express, and to get their free shipping (on orders >$98 at time of writing) as well as take advantage of a $10 off $100 rather than $5 off $50 promotion, I ordered some of these cool looking drivers to fill out my order. They effectively paid for themselves. A few days later, I had a couple big boxes and a lot of these little drivers. I started playing around with them to see how to mount them. First I pictured using some mounting points on them to bind a dipole array together with dowel rods for use as surrounds. Before I got too far down this path, however, I decided to try simply gluing them to a wooden frame. As it turned out, the Polyethylene (which is tough to bond) responded well to hot-melt glue, and was easily affixed accordingly.

Given their low cost, I decided to toss something together based upon scrap. I had some leftover rectangles of pine from the disassembly of cabinets on my passive radiator subwoofer project and so glued a quad together on one of the frames, and leaned it up against the wall to simulate an on-wall mounted dipole surround. Some quick tunes run through confirmed that the drivers actually sounded good in this arrangement. In fact, they even sounded solid without the support of the large plane presented by the wall. Not up to what I demand of mains (duh) but filled a lot of space, and didn't have notable sonic warts, when listened to along the axis of the slots. The frame, at that time, was open, and I noticed that the sound emanating from the "front" of the enclosure was not nearly up to the output from the slots. Not only did it give up much of the bass and midrange energy (they're wired as dipoles, so there was cancellation at lower frequencies) there were some harsh resonant sounds.

Fortunately this issue didn't persist on the axis firing along the wall, and thus could be dealt with by enclosing the front of the system and leaving the slots as the primary sound sources. Since this was intended to be a dipole surround system, this was the half-baked plan anyway. I started by finishing the driver mounting for two systems, each with two drivers firing to each side. They were then wired as dipoles- meaning that while the ones facing to one side push, the others pull, and vice versa. I added some carpet underlayment (comprised of a variety of scrap textiles, bonded together). This is a pretty effective acoustic treatment, intended to absorb the highs inside the frame. At this point it's looking pretty ugly.

Good thing that your scrounger in chief had a plan for that, too. I work at a company that sometimes has events requiring signs for various purposes. After they've served that purpose they often find their way into a corner or somebody's cubicle, until such time as they can be disposed of with examples of each saved for posterity. The material utilized for these is known as foamcore board. Some DIYers may have heard of this material, there are people doing some great work with it over on DIY Audio. The variant I have been collecting over the past few years is actually a relatively high-end version, with plastic facer layers on either end, almost like a 0.25" version of 'Gatorboard'. It has some very desirable properties, in that it's well damped via the foam polystyrene core, and very rigid for its weight. The glossy backside of them is quite attractive too- bonus!

I've been gathering this since 2009, and have made a turntable platform for my Linn LP12 from it, with which I'm very happy. This was a perfect opportunity to burn through a little bit more of my supply, staying with the theme of very inexpensive or found materials. I decided to improve the performance of the system by applying roundovers to the side of the slot that's not up against the wall surface. This will not only add a little more directionality to the system, keeping it firing to the sides and not the front, but also dramatically reduce diffraction, eliminating the location cues that those early reflections cause. To do this, I had to remove sections of the facing, after which it could be curved into a 3" diameter bullnose.

Basically, the slots are aligned with their long dimension perpendicular to the floor, with one side of the slot butted up against the wall and the other along the rounded bullnose. The foamcore was affixed to the edge of the wooden frame, again with hot melt glue. To minimize any panel resonance from the rather large front of the speaker, two layers were used in a constrained layer damping "sandwich" using cheap weather stripping between them. They were fastened together to the wooden frame using hex-head screws with bonded washers. This pulls the two panels together enough to compress the weather stripping and prevent resonance.

Bonded washers are a rubber washer glued to a metal washer, which allowed me a nice large, soft clamping area from each of the six screw locations. The hexhead screws allow you to use a magnetic driver bit, that makes them extremely easy to use- I didn't even have to drill pilots. Care was used to ensure that I didn't tighten them down too much and destroy the foamcore either by torsional force at the surface or overly compress the foamcore. There's damping (again, weather stripping) between the drivers and the front panel, as well as between the polyethylene waveguides and the wall. I left the gaps in place- there's enough absorption, cancellation and whatnot that the soundfield is comprised nearly entirely of output from the slots. Mounting of these 10 pound (maybe 10.5 lbs, I weighed them before the front panel was added) speakers was exceedingly easy due to their light weight- some picture wire was knotted and soldered for durability, then mounted with a small picture hanging hook. By mounting this way, the frame is balanced flat against the wall, damping the driver assemblies via the weather stripping.

The combination of the directivity imposed by the slot and roundover "waveguide" and the dipole cancellation means a minimum of direct sound arriving at the listener, and two rather well defined lobes developing along the wall. After running some 4x16 AWG surround wire up through the ceiling and getting nice and itchy with attic dust and fiberglass, the installation was complete. They are on the rear wall, but would perform nicely in any on-wall location, preferably near the ceiling reasonably far from the listening position.

Sound is what I'd hoped- they are very diffuse, providing a very impressive level of spaciousness like that of multiple pairs of surrounds. They disappear very effectively, and are a very welcome addition to the bedroom system. I'd previously only made monopole surrounds, and never achieved the level of performance from the surrounds as I have with these. The whole of the assembly could easily be completed in a single day of work, though I spread it over a bunch of sessions. These sessions included a lot of messing around to figure out the best solutions to the various challenges associated in using salvaged materials, etc. There is expected comb filtering and other behaviors that I wouldn't like in a "main" speaker, but because the sound is dominated by reflections, these deviations average out and the overall lobe has a big midbass from 70 to 300 Hz or so, about 3dB above the 300 to 10 kHz band which is surprisingly flat (+/-2dB), and then down about 6dB from the nominal 90dB sensitivity per lobe achieved in that midband by 20 kHz. These are perfectly acceptable numbers for a surround, but one could do some more refinement and filtering if so desired. Some of that oversized midbass is pulled out by using the surrounds set as "Small", but the tonal balance is definitely quite warm.

Impedance load is very easy, basically that of a single driver, as they're wired series-parallel for an 8 Ohm load. It does reflect a little ripple from the waveguide loading, but it's pretty minor compared to the vast majority of devices out there, and many unloaded drivers have as much or more ripple. Note the very tight scale- this is a smooth plot with the exception of the one ripple at about 350 Hz.

Foamcore board with paper facers is widely available at hobby shops and even some dollar stores, and is a pleasure to work with. I had so much fun and was able to make these so quickly, that I plan to do quite a bit more with foam core in the near future.