Cloud

Driver Choices

Pretty arbitary, The seas tweeter was relatively low in cost, and high in performance under objective testing. The peerless woofers seemed pretty decent. In hindsight I would have gone for the phased plugged versions. I didn't because I was a little scared of the crossover design and they were more expensive.

Baffle Design

The speaker was early days for me. Little thought was put into the baffle design. The tweeter suffers from a number of ripples around the crossover frequency for this reason. I'd recommend the use of a baffle diffraction simulator (I later used The Edge to avoid this potential problem). The crossover removed most of the effect of the lower dip, and as physics dictates, the upper dip becomes insignificant off-axis.

Enclosure Design

I didn't take any photographs of the box during construction. The internal volume was about 25-30 litres IIRC and simulated with winISD (don't take my word for it, just do your own simulations). There is a single flared port on the rear of each speaker, they are placed away from the wall. Length is 13.5cm (to match whatever the flat tuning was), flaring was done with a router. All panels were 19mm MDF thick, except the front, which was 38mm. The speaker was internally braced with a honeycomb structure (12mm MDF with lots of holes, took out most of my thumb on the drill press). The outside was veneered with very, very shit iron on veneer. I now have a huge supply of the proper stuff, which is much better. The finish was oil, which ended up looking pretty shoddy due to the iron-on veneer. Thanks to the extensive bracing, there is no room inside the enclosures for the crossover, so it sits on top (allowed easy tweaking over a few months)

Crossover Design

Decided upon a 2nd order network based on what others had done at the time. Figured the seas tweeter should handle quite a low crossover frequency, improving the off axis frequency response and neting various other less significant benefits. The main disadvantage is that the tweeter recieves a greater amount of power.

The impedance and frequency response of each speaker was measured with speaker workshop. To simplify things the woofers were treated as a single speaker. The network was simulated prior to construction. The woofer consisted of baffle step correction and a second order low pass filter. I'm a bit vauge, but I think the woofer inductor was tweaked to counter the peak in it's response. The tweeter was just a second order high pass and a voltage divider (to match the levels.) Subjective listening tests combined with measurements and some feedback tweaked the final crossover design. The small capacitor was in parallel with the big one on the tweeter to get the value I wanted. The inductors were hand rolled and measured during construction to get very precise values.

Gallery

The measured response includes measurements every 30 degrees off axis (Pink on axis, red 30, blue 60, green 90).

One Speaker

Measured Response

Simulated Impedance

Simulated Response