Originally Posted by rock_bottom
For an example of what is allegedly the result of "design by listening", see this interesting blog entry about a NuForce DAC
. The analog section of the DAC clips before the digital signal reaches full scale! That's just flat-out incompetent design.
There's no doubt that measurement is a far more efficient and reliable means for finding technical faults, and therefore avoiding or correcting them. It seems to be just common sense that a product has to be free of obvious measurable faults that can easily be audible.
I still remember when the golden ears were oohing and aahhing over electronics designed (by ear) by a young designer named Andy Rappaport (sp?). His backers sort of fled the scene when his power amps allegedly started burning their customer's houses down!
At the risk of walking on gravesights, I had more experience than I would
like with Mr. Rapaport when I was a high-end dealer back in the '70's.
His main product back then was a smallish, minimalist preamp.
Operationally, it was a disaster, huge bangs when switches were moved,
and, despite some uninformed participants in this groups claims to the
contrary, it EASILY sounded pretty lousy. Its cosmetic finish was crude,
at best. Inside, were black "bricks" containing the actual circuitry.
Finally had the opportunity to measure one, and it measured pretty awful
as well. Even on the easy stuff, like conformance to RIAA contour, it was
Despite that, a significant portion of the buying cognizeti at the time
thought Andy was just the darling designer.
He was prone to make some of the most ridiculous claims. For example, he
made one statement about the rear enclosure of the Quad ESL causing
severe box resonances. Well, he made that claim without having a clue as
to what a Quad ESL was, because he did some serious backpeddling when I
showed him a pair, minus the rear enclosure that it never had.
He was obnoxious, ill-informed, basically considered all his opinions as
universal fact. He claimed his preamp circuitry was the result of decades
(or some such nonsense) of research, which was interesting considering he
was something like 22 at the time, meaning, I guess, he was weilding a
soldering iron when he was 2.
One day, I got so pissed at his blustering, I took the demo he had
provided off the shelf and into the lab. With a combination of a small
ball-peen hammer, a cold chisel and some methyl ethyl ketone, I proceeded
to open up his magic bricks.
What was inside? Nothing more than a Motorola dual 741-style opamp
surrounded by the IDENTICAL circuit found in the Motorola apllications
note of two years previously INCLUDING the mistake in the original
printing that lead to the awful RIAA performance. What was missing was
reasonable high-frequency compensation to take care of the opamp's lousy
phase margin at about 20 kHz and any way of correcting the fairly high
input offset current and voltage. There was no innovative design, there
was no magic, it was pure plagerism, and plagerism of one of the poorer
examples of public-domain design at that. The lack of DC offset
compensation led to a much as 3/4 volt of DC floating around so that when
you switched input, you were switching pretty hefty damend voltages,
generating pretty serious pops to the power amp.