These cabinets were built by the Heintzman Piano Company in the Junction neighbourhood of Toronto. Hard to tell exactly how old they are... probably pre-1960, possibly even pre-WWII. When I found them they had flimsy, clumsily made baffles that were clearly not up to the same level of craftsmanship as the shells, so someone decades ago had ham-handedly repurposed them. What was their original configuration? Church organ speakers? A PA system at a dance hall? Still in fantastic shape, they were ideal for the Saba Greencone project I'd been itching to put together.

Listening tests of the famous Greencones (20cm midbass: Permadyn 5298 U 8 and 10cm tweeter: 1670 CU 15)had been tantalizing. They had a bit of a high-mid/low treble shout and rolled off below 100Hz, but their subjective presentation was musical, detailed, lifelike, and involving. Astonishingly so. Brainstorming solutions to extend the bass, I had to rule out sealed and reflex designs because these have to breathe, as they did in their original semi-open or (distributed port?) implementation in Saba radios. I was almost ready to just fit them into the Heintzmans and live with the lean bass when I found the Saba 30cm units (5898,920,002). These don't quite go all the way down to 20Hz but definitely below 50.

The 1.4mH "filterdrossel" inductors came from a set of Telefunken RB46 speakers from the mid 1960s. Combined with a set of huge pop-can paper-in-oil capacitors (56uF), this created a gentle first order crossover between the 30cm Sabas and the 20cm Greencones centred around 580Hz. Now between them they could handle a little more wattage with a firm response. I used 2.2uF ERO MKT capacitors to cross over the Saba 10cm bakelite tweeters. The crossover components were mounted to the baffle on a panel of neoprene to reduce microphonics. After listening tests, I chose to bypass the 56uF with a further 5uF of Dearborn paper-in-oil, and bypass the 2.2uF with a further 1uF of polypropylene (MAXcaps). All wiring is milspec high-purity silver-plated copper with teflon insulation, with each set of leads suitably sized to its driver (30cm: 16AWG, 20cm: 18AWG, 10cm: 22AWG).

For grillecloth, I found some open-weave silk in the old garment district on Queen West. I experimented with perforated masonite pegboard for the back panels (similar to the old radio backs), and found that they still sounded better open-backed, as a U-dipole. I settled on velvet curtains to slightly tame the back reflection without impeding airflow. (The huge curtains these were cut from once hung behind the stage at the old Top O' the Senator Jazz Club). Minimal asymmetrical damping on the interior (the radios had none)... to measure acoustic deadness and resonances, I like to stick my head in and see how it sounds with my voice, knock the sides, etc. I'm not against the laboratory approach to speaker-building, but I am a believer in the care-oriented builder's instinct. To preserve the finish, I had 1/4" glass cut to fit the top surfaces.

Sonic impressions: rich and warm presentation. Voices especially sound incredibly real: watching an HDTV OTA news broadcast, you become aware in each segment not only of the unique resonance and timbre of each person's voice (head/throat/chest), but also of microphones, recording techniques, and environments. These speakers let you listen into the program material. They sound best with the tweeters on the outside. The stereoscape imaging is not as focused as the single point source Philips 9710M fullrange, but the Sabas beam less and are a bit more efficient, with a higher, silkier top end. (the Philips in their tuned bass reflex cabinets have a bit more "whump" in the 50-100 Hz region, which works well for reggae and dub).

*UPDATE: Stevie says their sound is "caramel", and listening to them is "like wearing a huge set of headphones"*

Click here to go back to the main page.