The satellites were conceived as 6" two way speakers, utilizing the Foster/Rank ribbon tweeter from my monster dorm blasters built for college. The midranges were 6.5", four-ohm Peerless units. In the cabinets, the midranges have a smooth response down to 90 Hz, with a slight (< 1 dB) hump centered around 150 Hz. The woofers have a natural rolloff around 5 kHz. The tweeters have a response from 5 kHz up to around 30 kHz (as measured by ERD). A comparable Pioneer ribbon tweeter has specs listed as 6 ohms, strontium magnet.35/50w -Freq: 2000-30khz - SPL: 91db@1w/1m. The 2 kHz lower limit is misleading, as the units do not play well much below 5 kHz.

Crossovers were a "chance" design, which is one of the primary drivers for the redesign going on at present. The tweeter section is the 2nd order, 5 kHz section designed for the monster towers. There is an L-pad allowing tweeter level adjustment. The woofer is run full range; it's natural 5 kHz rolloff is at least kind of complimentary. The result is a mess, with upper midrange irregularities and cone breakup of the woofer the major flaws. For all that, the sound is surprisingly good at lower volumes, but it gets fatiguing and somewhat ragged when pushed hard.

Cabinets were constructed as sealed, 6.9 liter boxes. Front baffle is 1 1/2" thick, made by laminating two 3/4" sheets of MDF. Side and rear panels are 3/4" MDF, with multiple braces between panels. Top and bottom caps are solid oak. Binding posts were RS spring terminals, and a fuse holder was installed to protect the tweeter. The front corners were rounded over using a template and a lot of filing. Veneer is a plastic-backed African Mahogany, rolled right around the corners for a solid-block-of-wood effect. The back panel is also veneered. There are no grilles at present.

I learned a couple lessons while working with this veneer. First, when rolling out flexible veneer, work across the grain, not with it. This allows the veneer to stretch a little to correct directional errors. Second, with this kind of construction, the veneer should have been applied and trimmed before the endcaps were glued in place. I had to trim the veneer to fit next to the caps, which led to some really bad blemishes where the two meet.

REDESIGN PROGRESS

  The satellites are currently undergoing a significant redesign. There have been several changes over the years, including new binding posts (real ones) and new woofers. The foam on the original drivers eventually rotted, and instead of re-foaming I opted to buy better drivers inetead. A Peerless 831732 (now supplanted in the Peerless line by the 832732) was installed.

  The first task in redesigning the satellites will be construction of a proper crossover, and installation of new tweeters. For this, measurements of the drivers will be required.

  The 6 1/2" Peerless CC line 831732 woofer has the following spec parameters:

Used in the current 6.9 litre box with fiberglass stuffing, this results in a smooth response down to 90 Hz (Qtc approx. 0.63). These drivers are actually a better fit for the cabinet than the originals. Some experimentation with stuffing and filler will be performed, along with inclusion of DCR for the yet-to-be-designed crossover, to bring the Qt closer to .707. The electronic crossover will be redesigned to account for this rollof as part of a 4th order Linkwitz-Reiley crossover at 90 Hz, filling in an existing 80-90 Hz hole that exists in the system at present (the current crossover is at 100 Hz and is 4th order electrical for both sub and satellite, leading to response errors due to the satellite rolloff).

   The frequency response and distortion curves of the  831732 unit, as provided by Peerless, are provided in the right hand photo collection. An in-box measurement on-axis is shown just below that. These measurements were taken with a Radio Shack sound level meter (analog variety, non-modified), and corrected with the inverse C-weighting curve. Test signals were supplied by Speaker Workshop's sound funtions: warble tones of 1/8 octave were used for this measurement. Note that there is a huge peak at 5.7 kHz - just exactly what would cause the ragged and fatiguing sound. It is visible also in the full-range measurements (see Home System). Taming this peak may require including a notch filter in the new crossover, since it is only about 1.5 octaves above the target crossover frequency. A proper microphone/preamp combination (probably of the Wallen variety) will be constructed and used to verify these measurements.

   The current ribbon tweeters require a 2nd order crossover no lower than 5 kHz (the current configuration), so improvement in the crossover using the current drivers would require a third or fourth order alignment around 4 kHz, leaving the response peak of the woofer as problematic at best. Rather than invest time & money in mating these not-very-compatible drivers,  a new tweeter has been obtained and will be added to the system over the summer. The new units are Raven R1 ribbons, which can be crossed over at 2 kHz. Parts for an assymetrical 3rd/6th order (electrical) crossover have been obtained.

    Grille fabric has been obtained (from Parts Express), along with snap-in grille retainers. Acoustic ceiling tile, felt and foam have been procured for testing purposes. New dual binding posts (also from Parts Express) have been obtained to allow the crossover to be relocated external to the cabinets. Given that there are 5 coils, 6 capacitors and a handful of resistors for each speaker, I doubt it would be practical to try to place all the parts in the cabinets anyway.

    Additional progress on the redesign will be posted here as it happens.