> So, here is a case where an industry found it desirable to describe signal
> voltage in decibel terms without requiring that the voltage be measured
> across the same resistance as the implied reference: 600 ohms.
That's right George. With the proliferation of transformerless audio
designs during the late '60s and '70s, it no longer became necessary to
"power match" the source with the load. For decades, dBm (as previously
noted is the equivalent of 1 mW across 600 resistive ohms which nets 0.775
volts AC RMS) was used since the majority of line-level audio circuits
demanded a 600-ohm termination. Due to the highly reactive nature of the
transformer, it was (and is) necessary to properly terminate the secondary
of the transformer for optimal transient response, but also...matching the
designation load with the transformer's source impedance was critical in
older power-matching circuits.
When solid-state electronics began replacing transformers (e.g., discrete
and LSI op-amps), the source impedance of audio circuits plummeted to near
zero. As a practical matter, even the best audio op-amps require a nominal
resistance at their output.
Approximately twenty years ago, Richard Cabot of Audio Precision Corporation
drafted a white paper on the ideal output impedance of an audio circuit.
Interestingly, it is not near zero ohms. Rather, under long audio paths,
the audio cabling begins to take on transmission line qualities. And if I
recall correctly, the optimum output impedance for most line level
applications is...near 50-ohms. There is a direct correspondence between
the optimum audio output impedance and how we derived at 50-ohm transmission
lines. Another topic of discussion which is OT to this reflector is
exactly why do we have 50, 52, 70 and 75-ohm RF transmissions lines. AT&T
never opted for 50-ohms on their landline audio circuits and for some reason
settled on 150-ohms as evidenced by their 600:150-ohm "repeat" coils. How
many of us remember the Western Electric 111C transformer? I imagine there
is probably some history associated with the 150-ohm twisted-pair circuit...
Today, we can pretty much load the output of a solid-state device without
much worry. Provided that the load impedance is at least ~ 10 times greater
than actual source impedance, dBu is a relevant standard....we no longer
care that it is 0.775 volts across 600 ohms...we only care that 0.775 volts
is capable of being placed across the destination load. Most audio studio
equipment is NEVER terminated to the source's low output impedance. Nearly
everyone bridges the load with ~ 10K-ohms or greater. In a transfomerless
environment, we no longer care about transient ringing and overshoot when a
complex wave is impressed upon a reactance as is the case with transformers.
Even the best Jensen transformers must be terminated for optimum transient
response.
Circling back....dBm had its roots with 600-ohm circuits. Today dBm is used
universally to represent 1mW across any load. As an example, commercial
lasers are almost always specified in dBm, irrespective of the actual
terminating impedance.
-Paul, W9AC
|