Loudspeaker Projects
I seem to have an innate fascination for anything audio-related. As a kid I could play endlessly with a cassette recorder, constantly abusing the controls in attempts to speed up or slow down the recording or chop it up to get the weirdest possible sounds. When we got our first Macintosh computer with built-in audio recording in the early nineties, the art of making noise evolved to a whole new level (and I had a short try at making music). Eventually this fascination evolved into building my own electric guitars and loudspeaker cabinets. Those cabinets have come a long way, starting from nothing but a cardboard box and gradually evolving to high-end MDF designs and exotic things like miniature subwoofers.
Cardboard
My very first experience with the benefits of mounting a speaker into a cabinet instead of letting it work in free air was when I was a kid, with a cardboard box and a simple PC speaker connected to a radio. When covering a hole in the box with the speaker I noticed that the overall sound improved. There was more bass, as far as the term is appropriate. One day a friend gave me some car speakers he salvaged from a wreck. Having no money to buy wood and tools nor skills to do much with wood anyway, I took the cardboard idea to the next level and mounted the speakers in larger boxes made from thick cardboard. The speakers were powered by a classic solid-state amplifier. Of course the result was pretty horrible as the cardboard resonated like crazy, but it was a large improvement over the PC speaker, and it was glorious stereo.
The next upgrade was to implement a crude two-way system by adding ‘tweeters’ that were actually headphone speakers. There was no filtering, I only relied on the fact that the headphone speakers had a much higher impedance than the car speakers, hence would not easily burn out. It was all incredibly messy but it worked in some way. There was a noticeable, albeit small improvement in sound quality after all.
Second Generation
Eventually the resonances of the cardboard started to annoy me to such a degree that I applied even more cardboard to the sides of the ‘cabinets’ and eventually hardboard to make them heavier and stiffer. I also lined the inside with packing foam, some of which had a sound-diffusing profile but most of it was just flat. Every one of those hacks improved the sound bit by bit but eventually I concluded that cardboard is a horrible material(1). I had developed some woodworking skills in the meantime so I rebuilt the cabinets with scrap particle board. I also replaced the headphone speakers with actual high-quality tweeters, which I had bought in my euphoria of entering a real electronics store for the very first time. The things were damn expensive for my small allowance at that time but as the future would show it was a good investment.
My knowledge of speaker systems at that time — albeit still very basic — had improved enough to know that a proper two-way system needed filtering. Otherwise I would quickly make the expensive tweeters go up in a puff of smoke. Being somewhere in the middle of highschool with only basic courses on electronics I still did not have much knowledge about circuits to speak of. I went to the library and found a book about filters, but it was so far above my level that it might as well have been a book about dark magic. Yet, from experiments I had discovered that a capacitor in series with a speaker will only let trebles pass through. I had no idea how to calculate the correct value so I did some wet-finger experiments and ended up using a value of which I can now easily tell that it was so small that the tweeters only reproduced the very highest trebles. However, as I did not perform any low-pass filtering on the other speakers the overall response would have still been pretty OK, because the tweeters only needed to supplement the somewhat poor high-frequency response of the car speakers. But overall it was still a mess. Moreover, I still noticed resonances. The car speakers were hardly hi-fi, the particle board was too thin and too light. The only solution was to restart from scratch.
Third Generation
Highschool was nearing its end and in the meantime my woodworking skills had further improved, and I had built my first electric guitar. I still did not have extensive knowledge about electronics and loudspeakers though. I knew there were formulas to design them but at that time the internet was still in its infancy and speaker design software was expensive. So my next move in an attempt to build good cabinets was to try copying existing good cabinets. I had saved enough money to buy all good materials this time. I meticulously measured the dimensions of my parents' expensive hi-fi cabinets(2) and built new cabinets from MDF. This time I used even more expensive tweeters, new woofers and a prebuilt cross-over circuit with the right characteristics. I was still naive enough to think that by copying the dimensions and mounting speakers of similar sizes the end result would be almost identical to the high-end original cabinets. Of course I know better now but they were actually very good after all. The cabinets are actually three-way and consist each of a ported two-way cabinet and a passband subwoofer. In the original cabinets these two were combined in one tall tower, but I kept them separate in my ‘clones’.
