> > My theory is that when we turn it on and calibrate it, the
> > amp is cold and behaves as such.  After it's been on awhile,
> > and especially after it's been in use, it warms up and the
> > actual transistors used to do the amplification start to
> > allow a larger current into the load, as a warm transistor
> > will do (right?  That's why parallel BJTs are bad:  one heats
> > up, conducts more, heats up more, conducts more, etc.,
> > until one has 95% of the load and POOF!  Do I have my
> > facts straight?)
> 
> Yes, but that doesn't really count:  Your amp surely uses negative
> feedback, so it's the ratio of two resistors, rather than individual
> transistor parameters that determine the gain.

One would hope.  But then, one would hope that repeatedly returning
a volume knob to the same spot on an amplifier of reasonable quality
would produce the same results, as well.

> You can easily measure if the amp is at fault by measuring the AC
> voltage at the amplifier's output.  85->95 dB is a ten-times change in
> output power, or a 3X change in output voltage.  There's no way your
> amp is changing in gain by 10dB.  And your signal source shouldn't
> either, if it's at all good (test by watching the voltage at the input
> of the amplifier)

We've watched the signal source, and frankly, I doubt an Agilent
signal generator is going to have much drift in frequency/amplitude.
I'm aware of the scale of the jump here.  One or two dB I could buy
into, and accept as imprecision in the position of the volume, but
8-10 is unbelievable.

> My theory is something is changing acoustically in your setup. Are you
> moving things around in the room?  Maybe adding some reflective
> surfaces near the transducers?

Impossible.  The "room" is a 1 meter (approximately) cubed box,
built to be (within reason) acoustically isolated.  Things aren't going
in and out of it.  The calibration is performed under reasonably
repeated circumstances (meter similarly positioned, doors closed,
etc.).  Besides, the difference in amplitude is easily perceived outside
the box, and if it were a measurement difference, caused by a "hot
spot" inside the box where the echo was *just right*, the tone would
be the same to an observer outside the box even as the measurement
changed with moving the meter.

I still haven't ruled out observer/user error, since of five similar setups,
only one is really having this problem.  However, a bad amp could
cause that as well.

As to buying some high-end professional amps, that's not completely
out of the question.  It would be a tough sell, though, and some of the
features of a home stereo make it very well suited to this indeed (for
example, balance control and multiple output channels [front/rear]).

Mike H.
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist