My ultimate monitor

 

Introduction

Despite the fact that I am still very satisfied with the small Peerless/SEAS monitors, I knew that I could make them better in a number of ways. I found three main areas for improvement:

  • A quieter and more spacious high frequency respons by building an enclosure that gives much less diffractions effects. Existing enclosures that do this well and are also beautiful for my eyes, have large facets such as for example the loudpeakerboxes offered by Avalon Acoustics.
  • A more transparent midrange by developing an enclosure that needs no acoustical damping. The ónly closed box enclosure that is able to meet this requirement across the entire midrange is the sphere/tube invented by B&W. I do not know of other concepts that are able to prevent internal standing waves so well and therefore need almost no acoustical damping inside the enclosure.
  • Reduction of mechanical resonances by decoupling the edge of the frame of the midwoofer from the enclosure and by magnet mounting of the driver. This is an approach that I found on the website of Siegfried Linkwitz, see this link.

    • I have built a test monitor to verify by my own measurements a working implementation of these three concepts. Doing this I changed a little bit the concept of B&W. My tube is not straight backwards but curls up. Another change is a cone behind the midwoofer which doesn't float. I made the construction in such a way that the cone becomes an integral part of the housing and is able to support the magnet mounting of the driver. Based on this test box I designed my ultimate monitor. I really wouldn't know how I could improve on this one. It is not of concrete, but I used different layers of MDF and plywood which in addition have different thickness thereby obtaining an extremely dead box. The total weigth of enclosure and the loudspeaker drivers is 16.1 kg at a net volume of 4.5 liter. Below a picture of the test box at the stage where it still had a straigth tube behind the internal sphere.

    At first instance I tested the enclosure by using the Peerless HDS 134 midwoofer and the SEAS Millennium tweeter, in that way I was able to make a fair comparison with the 'old' monitor. The sound was significantly better, a nicer high frequency respons but mainly a better lower midrange. Especially the reproduction of piano music was surprising in how much less the sound was smeared and how much dynamics were improved. This gave me the motivation to build two of these very laborious boxes, in such a way that it is not only pleasing for the ear but also for the eye. In fact, I think that the box is so laborious that I don't expect anyone to build it according to the instructions on these pages. My main aim is to inspire people to deviate from standard enclosures and develop their own, maybe simpler version of this concept. On the forum of 'zelfbouwaudio' (in Dutch) are a number of fanatics who are going to use the sphere/tube in their own designs.

    Back to the top

    Drivers

    My aim was to select drivers without compromising quality while maintaining a very friendly price level. Fortunately, in 2005 such drivers have become available for the do-it-yourself community.

    The drivers are not selected to obtain a very large sound pressure, but for the best dispersion. I try to create a point source, which, above the bafflestep, radiates evenly forwards independent of frequency. I believe this behaviour is important especially in the midrange and the best loudspeaker systems available are prove of this with their terrific spaciousness and imaging. This starting point limits the size of the midwoofer. The compromise chosen is the Peerless HDS 152 Exclusive. This driver exhibits an almost ideal dispersion up to 2000 Hz and there are enough tweeters available that can be used from this frequency. The midwoofer is usable from a low 100 Hz if the demand on sound pressure is not too large. If large sound pressures are needed, without real limitations in a normal living room, the crossover could be chosen to be at 200 Hz. For this application, the monitor could be used as head in a three way system with a separate bass unit. See for a practical examle the Wilson WATT/Puppy.

    In my opinion the Peerless HDS 152 Exclusive is a loudspeaker that is a challenge to other, more expensive midwoofer brands. The sound is very clear and transparent, while the roll-off at high frequencies is friendly and well under control. The advanced electromagnetic motor design guarantees an extremely low distortion. I consider the driver to be of top quality, however, I suspect the looks will not be for everyone.

    For the tweeter I ended up with a driver, the SEAS 27TDC (noferro 550), that does not use ferrofluïd and still can be used up to large sound pressures when the crossover is no lower than 2000 Hz. I noted this driver on the testing pages of Mark K. This tweeter tests equally well or better than most top tweeters. The 2nd harmonic distortion component is at an inaudible 1%, the 3rd harmonic distortion component is lower than 0.1% and the higher harmonic distortion products are even lower than 0.01%. I would not dare to state that the absence of ferrofluïd is the cause of the very smooth sound, but my feeling tells me it does.

    Back to the top

    Enclosure

    This entire project was ofcourse mainly about the enclosure. In the introduction I already pointed out the three main design features. I made two links pointing to the design drawings. First a Microsoft word document containing the front, side and upper view and all sections belonging to the internal sphere. Secondly, a link to a page with all sections of the internal inverted horn, calculated with AutoCAD and printed out after that.

    1. Microsoft word document with drawings

    2. Page with the sections of the inverted horn

    Below the entire building process is illustrated. By means of a little explanation below the pictures I hope that all is clear.

    First al 36 panels were sawed out, except the material for the finishing layer which consists of 5 mm bendable plywood and 4 mm MDF. After that all section drawings were cut out and fixed to the panels with a gluestick. The panels are alternating MDF and plywood, depicted by darker yellow and light yellow in the Microsoft word document. To zoom in, click the picture.

    The two pictures above should make clear how I sawed out the sections of the sphere.

    Using another panel as an example, here it is shown how the remaining yellow parts were sawed out with the jigsaw at an angle of 45 degrees.

    Hollowing out the front of the cone, on which the midwoofer will be screwed...

    ...and checking if it will fit by placing the trafo disk, on which the midwoofer will be glued.

    All panels are sawed out.

    Close-up of the panel out of which also the last piece of the cone has been sawed.

    Here it is shown how I have glued the panels. First the panels have to be positioned, three small holes are bored and through these holes the panels are screwed together.

    Then the panels are taken apart, sanded and glued.

    Screwing the panels again for correct positioning, after which a lot of clamps are put on.

    After drying the screws are removed. When glueing the thin panels, 3 or 4 panels can easily be glued at the same time with a set of screws of sufficient length.

    Groups of 3 + 3 + 3 + 4 glued panels. You see the last 13 panels of each enclosure. On the panels of the second enclosure circles of paper can be seen. They are glued on the panels to prepare for sawing out the crossover compartments.

    The crossover compartments are sawed out...

    ...and the 13 panels are glued together.

    On the left side of the picture you can see an 18 mm hole through the back of the enclosure. This hole is necessary to be able to screw the midwoofer to the cone. On the right side of the picture you can see the next block of glued panels, numbers *5-*5 up to and including 8-8.

    Front and back side of the inverted horn and in the back side you see the compartments for the crossvoer.

    The arrows point at places were I have drilled holes for the cables between drivers and crossover.

    Three strings of long hair wool, fluffed and bonded with a rubber. The weight is about 10 gr.

    The wool has been pulled through the inverted horn by attaching a piece of string to the rubber, after which both the piece of string and rubber have been removed. Numerous experiments and measurements have shown that this is the only damping needed to remove all resonances. The part of the enclosure with the internal sphere, see the left part of the picture, remains empty. The midwoofer is therefore able to move freely without loss of energy and storing of energy in damping material.

    To be able to tightly fix the midwoofer with a coach bolt, I glued a washer to the back of the cone.

    After that, the two halves of the enclosure with the sphere and the inverted horn are glued together. Onto the front I have put the last 6 mm MDF panel (section *15-*15).

    Now the enclosure can be finished. First a piece of 5 mm flexible plywood is glued at the bottom.

    After the glue was dry the parts sticking out were removed by sanding.

    After that, the lines for the large facets were drawn on the MDF, click on the picture to see the measures.

    To take into account the finishing layer of 4 mm MDF, lines are drawn within these measures at a distance of 5,5 mm.

    Along these lines large facets are sawed at an angle of 45 degrees. The picture shows the result after sanding.

    The thin backpanel is also sanded at the vertical edges at an angle of 45 degrees.

    The first side layer of flexible plywood (5 mm) is glued.

    The result after sanding the parts that sticked out.

    In the picture above also the top of the enclosure is provided with a layer of 5 mm flexible plywood.

    On second thought I found the baffle a little bit too slim, therefore the large facets left and right were provided also with a layer of 5 mm flexible plywood.

