Construction of the enclosure
The woofer/midrange driver is a Peerless HDS Nomex 832873 (= 830873). The previously used 850488 is no longer being manufactured by DST/Tymphany and inventory of this excellent driver is depleted. The 832873 is its drop-in replacement for PLUTO.
An alternative woofer/midrange driver was developed by Seas for this application, L16RN-SL (H1480-08). It is the preferred driver when PLUTO is used without additional subwoofers. It can be purchased from Madisound. A few circuit component changes are required for its use. They are listed on the PLUTO owner page.
The tweeter driver is an AURA NSW2-326-8A. It may be difficult to obtain at times, but I do not have nor know of a substitute. It was actually this tweeter that inspired the PLUTO design.
The use of pipe K as enclosure for the woofer has advantages over a closed box. A pipe is inherently a very rigid structure and will not contribute secondary sound radiation from its surface. This is very difficult to avoid with a box construction where often wall vibration modes store energy and radiate more sound at certain frequencies than does the speaker cone. Pipes have a serious problem, though, in that they form acoustic resonators whenever their internal length is a multiple of a quarter wavelength. For PLUTO this is at multiples of 90 Hz, but can be well damped with stuffing as seen in a terminal impedance plot for the woofer.
It took some effort arrive at this result. The first prototype had a coloration of sound that could be detected with some broadband signals even though I thought the impedance curve looked acceptable. When I checked in the time domain the response to a 4-cycle Blackman windowed burst with the microphone very close to the dust cap of the woofer cone, I saw the picture at left. The delayed reflection off the rear end of the closed tube K is clearly visible as it is transmitted through the cone. Experiments with partially opening the rear of the pipe did not lead to improvements.
A pipe that is closed at its far end has an extremely high acoustic terminating impedance, p/v, which presents a large mismatch to the pipe's own acoustic impedance. Openings at the end lower the terminating impedance, but also make it reactive due to local acoustic mass. The best result was obtained by leaving the pipe closed at its end and using just the right amount of stuffing material to absorb and not to reflect the rear wave coming off the cone. As seen in the right picture the reflection is now over 40 dB down. This represents acoustic waveguide propagation with over 40 dB return loss or less than 1% reflection. The long length of the pipe allowed for forming a very good acoustic resistor. The earlier sound coloration is no longer noticeable.