Although I have been using those cabinets for more than ten years, I never really finished them. I intended to give them smooth round corners (which I finally did not so long ago) and paint them with lacquer but they have been sitting as raw naked MDF boxes in my room all the time(3). In retrospect this proved to be a good thing. Being unfinished, I can now easily upgrade them with my further improved knowledge and then finish them after all. The cabinets themselves are well built, complete with reinforcements and proper sound-absorbing foam this time, and the speakers mounted flush with the front. So it would be stupid to have to start from scratch. The biggest problem with them is that although the tweeters and subwoofer drivers are high-end, the woofers are still budget speakers (Monacor SP-150) due to my still limited budget at the time I bought them. They are low power and the manufacturer does not even give all T‑S parameters needed to calculate the frequency response. They are actually meant as general-purpose drivers for mounting in ceilings etc. Yet, as far as I can judge they are a pretty good match for the dimensions and tuning of the cabinet. I plan to replace them with high-end woofers and build a new cross-over filter or fine-tune the existing one.
Fourth Generation
The literally biggest problem with those “third generation” cabinets is that they are big. The room in which they stood for more than ten years is actually disproportionally small in comparison. And when I finally went to study for engineer at the university (at last, proper electronics courses!
) their sheer size made them unsuitable for my student room as well. If there is anything you should learn from this, it is to never build larger cabinets than you really need. Start small and make bigger ones when you see a need for them. My solution was of course to again build all new cabinets in a ‘bookshelf’ form factor. This time I was armed with actual knowledge of electronics, cabinet and cross-over design formulas from the now matured internet, even better woodworking skills, and a much larger budget. So I went a little crazy.
I discovered that those very first tweeters I had ever bought, the DTW72 from Visaton, were actually damn good. Hence I salvaged them from the scrapped second-generation cabinets. They had never worked under any significant load anyway due to the conservative filtering. I also still had a lot of surplus 18mm MDF left from the third-generation cabinets. It is pretty thick for cabinets of the small size I wanted to build, but when it comes to cabinets there is no such thing as too thick panels. I have had my share of resonating panels so I was determined to eliminate virtually all resonances this time. I did not skimp on the woofers either and bought a pair of SPH-135C from Monacor. The whole idea was to build cabinets with no compromises. I had to design my own cross-over filters anyway so I used every trick in the book to compensate for inductance of the tweeters with Zobel filters etc., and I used coils with a higher power rating than strictly needed to minimise parasitic impedance. The filters use a second-order Bessel alignment. The cabinets are a plain sealed enclosure design, again with rounded corners on the front panel to reduce reflections of treble frequencies. The box dimensions correspond to the Golden Ratio and of course the inside is again lined with sound-absorbing material.
The end result is that these cabinets are awesome. Their frequency response is very well balanced and the stereo image they produce is crystal clear. The SPH-135C is also able to output an amazing amount of bass for such small cabinets, I can highly recommend them. Of course there is an inevitable roll-off at the lowest frequencies, but thanks to the closed cabinet design and the quality of the woofers, it is possible to boost the lowest frequencies through careful equalizing to further rectify the frequency response and get very satisfactory deep bass. And finally, the cabinets also look great. I painted them with a black satin gloss lacquer and the carbon fibre cones of the SPH-135C make for a great finishing touch. The only minor eyesore are the shiny pale screws, I should try to find black ones or do as Mick Jagger says with the current ones.
These cabinets actually look a lot like the AudioEngine 5 speakers and are probably quite similar aside from the built-in amplifier, bass reflex port and high-gloss finish. Mind however that I built them in 2001, before AudioEngine was even founded (2002). The AE5 was their first consumer product and was released in 2005.
