Introduction: BI-AMP THAT SPEAKER

About: Freelance audio designer. Into valves and hybrid gear. Dabbles in astronomy.

Always wanted to build an active speaker but put off by the exorbitant price of commercial crossovers? Well now help is at hand. For 20pounds you can simply build the circuit kit described here and you're ready to go. All you need is a pair of stereo amps and a signal source. As an audio engineer of 35years standing I know that the only good sounding speakers are active. But going active usually requires spending big in the equipment department. Really all you need is a good filter and a spare stereo amp. As the tweeter channel only needs a few watts of power this spare amp doesn't need to be expensive either.
If you check out the net you'll find lots of active crossovers for sale. Many use esoteric technology but all are overpriced and unsuitable for experimenters. So here I will introduce a new concept for audio experimenters, the core kit. I believe it's counterproductive to try sending heavy items overseas. Especially as most can get the heavy and expensive parts locally and at a lower price. It makes sense then to design high quality kits which can be operated by a wide range of locally available power supplies. By locally I mean everywhere with a mains supply . The circuits are designed to operate from battery eliminator type plug in power supplies. Current requirement is 20mA. If plug in supplies, or the mains aren't available the circuit can be operated from batteries, a car battery is ideal.

Similarly a basic housing is provided. If the builder wants more he can make his own arrangements locally. The payoff is less expense and higher performance for the customer and, hopefully, higher sales for us. Why go active? Audiophiles know that so called passive speakers, those driven from a single power amp passive filters will never sound as good as an active speaker. Putting a passive filter in series with the speaker drivers will rob the drivers of the full damping factor of the amplifier driving it. The result, muzzy sound. Furthermore most of the energy in music occurs below 1kHz so a single amplifier can be driven into overload passing the unwanted high frequencies distortion components into the tweeter via the passive crossover. In the active system the tweeter will be driven from a separate amp so that even if the woofer amp is overloading no damage can be caused to the tweeter and the whole system will still sound sweet.
Another advantage of the active system is that the required response curves are easily generated by standard filter stages and are independent of speaker parameters. Anyone who has tried to design passive crossovers will be well aware of the difficulties, and compromises, involved.

Step 1: GETTING THE PARTS

Getting the parts. A full circuit diagram and component list for this stereo filter is available from my website http://www.macaulayaudio.co.uk

Step 2: HOW IT WORKS

How it works. Versatility is the hallmark of this filter. It is intended to be universal insofar as any filter system can be. The turnover frequency in the standard version has been set to 2.5kHz. Why? Well this is within the linear range off most small <10 inch diameter woofers. It is also safely above the resonant frequency of most dome tweeters. The turnover frequency can be altered as required by changing 4 resistors, R4,R5,R6 and R7. Knowing the required turnover frequency, f, is expressed in kHz then the required resistor value R will be in kohms.
R = 15916 / f
Technically the circuit is based upon the well known and respected TL074 quad op amp. The filter circuit is based on these op-amps operating in unity gain mode. This means that the maximum distortion will be less than 0.003% across the audio band. Similarly noise levels will be <100db below a line level input. In practice noise is inaudible.
Standard Sallen & Key filter circuitry is employed for both high and low pass sections. Both of which operate as 2nd order filters giving 12db/octave roll-off in the stop band. Filter 'Q' has been set at 0.5 giving a Bessel roll-off characteristic as used in 'Linkwitz-Riley' crossover filters. The attraction of Q=0.5 filter sections is because they are 'dead beat', that is they don't suffer from overshoot on transients. This makes them ideal for audio filters.

Step 3: PUTTING IT TOGETHER

Assembly. Most of the components are mounted on the strip-board panel. The layout of this is shown in Fig 1. First order of work is to make the breaks in the strip-board tracks, shown as 'X' in the drawing. These are best made with a proprietary tool or alternately by twisting a 3mm drill bit in the appropriate holes. Once the breaks in the tracks have been made the components can be mounted. First though insert the wire links between A2-J2, B1-C1, B16-C16, G8-H8 and G19-H19.

It's a matter of choice what order the other components are inserted and soldered in place. Personally I would start with the 14pin DIP socket as this then makes positions of the remaining parts easy to find. The main thing that needs to be carefully observed is the orientation of the electrolytic, C1 the diode, D1 and of course IC1 needs to be correctly mounted in its socket.
Once satisfied that the board has been correctly built check it out! For this you'll need a multi-meter switched to the low ohms range. Systematically check the resistance between adjacent tracks to test for lack of continuity. Except across the tracks where there are wire links you shouldn't find any. If you do it means that you have a solder bridge between the tracks. These need to be hunted down and eliminated. Assuming all is well the next stage of construction can commence.


Having built and tested the board attention can now be turned to the mechanics. The circuit is mounted on a 3mm thick plastic panel which measures 157 x 61mm. The mechanical drawing shows the position of the mounting holes for the input/output phonos and dc socket etc. Once these have been drilled out as shown mount the sockets into place.

Now solder flying leads to the board as shown in the schematic. Leave these about 200mm long to allow for easy soldering. As the device will be fed from a low impedance source there is no need for screened cables here. Simple hook-up wire, say 7/02 guage will do. A special word though about the power supply. As mentioned earlier this device and further projects have been designed to use readily available power supplies. Any voltage between 9-35VDC will do. Cheap mains operated dc battery eliminators are ideal and they are available the world over. The voltage range also ensures that operation can be had from a battery. Anything from a PP3 upwards, a Car battery is ideal.

Part of the power supply for the board is the LED D2. This is operated in series with the circuit the quiescent current consumed by the circuit lights it up when in operation. Here I've used a very simple but surprisingly effective method of attachment. A small piece of masking tape! Although it's well known it bears repeating. A simple way of determining the cathode (k) terminal of the led is to hold it up to a light. Inside the led bulb you will see a horizontal bar connected to one side. This side is the cathode. To mount push the led into the 3mm mounting hole. Bend the leads down flat to the panel. Secure by covering the rear of the hole with a piece of masking tape.

Now comes the problem. How best to fix the board into position without unsightly drill holes? The solution chosen is to use double sided self adhesive pads available from all good stationers. Position one of these pads at the four corners of where you want the board to be. Note that the adhesive surfaces are protected by a paper tab. These need to be peeled off to expose the adhesive. Leave the upward facing tabs in situ. You don't want to permanently fix the board until you have tested it!

The last bit of construction is to wire the board and sockets together as shown in the schematic.

Step 4: FINAL TESTING

Once you have done, test it! Apply power and a signal source to the circuit. On power on the LED should light. If, after a couple of seconds nothing goes awry you can be confident enough to fit a signal lead to each one of the outputs in turn. You should have two bass/midrange and two treble channels. Finally fit he board into position on the adhesive pads. Good listening.

Step 5:

The graph below shows the frequency response of the filter. Turnover frequency is 2.5kHz.