Lightning

Driver Choices

I wanted to build some speakers to handle from 80hz up for parties which would compliment the output capabilities of my subwoofers. I also wanted to keep the cost relatively low, since it was likely that they'd be abused and I didn't want to replace expensive components. I chose a P-Audio Challenger 12in P12-C300MB for the woofers. They were of reasonable cost, impressive build quality, high sensitivity and reasonable power handling (300wrms). I got them from www.cannonsound.com.au/. I looked at compression drivers. Most of them were rediclously priced for what they are. After buying a compression driver you also need to select a horn and pay for that too. I decided on a pair Peerless 850439 tweeters. The sensitivity was reasonably high etc. The main downside compared to a proper compression driver was the lack of directivity at lower frequencies. The result was fairly ragged off axis response, but it wasn't worth doubling the price of the speaker to fix it.

Baffle Design

The woofer was placed at the bottom. The tweeter was placed at the top. The position of the water didn't really matter as there were minimal baffle related effects within it's passband, excluding baffle step of course. The tweeter's position was optimised using the edge (see pictures at the end) such that the response was pretty much flat down to 1500Hz or so (note the 1db scale on the graph. Enough room was left at the top of the enclosures to allow placement of compression drivers should the horns be a failure.

Enclosure Design

The box was wedge shaped to facilitate use as a foldback or whatever, should it ever be required. It was 45 litres or so and ported to provide a fourth order roll off at 80Hz. You can work it out with winISD, I don't have them anymore so can't measure the port. I carpeted the boxes after construction to provide a relatively durable finish. There was a single brace across the two largest panels. See pictures for the rest.

Crossover Design

I kept it simple. A second order high pass was used on the tweeter at 3khz or so to work with the natural second order roll off of the woofer. The voltage divider was tweaked until things sounded reasonable. A PTC was later placed in series with the tweeter. The inductors were hand rolled and measured during construction to get very precise values.

Gallery

I can't seem to find the frequency response data unfortunatly. I did measure the final result at one stage. It was relatively flat, with a fairly significant dip at 2-3 khz off axis.

Partially Constructed

On top of subwoofers

Crossovers

Bad Carpet macro

Simulated effect of baffle on tweeter response