The appeal of DIY speaker design

diy speaker couple


On an audiophile blog recently, a commenter was saying that audio DIY must always remain a ‘niche’ activity that very few people could ever be interested in. Indeed, the idea of spending evenings and weekends wrestling with soldering irons and circular saws in order to produce something inferior to that which can be bought ready-made for less than the price of the raw materials, would seem very silly indeed – if it was true. But I would say that hi-fi is unique in the way that people get drawn into a very costly addiction with ever-escalating prices, and DIY could offer them a better fix while saving money.

For a start, the majority of ‘active’ audiophiles do a substantial amount of ‘soft DIY’, already: continually installing new acquisitions then changing their minds and packaging them up to sell on eBay, fitting room treatments, experimenting with special ‘cable lifters’, building shelves to hold their collections of 3000 LPs, setting up their turntable azimuth, keeping their record cleaning machines topped up with special fluids, demagnetising their tape heads, and so on. These things are often motivated by technical or scientific claims, so the average audiophile is immersed in a world of technical tinkering anyway. Each little job completed gives a small sense of achievement, even creativity. In fact I think the audiophile addiction is not just one of seeking out better musical experiences, or better bragging rights, or craving continuous change; I think there is also an element of creativity, similar to some people’s compulsion to build musical instruments. Assembling the perfect system and listening environment is a creative act, and so I don’t think the active audiophile can ever reach a state of complete contentment, and nor would they want to – despite what they say. Continually changing the system is not an unhealthy thing in itself, in other words. The thing I don’t like about it is the way the cost of achieving each creative ‘hit’ ramps up rapidly into the stratosphere.

If an excellent commercial hi-fi system cost just a few hundred pounds (and was acknowledged to be excellent so that the owner could feel good about owning it) then the ‘hard’ DIY imperative would be reduced. In the world of hi-fi, though, this cannot happen: quality is perceived to be synonymous with cost and he who takes the first steps in this hobby must accept that he is on a path to ever-increasing expense, regardless of the technological progress that reduces costs and increases quality in the ‘civilian’ world.

DIY is a way of cancelling the rising expense factor from the equation. By embarking on the path of audio DIY, the addicted audiophile opens up an infinite number of creative possibilities and the potential for great satisfaction at affordable prices. My claim is that by choosing an intelligent sub-set of audio ‘targets’, they can produce something that is actually better than can be bought commercially – if only because the system is infinitely tunable for their specific room and ears.

Without a heavy technical background in physics and maths it may look as though the average audiophile might not be able to achieve all that much, but I would claim otherwise. I think that an awful lot of the heavy technical stuff is necessary only for overcoming the challenges of already-obsolete technology or for perpetually tinkering with problems that have already been solved, such as amplification. From my viewpoint we can place certain stakes in the ground, and worry about them no longer.

First, I would consider using no source but a digitally-derived one: a CD player, PC, streamer, they’re all equivalent. I wouldn’t spend my time building DACs using DIY – it’s all done with chips that are churned out for pence, and all I could achieve would be a prettier box or to try to persuade myself that ‘higher quality’ (more expensive) power supply components made a difference. I could also get problems with massive speaker-damaging thumps at power-on or -off, or hum loops, and (as in real commercial engineering) find that the largest effort in these sorts of things is spent in circumventing ‘fault conditions’ rather than anything truly creative.

Next, I would go active with my speakers, using solid state amplification. For amplifiers the same aversion to DIY would apply – I’ve been there, done that, and it’s completely pointless. Unless I know enough to design my own amplifier, all I am doing is packaging someone else’s design. Even if I design my own, it’s only a variation on a standard topology, and is it going to sound any different from all the other amplifiers out there? Again, the greatest part of the effort may be spent on apparently-peripheral aspects like avoiding hum, or thumps at switch-on and switch-off. This time, the thing can actually be dangerous and the more you think about it the more you worry. The first amplifier I built as a teenager, I didn’t know enough to worry about anything, but the last one I built bristled with soft start circuits, ‘mains lifters’ and crowbar protection, and yet I still couldn’t be absolutely sure it was safe, and I knew that it was not strictly legal. Nor could I be sure that it couldn’t destroy the speakers if the power fluctuated, for example. I have experienced a tweeter going up in smoke due to an unstable amplifier, and the slight sense of ‘anticipation’ every time I applied power to a DIY amp never left after that! If my DIY efforts were spent on building amplifiers and not speakers, would I be happy to connect commercial speakers worth thousands of pounds to them? In effect, I know too much about what can go wrong, and I am more than happy to hand that worry over to someone else!

Which leaves the speakers, and I think this is where it gets very interesting indeed. Speaker design is still an open problem, and unlike the dry requirements for DACs and amplifiers, it involves real physics, trade-offs and compromises that may work better in some rooms than others and on particular recordings – the DIY-er can produce something unique. At the same time, the techniques involved in building a valid speaker are relatively ‘agricultural’ and it is possible for anyone who can do simple woodwork to make something that works just fine, even if it is not beautiful. The possibilities of where to go from there are endless: set up a beautiful workshop and learn the skills of fine joinery, or learn to use a CAD package and contract out the machining and construction to professionals. Create speakers that look like scuplture or, like commenter Serge Auckland, restore and modify an existing commercial speaker and have it finished in a new veneer to your taste. Form genuinely follows function with speakers – it is not just a case of dressing up a circuit board with a fake, or pointless, massive chassis. Also, the DIY-er’s input is real, measurable and audible rather than imagined. I think that the key is to put in place a DSP ‘core’, and from then on it is pure creativity. No matter what mistakes are made, no one will die, and the potential for damage is small. MDF comes in huge sheets for £30. Raw drivers are usually in the tens of pounds range, not thousands, so even if damage does occur and the DIY-er sticks a screwdriver through a cone, it’s not the end of the world.

