[Amps] The Pin One Problem

[Paul Christensen]

> One word ........ transformers.

Audio transformers can be a convenient way to break-up grounding and RFI 
problems (owing to generally high CMRR figures) when interfacing equipment. 
Field engineers often prefer transformer mic inputs in mixers since the 
device is often taken to temporary or remote locations where it's difficult 
to quickly control cabling and grounding issues.  But, it's still possible 
to have a "Pin 1" problem with a transformer or active balanced input - if 
the audio cable shield is connected to the equipment circuit ground rather 
than bonded at the chassis.

Good transformers are expensive.  But there's no reason why today's 
transceivers and audio accessory equipment cannot be designed using active, 
balanced inputs using a simple 3-stage instrumentation amp (discrete or 
on-chip) where CMRR figures approach a transformer across the audio and HF 
spectrum.   It costs the manufacturer little more to design this way.

For short and moderately long cable runs, it is much more important to have 
a high CMRR balanced input than a balanced output.  If the instrumentation 
amp is doing it's job, it doesn't care if the source is balanced or 
unbalanced - it's simply bridging a source between two terminals, 
irrespective of a ground reference.  This performance cannot be achieved 
with a single op-amp.  Although a single op-amp input can be labeled as 
having a differential input they cannot be called "balanced," and CMRR is 
never as good as can be achieved with the 3-stage circuit.

Getting pretty far OT here, but I'll finish with this thought.  As cable 
runs lengthen, they should be treated like transmission lines and the source 
Z becomes critically important due to distributed cable capacitance.  One 
broadcast equipment manufacturer has broken the sacred rule for shielding. 
They use a wire management system composed of hyper-twisted CAT5/6 cabling 
with tightly controlled source and terminating Z.  When CMRR and Z is 
controlled, there's little need for a cable shield to perform against most 
common forms of interference.  Bell Labs/AT&T knew this probably back in the 
1920s.  With the use of step-down/step-up repeat coils and twisted pairs, 
they survived decades without shield cabling, and often circuits were 
bundled tightly together and yet crosstalk remained acceptably low.

Paul, W9AC