Party Box
One would think that having built near-perfect speakers I would call it quits, but my audio obsession creates an unstoppable urge to build something audio-related from time to time. My next project was to create a kind of ‘boom box’ from nothing but scrap materials, with a sound quality as high as possible. The ingredients were a set of old Altec Lansing PC speakers, a loose high-quality 80mm driver from another PC speaker, a Velleman K4001 mono amplifier kit I built long ago, and particle board. The idea was to get rid of the ugly resonating plastic housings of the PC speakers and build everything into a wooden cabinet. The third speaker and mono amplifier would make a miniature fourth-order subwoofer to supplement the poor bass of the stereo speakers. The ported chamber of this subwoofer would also double as the housing for all the electronics. The entire thing is powered by an external switching laptop power supply.
Because I had no parameters at all for any of the drivers, it was back to wet-finger work. My idea was to first build a prototype from the particle board, experiment with it and tune it, and then build a final design from MDF. I dimensioned the enclosures according to the best guesses I could come up with and started building. I even did an all-nighter on this one just to have it ready for an actual small party. 
The result was damn ugly but my guesses seemed to have been pretty good because it produced an impressive sound with much more bass than anyone would expect from a heap of scrap materials built into something the size of a shoe box. There was little left to tune, aside from the low-pass filter of the subwoofer and the frequency of the ported enclosure. I never expected the puny 7W mono amplifier to be able to produce so much bass output for the subwoofer. It probably clips considerably but the fourth-order design forms an inherent band-pass filter, and the result is good bass overall.
This design has problems though. First, the switching power supply often causes noise if the audio signal originates from another device with a mains power supply. I'm still not entirely sure about the exact cause but it is either due to ground loops, RFI or a combination of both. It can often be solved by connecting the source through an isolation transformer. Another problem is distortion due to both air leaking from the tiny holes for the cables and potmeters in the ported chamber, and the small diameter of the port. It is amazing how much pressure the subwoofer generates and how such tiny holes can produce so much noise. I managed to eliminate most of this noise by sealing the holes as much as possible. The noise due to the narrow port is less easily solved. Eventually I converted the straightforward cylindrical tube to a kind of dual exponentially tapering tube (made from nothing less than rolled-up pizza boxes), which provides a good compromise between a short tube and low turbulence. A final problem is that the bass is not very evenly distributed. In the end I replaced the subwoofer driver with a commercially available driver for which I had the Thiele&Small parameters such that I could calculate the frequency response, which proves to consist of a rather sharp peak. The possibilities for changing the dimensions of the subwoofer are limited however, because they are mostly dictated by the overall box size and the space required for the electronics.
The box actually works so well as-is that I haven't yet progressed to the second stage of building the MDF version. The main reason to do it anyway would be to make it look less ugly and get rid of the constantly chipping particle board. Also, in the final design I would put the speakers at an angle to get slightly better stereo (see the diagram above).
I did however use the experience gained from this prototype to build my next project, which is a miniature subwoofer for a portable 5.1 surround system. The photo at the right shows some satellite speakers I built to accompany that subwoofer/decoder, made with Altec Lansing drivers salvaged from a dead surround system. Each box has two 8Ω drivers in parallel to match the optimal 4Ω of the decoder's built-in amplifier.
Last but not least, here is a silly experiment: the tiniest (and also ugliest) two-way speaker cabinet I have ever ‘built’ at 35×39×64mm. It uses two headphone drivers, one of which is actually one of the ‘tweeters’ I had originally used in my first cardboard speakers. The other one is better at producing midtones (‘bass’ would not really be the appropriate term), effectively making this a real two-way speaker (no crossover circuit though). The maximum power handling however is probably to be expressed as a fraction of a Watt.
(1): With my current knowledge I could probably make a pretty decent cabinet from nothing but cardboard, but it would take so much effort that it would still be an exercise in futility.
(2): These cabinets were from a small local company that as far as I know has gone bankrupt in the meantime. I can't remember the name.
(3): A piece of advice: do not do this. Do not expose yourself to untreated MDF during an extended period unless it is of the kind that is free of urea formaldehyde. Also by all means avoid inhaling the dust while processing it. I have probably inhaled enough dust and urea formaldehyde fumes the unpainted MDF has been diffusing over the years that my immune system has gone bonkers. When working with MDF nowadays I get nauseated no matter what I try.