A few basic (and genuinely interesting) physical principles apply to the choice of drivers and the dimensions and shapes of enclosures but, I would claim, not a great deal of mandatory maths. I think there is a large difference between knowing the maths that describes something (and anything can be described with maths), and relying on maths to design, modify, or optimise something – which can only work perfectly when all the requirements are known perfectly. Maths may be able to perfectly describe the performance of a particular 3″ driver in a box, and the same for a 4″ driver, but it can’t tell you unequivocally which to use in a particular design. Nor can it tell you how much better it will sound when driven actively than passively, or how it will be affected by correcting it with DSP. Maths cannot say with precision what the crossover frequency and slope should be, nor can it tell you just how forgiving any of these things may turn out to be in terms of audibility. One thing is for sure, though: by going active with DSP, the problem is simplified beyond recognition. The designer can play the speaker’s parameters like a Stradivarius, unlike the passive designer who must do it wearing mittens. All designs are exercises in conflicting requirements, and where to place the multi-dimensional surface that describes the final design is not going to be completely scientific, so an ‘artist’ may easily get results that are better than a wannabe ‘scientist’.

Fortunately, using DSP and possibly a little woodwork, a speaker design can be modified at any time. Any hard maths that is necessary can be performed by software packages (that tell you the preferred dimensions of enclosure for a particular driver, for example) and by the real time maths being performed by the DSP and the settings and measurements it is fed with. Measurements may (or, I am beginning to think, may not) be necessary, but a microphone and pre-amplifier are very cheap these days. Learning to use the microphone and software is of the same order of complexity as learning to use Photoshop well – not trivial, but fairly commonsensical, fascinating and, to use a cliche, ’empowering’.

What I am getting round to, is contrasting the tantalising range of options that the DIY-er has, against the bleak and unforgiving prospects that await the typical audiophile when trying to score his next fix. The normal audiophile is going to be changing something in his system, that’s for sure. My vision of what he goes through is something like this:

He may buy some pocket money ‘tweaks’ (order of magnitude £10-£100) and hear no real difference. He may splurge some hard-earned cash on some new cables (order of magnitude: £100-£1000) and spend a few days straining to hear or imagine a difference. He may make a much bigger decision and do some heavy research, trailing round various hi-fi dealers to try and establish which new amplifier he should buy (order of magnitude: £1000-£10000). If it’s solid state, he’s unlikely to hear any real difference. If it’s valve there may be a difference but not in a good way. Most difficult of all, he may decide he cannot live without a new pair of speakers. These will all sound different, but it will be almost impossible to decide which sounds best. However, he has to have his fix, and at some point a few weeks into the process he will make the decision, pay the money (order of magnitude £1000-£10000) and wait for delivery. In the meantime he will read a review of a new brand of speaker and wish that he had heard it before putting down the cash – he could change his mind and cancel delivery of the ones he’s ordered, but it might be weeks until he gets to hear the new brand, and he knows he can’t wait that long. When the speakers are delivered, now-tainted by the knowledge that there are some others he might have held out for, they don’t sound so good in his room (he thinks), and he spends a long time shuffling them around and ‘burning them in’. Possibly, he thinks, he also needs a different amp, or some new speaker cables. But two weeks in, he’s finally happy and spends the best evening listening to music that he can remember. But he realises he’s turning the volume up a bit high at one point and decides to call it a night. Next morning he decides to have one more blast of them before going to work, but they sound different – not as good. He worries about it all day. That night his worst fears are confirmed: they’re sounding pretty average. Could he have damaged them? Probably not, but they cost £5000 and he hadn’t budgeted for upgrading so soon. Damn, he knew he should have gone for the next model up with the beryllium tweeter. Or is it the amp? His wife asks what’s wrong because he looks ashen. He logs onto a forum in the hope of some answers… etc. etc.

The sensible DIY-er knows enough to realise that an adequate piece of wire is just a piece of wire, and that all decent solid state amps sound the same, especially when used in an active configuration. Ditto for DACs, so he can relax about those things. The really interesting part is the speakers, and here he can indulge his creativity. Instead of buying some new cables or another inaudible tweak, he can actually change something real: maybe try a different crossover frequency or slope between mid and tweeter. It will affect the in-room response and the extent to which the drivers have to handle frequencies they’re uncomfortable with. They’re all ‘valid’ settings over a fairly broad range, but they all sound ever-so-slightly different. Maybe a slight tweak to the tweeter delay or gain will yield another tiny, yet measurable, change that may, or may not, be preferable. Such fine, independent adjustments are simply impossible with passive crossovers.

If it’s a free Sunday afternoon, the DIY-er can do something more ambitious: improvise some rounded ‘teardrop’ housings for a couple of small tweeters he has lying around and cable-tie those on top of the housings he made for the mids last week from plastic pipe stuffed with wadding material. If the experiment works, then next weekend he can design a complete housing featuring organic shapes using CAD and send it off to have it machined from solid plywood or 3D-printed, finished in a fine gloss coating in white… or red? etc. etc.

In the old days, the vast majority of the effort (and frustration and fear) was spent wrestling with a crude crossover that could never yield the necessary finesse. DSP on the other hand turns DIY speaker design into nothing but fun. I recommend it.
As far as I can tell, this package for $60 should be enough to do full DSP-based active speaker processing using a PC:

You’d need a PC with a multi-channel sound card, and it may be compatible with USB-based cards, or maybe an HDMI-based output, which opens up the possibilities of which types of amplifiers you could use. A 6.1 AV amp with HDMI input and a direct multichannel mode (if such a thing exists) would be pretty much the ideal system for a three-way design.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s