    In the picture above the curved enclosure sides are covered with a second layer of flexible plywood. Furthermore there is the finishing 4 mm MDF layer on the top and bottom.

    Covering the large facets with the 4 mm MDF finishing layer.

    Here the small facet at the top front is covered with 4 mm MDF and all sticking out parts have been removed by sanding.

    The covers for the front, with the necessary cut-outs for the drivers.

    Carefully positioning this last finishing layer.

    Letting dry this finishing layer using the own weight of the enclosure.

    The enclosure has been primed.

    The result after sanding, filling of all cracks, sanding again and painting.

    Direct underneath the frame of the midwoofer a rubber hose has to be placed. I discovered that a rubber gasoline hose of 8 mm outer diameter and 4 mm inner diameter fulfilled my needs. It has to be cut at the right measure and glued together at both ends with special glue (the one that dries within seconds) to make a circle.

    The midwoofer has not been mounted traditionally at the edge of the frame, but by mounting the magnet to the cone within the enclosure. To this end I bought trafo discs which, after sticking a screw bolt through their holes, have been glued to the back (magnet) of the midwoofer by metal glue (two component).

    The screw bolt has been guided through the hole in the cone. By sticking a spanner through the hole in the back of the enclosure it has been tightly fixed with a nut. This work requires precision and patience. Also do not forget to solder the loudspeaker cables to the midwoofer and lead them through the enclosure towards the crossover compartment!

    And here it is! The tweeter also floats in the enclosure, lightly fixed with a rubber hose around it.

    Back to the top

    Crossover

    The active filter that I built for My own system has been adapted for my ultimate monitor to optimally work with the subwoofer.

    The passive crossover has been designed such that a 3rd order Butterworth response will be obtained.

    I have noticed, after a long period of listening, that the tweeter should be connected in phase. When the tweeter is connected out of phase, the response curve is very flat on-axis and at 15 degrees off-axis and somewhat less flat further off-axis. But if the tweeter is connected in phase, a slight dip occurs around 3500 Hz on-axis and at 15 degrees off-axis. Further off-axis, however, the respons is flatter in this case, apart from the beaming of the tweeter. This response sounds more natural and with a little bit more depth.

    The electrical respons of the passive filter at the loudspeaker terminals, as predicted by SpeakerWorkshop. Green the amplitude and phase of the tweeter filter, red the amplitude and phase of the midwoofer filter.

    The acoustical respons at 15 degrees horizontally, as predicted by SpeakerWorkshop. I have found through experience that the 15 degrees response is a very nice design target for correct tonal balance. Green the tweeter amplitude, red the midwoofer amplitude and black the resulting total monitor response. The 15 degrees angle is not a simulation, I just imported measurement data of midwoofer and tweeter as obtained when measuring these drivers unfiltered at this angle. The measurement data was obtained by placing the monitor upon a bass unit.

    Back to the top

    Measurements

    After mounting the crossover into the monitors I have performed a number of measurements with DLSA Pro. The monitors sound very well and the measurements confirm what I hear.

    Above the electrical measurement is shown, meaning the amplitude and phase of the impedance. A 3rd order Butterworth crossover gives a low impedance peak at the crossover point, e.g. when compared with a 4th order Linkwitz-Riley crossover.

    The acoustical amplitude and phase at 15 degrees off-axis horizontally, measured at a heigth precisely in between the midwoofer and tweeter and at 1 meter distance of the box. The monitor has been measured as is, without the active bafflestep compensation. Therefore a slight amplitude decrease can be observed below 1000 Hz.

    The horizontal dispersion: the amplitude response at 0 degrees (black), 30 degrees (blue), 45 degrees (red) and 60 degrees (green). Diffraction at the lower sharp edge of the monitor gives a broad bump at 1000 Hz. I didn't correct this in the crossover, it sounds the best this way. The bump disappears almost completely when the monitor is placed upon a bass unit.

    The step response (at 15 degrees off-axis horizontally) corresponds with a 3rd order Butterworth system.

    The waterfall plot at 15 degrees horizontally. Almost across the whole band the response is 30 dB down within 1 msec!

    Back to the top