Rather than try and redesign existing projects to incorporate unshielded drivers, I decided to develop a project for those of you wanting a very inexpensive DIY home theater system that provides excellent sound and is completely "home theater compliant", meaning that it uses only shielded drivers, and if properly finished meets the various "spousal approval" parameters that tend to muck things up a bit. As such, Part I of the Dayton Home Theater Project is the use of 5 identical small, sealed enclosures covering the frequency range of 100hz to 20khz. Since these speakers have an internal volume of about 6.8L (about .23 cubic feet), they can be integrated quite easily into most rooms with minimal intrusion. Moreover, they are small and light enough to be conveniently wall-mounted (and are design to take be placed 4"-12" from a wall without developing a "muddy" sound that many speakers designed for mid-room placement tend to exhibit.) Finally, I've tried to take some of the more tedious work out of this project by eliminating the need to flush-mount the odd-shaped 5.25" woofer, the Dayton 295-300.

Many of the most popular home theater systems on the market have achieved their commercial success due to having very small cabinets that are easily integrated into the room. I decided to target the smallest possible enclosure size with this project that would allow the drivers to not only gain high "spousal acceptable", but also to have sonic properties that meant the satellites were more than just small -- they're small for a reason. The DHT satellites have an internal volume of 6 L, giving them a cutoff frequency of around 88hz and, being sealed, a low-frequency rolloff rate of about 12 db/octave. What this means is that by relying on the high-frequency crossover of Dolby and DTS audio, which occurs in this frequency range at a rate of about 12db/octave, response should be relatively flat in the upper midbass region and no loss in performance would be observed in spite of the small, sealed enclosures.

The internal dimensions for the enclosure are 9" (H) x 6.5" (W) x 6.0" (D). Using 3/4" MDF, which is recommended to reduce enclosure-induced midbass coloration, its low cost, and to reduce cabinet vibrations, the panels for the enclosure can be cut as follows:

The resulting enclosure's external dimensions are 11.5" (H) x 8.0" (W) x 7.5" (D). The drivers are centered on the front baffle with a center-to-center separation of 5.5". I recommend flush mounting the tweeter, but the crossover has been optimized without flush mounting the woofer. Due to the odd-shaped frame of the woofer, this can reduce the difficulty of enclosure construction significantly. Further, this reduces the separation between the acoustic centers of the drivers, eliminating the need to correct for a large offset between the drivers in the crossover and providing better overall phase response.

Some of you will prefer the sound of these satellites with the woofers flush mounted, however. As designed, the ideal dispersion pattern for these speakers is +/- 25 degrees horizontally and +25/-15 degrees in the vertical plane. By flush mounting the drivers, the vertical dispersion performance shifts to +20/-25 degrees, making the speakers more suitable for placement above the listening position. Those who wall-mount their speakers on pivoting bases will be able to choose between flush mounting the woofers or angling the speakers downward into the listening area. As with any speaker, some experimentation and optimization should be performed when installing the speakers to determine the most desirable listening position and speaker placement.

The decision to optimize these speakers for performance along the woofer axis and above was made on the basis of the average listening room. Most people's rooms have quite a bit of "stuff" below the listening axis. In my own room, there are couches, tables, plants, children, and a few small animals (and even 95lb golden retriever). None of these things are conducive to better room acoustics, but none of them are likely to be "optimized" either. As a result, I chose to provide make one of the many "compromises" required during the design process in a manner that reflected the real-world situation. Based on listening tests, I believe this choice was well-considered, due to the smooth sound pattern these speakers develop throughout the listening room.

Rather than present the crossover for this project in the traditional, computer generated format. I decided to give you a small peek into the design process. During the design of these and other speakers, I kept a notebook of possible crossover designs. When I reached the final stages of the process, I would build and listen to each crossover and finally arrive at the best sounding version when considering complexity of construction and the overall parts cost. Normally, I use a computer program to draw this crossover. In this case, I've scanned the notebook page with my original testing notes (in this case, the extremely good impedance phase angle, something I think is very important for generating good performance when using inexpensive home theater amplifiers that can be very sensitive to difficult to drive loads. In this case, the speakers present an extremely easy load for any amplifier and have performed very well with my old, $300 Kenwood VR-309 as well as my current Denon 5800, which weighed in new at about $3500).

You probably notice that this is more complex than the average crossover and the parts count is not particularly low. Thanks to computer optimization, however, I was able to tackle a fairly difficult set of drivers to produce excellent overall response while maintaining an extremely good impedance profile for the amplifier. Further, careful computer optimization not only helped me improve the overall frequency response, but also helped me try various topologies that used less expensive parts, resulting in the great economy of the final design.

As shown to the left, the final frequency response is quite good from 1khz-20khz with a 3500hz in-phase crossover. Approximately 4db of baffle step rise is evident when the speaker is measured in half-space. When placed within 4"-12" of a wall, response if flat from 100hz-20khz +/- 3db and the overall sensitivity averages about 85db @ 1W/1M.