newton2@topaz.berkeley.edu (08/24/86)
In the late sixties, I started a sound reinforcement company and quickly discovered that even I could design/build equipment far more suited to the emerging requirements of post-Beatles sound systems than could be bought from the likes of Altec. (Actually, there really *wasn't* anyone but Altec making this stuff- maybe Langevin and the remnants of the Ampex-JBL-Cinemascope rpoject, and other film-industry derived stuff, but otherwise nothing advanced beyond what Bell Labs had designed for the talkies and record-cutting). It wasn't just a matter of attaining higher power levels- it was the fact that you could build a twenty-input mixer that was quieter than a five-ton (mostly iron) recording console for less than the cost of a store-bought stepped ladder fader for a single input. Anyway, building high-power (and indestructible) solid-state amps was of course a priority- one needed dozens, basically one-per-driver in the brave new world of hi-fi sound reinforcement. For tyro designers like me, the then-common output and driver devices were a problem- they were restricted to low-voltage regimes and therefore low (100 watts into 8 ohms with 2N3055's) output powers (given output transformerless design for all the obvious cost, quality and elegance reasons). Worse, you couldn't get power transformers wound for you locally in the heavy currents necessary for power several such low-voltage output modules from a single bipolar DC supply. So the question of how to build r eally hefty power amps was always an issue (we didn't wanna just buy the Crown DC300s that were just becoming proven- no $$). J ust then I became aware of a San Diego company that made kilowatt solid state amps for driving sonar transducers. I was most impressed with their careful, systematic analysis (what I'd now recognize as quintessentially engineer-like) of what were truly the worst-case fault modes their amps might encounter, and the circuit topologies they'd adopted to cope. They dealt with not just roadies who shorted speaker cables (or the undersea nuke-war equivalent), not just totally reactive loads, but the case where another transducer actively drives power back into the victim amp in the worst possible phase-way. Oh, yeah, the point: They directly rectified the three-phase (400 Hz, i guess) line power to generate the DC supplies, using a rectifier bridge two components (actually, however many are appropriate to the star or whatever configuration the were using) of which were SCR's, so the DC supply was efficiently regulated. BUT: they used a humongous OUTPUT transformer (many kilowatts at up to 100KHz) to couple to the load! Not a very cost-effective tradeoff for an audiophile these days. It was a lesson to me, though. Plodding through the careful engineering design of the whole amp, I learned that only dilettantes and amateurs would decree a priori that "you can't get x type of performance using y (e.g., transformers) type design-- an engineer, even in a field as close to the ultimate limits of human knowledge as --gasp-- *audio*, simply asks "What are the performance requirements", and then tries to build something that doesn't violate any of the constraints. By the way, I suppose you could safely build a line-powered "powered speaker". that is one where the amp was safely integrated into the speaker cabinet, but then you'd still be wise to isolate the input. Now *I* would be happy with a well-laid out instrumentation amp on a clean circuit board, but nervous nellies and product liability insurers would probably be happier with a transformer. And the kind of transformer that nut.audiophiles would accept would take PhilR twenty megabyte-long postings just to extol, and who knows how long to wind.....