Jump to content

WaveRider dipole speaker


Recommended Posts

I'm notoriously bad at documenting my design process, so this thread will be as much about personal development for me as sharing this design. 

 

Basis of Design

Goal

My goal for this speaker is to combine my love of music with my love of music to create speakers for an open living area in my home, which happens to be near a popular surfing beach. I want the visual design to be evocative of a surfboard.

 

Criteria

  • It must utilise some of the vast array of drivers I have in my attic;
  • To maintain a minimal presence within the room it must be a complete system, so have inbuilt active crossovers, filters, and amplifiers and be capable of full frequency response without requiring additional subwoofers;
  • It must be capable of reproducing dynamic peaks of at least 105 dB at 25 Hz;
  • It must approximate controlled directivity;
  • It should include FIR filtering for linear phase;
  • It should include inbuilt room correction/EQ; and
  • To avoid my usual time blowouts on such projects it must use of-the-shelf electronics rather than self designed, tested, and finished.

 

Given the criteria, I've settled on a dipole configuration with the lowest octave or so augmented by a monopole woofer, essentially transitioning to a combined cardioid in this range. To provide as close an approximation as possible to true dipole performance will necessitate a five-way design. For optimum dipole performance the dimensions between the maximum frequency any driver can play up to is determined by the formula F = 0.25(v/D) where v is the speed of sound and D is the path length, with an absolute upper limit of 0.5(v/D) beyond which the crossover should substantially attenuate further response downward to avoid comb filtering. That requires as short a path length as possible for the highest frequencies and such as small driver operating in the upper midrange / lower treble simply cannot provide the air displacement required for the midbass. So we need a driver for the monopole bass, a driver for the dipole bass, a driver for the midbass / lower midrange, a driver for the midrange / upper treble, and a tweeter for the upper treble, at which point the path length requires either abandoning dipole radiation or using two waveguide tweeters back to back with time alignment. I'm not convinced that rear radiation matters at the high frequencies the tweeter will be covering and in any case the later maximum dimensions available for the plate amplifiers precluded making this a six-way design.

 

Driver Selection

The easiest choice in driver selection was for the bass. I have four Scan-Speak Discovery 30W-4558T00 that seem perfect for this role.

image.png.4173d0087f210816f6c89dad529d793b.png

 

So now we have a start to the project.

Edited by Karri
  • Like 3
Link to comment
Share on other sites



Posted (edited)

The bass drivers were then placed into a an initial design draft. To get the required narrow baffle dimensions for the midrange driver required separating the midrange and tweeter section from the 'surfboard'. Next was to select a midbass driver that met the design goals and also fitted with the physical design. I had 200mm, 180mm, and 150mm options. To get the shape I wanted at the top of the 'surfboard' whilst keeping a close spacing between this driver and the midrange meant selecting 150mm. The best candidates I had on hand were SB Acoustics Satori and Scan-Speak Illuminator. Accounting for the required 6 dB per octave EQ below the dipole peak and needing this driver to meet the freq limit of the dipole bass driver meant linear excursion capability was a key criteria, and so the 15WU/4741T00 Illuminator was chosen for its +/- 9mm capability.

image.png.98a75ea81fa9a1246ac907bdbdf06c79.png

For the midrange I wanted it as small as possible to maximise the potential dipole upper frequency. The two candidates I had were the SB Acoustics SB65WBAC25 and Dayton ND65. Both are wideband units, so have relatively high linear excursion and both can perform well within the intended passband. I chose the ND65 both because I prefer the appearance and because I have a box full of them. At 65mm it allows for a short path length and therefore high potential freq.

image.png.a885dcebc1a41700ed7a6876cc20c55d.png

Lastly, I chose a Satori TW29RN tweeter for its excellent performance and that, with the faceplate removed, the 70 mm diameter of the motor provides a good match for the ND65 in terms of baffle dimensions.

image.png.e54981bd010b841a5d5bd19add322aa6.png

 

 

Edited by Karri
  • Like 1
Link to comment
Share on other sites

Posted (edited)

After much design work in FreeCAD, we have something starting to look like a speaker crossed with a surfboard - at least enough to satisfy me that it was evocative of a surfboard.

