Re: [ARSCLIST] FW: [ARSCLIST] Internet audio: What do you expect of it ?

[Scott Phillips <scottp@xxxxxxxxxxxxxxxxxxx>]

...just referring to PNP/NPN junction transistors as a group... :>)

I also hand built low power transistor amplifiers back then with similar
specs, but building one then with good specs and 'high' power was really
tough. Even the definition of high power was orders of magnitude
different. As time went on we also found ways to measure distortion
products that we didn't recognize before. It sure helps to know what you
are trying to minimize when you design things. Yep, tubes make lousy
transistors, and vice-versa. In those days tube designs had many years
of development behind them, and those lessons didn't apply very well to
transistor designs. Takes time, doesn't it..

-----Original Message-----
From: Association for Recorded Sound Discussion List
[mailto:ARSCLIST@xxxxxxx] On Behalf Of Mike Richter
Sent: Wednesday, December 19, 2007 1:31 PM
To: ARSCLIST@xxxxxxxxxxxxxxxx
Subject: Re: [ARSCLIST] FW: [ARSCLIST] Internet audio: What do you
expect of it ?

Scott Phillips wrote:
> A lot of the early transistor designs were attempts to use basic tube 
> circuit principles with transistors. Regular bipolar transistors 
> aren't all that linear all by themselves, they required new ways of
thinking.

I'm not at all sure what is meant here by "bipolar" (the transistor is a
three-terminal, continuous current amplifier of course), but let me
support the general argument.

Fresh out of school in June, 1960, I went to work for a small company
designing instrument servomechanisms. They had a solid-state amplifier
of their own design, the best then on the market, constructed to
apologize for the transistor's terrible performance as a vacuum tube. 
With no engineering background and only the GE transistor manual to
start with, I designed and had built a "High Fidelity Servo Amplifier" 
(the title of the paper I wrote on the design). THD under 0.1%, flat
from the cutoff frequency defined by the output coupling capacitor to
too darned high a frequency. We added a capacitor to roll it off before
100 KHz to prevent parasitic oscillation with external parts. This was
before integrated circuits (the TI Darlington cans were just coming
out), so with discrete transistors it occupied one cubic inch compared
with well over ten cubic inches (and more than ten times the cost) for
its predecessor.

It delivered two watts of clean power and sounded wonderful connected
directly to Sennheiser headphones with their 600-ohm impedance - and not
bad driving a reasonably efficient speaker until they hit clipping. The
design could never have been implemented in tubes - but we were no
longer designing for vacuum tubes. They were very poor transistors.

Mike
--
mrichter@xxxxxxx
http://www.mrichter.com/