In , on 10/15/03
at 11:19 AM, "John Templeton" said:

I've been working on tube amps for several years and always use a
dummy load when troubleshooting. I'm getting a Dynaco ST-70 back into
service and have all the required manuals etc.

What confuses me is that the book has very clear instructions for
testing the amp and biasing but *nowhere* in the instructions does it
mention hooking up a load before testing. I am confused by this. They
also do not mention running a test signal through the amp or examining
the output on a scope.

Can somebody tell me why/if they don't mention loading? Many thanks.

Most of the early tube amplifiers were designed by engineers who
learned their craft before the full math of feedback was taught at the
undergraduate level. Without this math it is difficult to guarantee
that your amplifier design will not break into oscillation at (then)
unpredictable times. The designer tinkered with the feedback till
things seemed OK or the boss insisted that the product go to market.

So many of those tube amplifiers would "run away" (break into
oscillation and sometimes physically burn) when operated into an open
circuit, everyone became gun shy and always loaded the amplifier. I
kept a pair of resistors in my toolbox -- just in case I had to operate
a system where the speakers might become disconnected.

The feedback math filtered down to the undergraduate level about the
same time as transistors became common. I'm not aware of any modern
amplifiers, tube or transistor, that have problems operating into an
open circuit.

The ST-70 was a relatively late design and David Hafler was sharp. The
ST-70 may be stable without a load.

----

While the "new" engineers' amplifiers were unconditionally stable,
there was a feedback nag that was not fully straigntened out till the
early 1980's. In the consumer market the wrinkle is known as TIM
(Transient Intermodulation Distortion). Now that they could easily
control it, engineers were using more and more feedback, which resulted
in amplifiers that measured lower and lower distortions, but somehow
sounded worse and worse (sometimes).

In retrospect, it was a simple oversight, but it took a bit more math
to make the issue clear to the design engineers. (If one looks in the
cracks, one will find a few earlier, isolated voices who were
empirically aware of the issue, but didn't formally prove their point.)


-----------------------------------------------------------
SPAM:
wordgame:123(abc):14 9 20 5 2 9 18 4 at 22 15 9 3 5 14 5 20 dot 3 15
13 (Barry Mann)
[sorry about the puzzle, SPAMers are ruining my mailbox]
-----------------------------------------------------------