image.png.fedb7a4b2b23a5ac5d013f7ecbe5b832.png

image.png.579314fbdfc1ce280a1423c2d34c25d2.png

image.png.6fa7762fbc528ebb5e1332af8e39f324.png

The spine exists to support drivers and also is internally routed to hide the wiring, with this coming out behind each of the dipole bass and midbass drivers as well as near the top for the midrange and tweeter. The midrange and tweeter baffle is supported by two 8mm carbon fibre tubes, which emerge behind each of the midbass and midrange drivers to route the wiring. The midrange and tweeter drivers are rear-mounted, with a 45 degree waveguide for the tweeter. At the rear of the monopole bass unit enclosure there is a sealed section and flush routing for the two plate amplifiers with inbuilt DSP.

Edited by Karri
  • Like 2
Link to comment
Share on other sites

For amplification, crossovers, driver EQ, room EQ, and FIR filtering I chose Hypex Fusion units. FIR has been a recent addition to these, and I know nCore well and they perform exceptionally. To maintain reasonable dimensions the subwoofer enclosure needed to be relatively undersized, which will be corrected for with a Linkwitz transform. This will require considerable amplifier power to drive the subwoofer to its excursion limits. A FA502 2 x 500 W RMS @ 4 ohm unit was chosen for the two bass drivers. A FA253 2 x 250 W and 1 x 100 W unit is more than capable of driving the midbass, midrange, and tweeter drivers.

 

Remember how I said I didn't have the available space for six channels? The FA503 would have provided an additional channel for a rear tweeter, but is too tall for the subwoofer enclosure. Had I really wanted to I could probably have made the subwoofer enclosure taller, but it needed to clear the dipole bass driver and in any case the midrange driver will be operating so high in frequency that I didn't feel it was necessary - its only purpose would have been to extend the upper frequency range of the reflected sound, as the wavelength is far too short to achieve a cancelation dipole effect and we already shape the forward radiation with the waveguide and inherent directivity of the tweeter, and with the surface of the front wall those frequencies are too high to effectively reflect in any case. It simply wasn't beneficial to the performance and would have created problems elsewhere.

 

FA502.jpg

HYPEX-FUSION-AMP-FA253-3.jpg

  • Like 1
Link to comment
Share on other sites

Posted (edited)

I love Australian hardwoods and have a few sheets of Jarrah glulam on hand. This is 26mm thickness and will be used for the 'surfboard' baffle. I've not tested it, but hope that by having the spine supporting the drivers that resonance should not be an issue. If testing finds that this is an issue then I'll likely resort to Jarrah or Karri veneer MDF with Jarrah hardwood edge trimming. For the midrange and tweeter baffle it needs to be kept thin to minimise the dipole path length, so 19mm Jarrah will be used. The spine and subwoofer enclosure are MDF, with two layers of 16mm for the spine to allow for internal routing and 16mm for the subwoofer enclosure with extensive bracing. The spine will have internal carbon fibre structural square tubing.

image.png.ca99d4718ed514dab6d7f1937535c3c8.png

The pieces have routing to ensure they glue precisely. The screw holes have been dimensioned for thread inserts, which will be set with epoxy on the wood threads. As can be seen, the FA253 and FA502 plate amplifiers have differing depth requirements. The front panel is tilted at 7.5 degrees to match the intended surfboard tilt angle. Of course, the midrange and tweeter baffle has no tilt and is vertical.

Edited by Karri
  • Like 1
Link to comment
Share on other sites



Posted (edited)
1 hour ago, frednork said:

You seem to know what you are doing so just wondering what is your objective measure/s of a successful build.

If I understand the question correctly, ultimately it is to reproduce music as realistically as possible within the room! In practice, that means that the speaker should meet these criteria as measured:

  • Be capable of lifelike SPL including dynamic peaks;
  • Be capable of reproducing 20 Hz to 20 kHz;
  • Reproduce a frequency response that is as flat as possible, and preferably within +/- 1.5 dB;
  • Have as close to constant directivity as possible;
  • Have as inert enclosures and baffles as possible, certainly below the point of audible interference;
  • Probably a string of other criteria that aren't immediately springing to mind.

I do find that for spatial perception some reflected energy of between 6 and maybe 15 ms or so is required, which is why I didn't choose a horn design. Pragmatically, something like -6 dB at 20 HZ can be acceptable in order to achieve the target SPL. How the speaker interacts with the room is important, and I believe that a dipole speaker has advantages here, as do horns. 

