Paul McGowan writes:
Designing any product for consumer electronics is a series of design choices – better known in some circles as compromises – giving one thing up for the benefit of another. Design choices are not unique to consumer electronics as compromise is just a part of the process of everything you can imagine: cars, food, clothing, medicines, recordings and so on. It’s just the nature of things.
As we’ve been focusing on amps for a while I am reminded of a design compromise we made years ago when we were developing the 200C power amplifier during the mid 1980′s.
The 200C was one of our best selling power amps of the day, reviewed by J Gordon Holt and Tony Cordesman, 200 watts per channel into 8 Ohms and double that into 4. We started with the intent of building a no compromise amp – our best effort ever and the amp did well. J Gordon remarked “This is, if not a world-beater, at least a world-class amplifier”. High praise indeed.
Some its design features included the use of solid copper bus bars to connect the power supply to the output transistors, including the output fuse in the feedback loop, JFET direct coupled input stage and a fully complimentary low feedback design. We went all out on this one.
Most of the circuit design work was a product of many years of work by Dr. Bob Odell, while I handled the construction and layouts.
We were darned proud of this amp and asked our friend Arnie Nudell, the founder of Infinity Loudspeakers, if we could travel to his home in LA and compare our amp to his best amp on the massive Infinity IRS system at his home.
After dinner and drinks we set down to do some listening and it didn’t take Arnie long to tell us that he felt the piano piece playing at the time was “closed in” and “lacked the upper openness” of his tube amps – and we had to agree – the difference wasn’t subtle.
Before we packed up the amp and went home with our tails dragging a bit an idea hit me. Dr. Bob had insisted on the addition of an output inductor that rolled the top end off at a very high frequency, something like 50kHz or perhaps even higher – clearly outside the band of human hearing. He relied on this inductor to keep the amplifier stable in a few rare cases.
I had always been a bit suspicious of this approach and grabbed a couple of clip leads Arnie had laying around and shorted out the inductors so they were no longer in the circuit. We played the same piano piece and instantly we all smiled. The veiled closed in sound was gone and we clearly approached the best qualities of Arnie’s prized tube amp of the day. Score.
It was clear from a sonic standpoint the filter on the output had to go – yet from a reliability standpoint it wasn’t a good idea because under certain circumstances the amp could go unstable and die – and when it died it wasn’t pleasant.
After much soul searching and investigative work, we decided to remove the inductor in production and put a big warning label on the back of the amp. That label cautioned users not to plug in or unplug the inputs when the amp was on – instead, shut it off before making any connections – a practice that should be followed with any amp under any condition. If users followed these instructions the amp was rock solid – but failure to follow them might result in flames coming out of the top. A risky gamble indeed, but we felt the ultimate quest for great sound trumped everything else.
That was the first compromise we made with the 200C and it turned out to be a bad decision – we perhaps lost 5% to 6% of the products due to customers failing to heed the warning – so in future production we slowed down the amplifier speed, made it safer, lost some of the openness that amp was capable of and called it good.
To this day I have often wondered if we made the right decision to back down on the amp performance to protect the few who messed up.
One way to look at it is we downgraded the instrument’s performance for 100% of the users to keep the 5% from having a failure.
Compromise in design is a tough call. If you were the designer and the technology of the day would not allow you to have your cake and eat it, which path would you have taken?