 

I find music a social experience, so I prefer a speaker that provides a smooth frequency response and sound stage over several seats, rather than only a single position.

 

Whilst somewhat controversial, I do find that linear phase is required for lifelike reproduction. At the very least, the purest in me prefers it.

 

Aesthetics are subjective, but the design has to be something I appreciate and that my wife at least finds acceptable.

 

I want these to last me many years, so one consideration is longevity. Typically, the electronics will have a considerably shorted lifespan than the drivers and hardware. Fortunately, due to my background I can often repair the electronics, but I did consider the case that the plate amps might fail, not be repairable, and no longer be available. In this case I would either adapt different amplification and DSP potentially by adapting them via milling plates to suit the current dimensions or, at worst, make new subwoofer enclosures that suit the new parts and simply unscrew the old enclosure and screw on the new one. Probably a moot point, though, as by then I may have designed a new speaker anyway.

 

Of course, these are all for this specific application. For other applications such as studio mixing, an 'invisible' lifestyle system, or where I want a separate and potentially advanced lower bass solution these may change.

 

As this is a personal project and will not ever be marketed as a finished product, I do not need to be concerned with cost, market price point, yield, manufacturing time, etc. These would be important criteria for a production design.

 

I don't feel that this is much of an answer, but I'd really need to put serious thought into providing something more thorough and I only do this as a an occasional hobby these days. To keep it in perspective, a total of two days have gone into this design. I simply needed speakers for this room so quickly designed something that I know will perform exceptionally using drivers I have on hand. The real work will go into the DSP/crossover programming and then measuring to confirm and fine tune.

Edited by Karri
  • Like 1
Link to comment
Share on other sites

13 minutes ago, Karri said:

 

I don't feel that this is much of an answer, but I'd really need to put serious thought into providing something more thorough and I only do this as a an occasional hobby these days.

No , really your answer is all good.

Sorry if my question seemed to be trying to elicit something specific. Just  generallycurious really,  and wondering how you may judge when it is done from an objective standpoint. Will be looking forward to how you progress. Thanks for sharing your project!!!

  • Like 1
Link to comment
Share on other sites

1 hour ago, frednork said:

No , really your answer is all good.

Sorry if my question seemed to be trying to elicit something specific. Just  generallycurious really,  and wondering how you may judge when it is done from an objective standpoint. Will be looking forward to how you progress. Thanks for sharing your project!!!

Not at all. It's a pleasure, and I look forward to sharing the process and progress between this early design stage and the final result! I've kept all of the above quite simple and high level, but am more than happy to provide greater detail for anyone interested. I think the aesthetic will probably particular to my own tastes and so the design may not appeal to others, not to mention that these would by no means be cheap to build from a DIY perspective, but should anyone be interested then I am more than happy to share the design and DSP files once everything is completed. More over, I look forward to demonstrating them for any WA members interested in seeing and hearing them once they're done. 

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

Posted (edited)

The next item on the design agenda is securing the bass and midbass dipole drivers to the spine. The intention is to have the drivers supported by their motors which, being the centre of mass where the force is generated, is preferable to having them supported by the baffle. This will be an iterative process requiring adjustment to the spine design to match and hopefully eventually achieving a pleasing combined aesthetic, but as a starting point I've tonight carefully measured the dimensions of the motor assembly of the 30W-4558T00 driver and designed the basic structure to clasp the motor each side of the spine, with the two halves being secured through the spine with 4 x M4 hex head metric screws with washers and nuts. These will likely be 3D printed in ASA then hand sanded and painted. This driver ventilates only forwards, so the clasp does not need to allow for motor cooling. 

image.png.8974660154de77e6e43ae814331d9c67.png

Tapering it back may both look cool and offer some additional lateral support, so I'll probably test that idea.

Edited by Karri
  • Like 1
Link to comment
Share on other sites



A little more CAD work tonight, and we're starting to see it come together mechanically and visually. The part toward the bottom of the baffle is the OLED display on the master speaker, which will show volume, selected input, etc. via the remote.

image.png.21f9cbd80d58157917b7dbcf9cab0de7.png

  • Like 2
Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.




×
×
  • Create New...
To Top