[BC] Audio Cable

[Tom]

Not wanting to contradict Mr. O... but umpteen years of hands-on 
experience says at audio frequencies, the few inches difference between 
the pairs when employing the break-out described is negligible in terms 
of audio propagation.  And properly constructed, L and R in the same 
connector is also a very minor contributor to crosstalk..

Consider the wavelengths for audio frequencies.  Even 20 kHz is an 
extremely LONG wavelength... I will agree that an IMPROPERLY-constructed 
cable will not perform to desired specs - but the fault is not in the 
connector(s)...

Now, regarding bridges, and low-Z outputs feeding high-Z inputs:  
SOMEWHERE along the line, there should be a terminated input. That much 
I'll agree with.  I've never experienced any significant crosstalk when 
PROPERLY bridging a terminated line - paying attention to both phase and 
LOADING.  A 600-ohm audio line can be bridged ONCE with a 10-k ohm load; 
rule of thumb is, as I was instructed many, many, many moons ago, bridge 
= 10x load; a second 10-k bridge across the same line would make the 
bridging load = 5 k ohms, which is less than the 6k minimum.  Multiple 
bridging of the same line requires higher impedances - 50 k ohms allows 
for multiple bridges (up to 8, if I did the math right in my head); the 
other concern here is, of course levels.  Most of the lines I've bridged 
over the years are program-level - +4 dBm to +10 dBm - and terminated at 
the primary load (telco, STL, Optimod, whatever).  If you stop to 
consider, though, what is a distribution amplifier?  The basics are an 
input amplifier, a resistive divider, and as many output amps as desired 
in the design.  IC technology generally puts the design for 
off-the-shelf DAs at some multiple of four; but there's no reason a DA 
couldn't have 25 outputs, for example.  It's just a matter of correctly 
matching the divider network and then supplying the correct number of 
output amps internally.  I've built out a DA with 10 outputs using 
off-the-shelf program amplifiers and a resistive divider, and had flat 
response out to the limit (20-20k Hz).  I've also gotten flat response 
WITHOUT the program amplifiers - same network; just remember that ONE of 
the loads must terminate the line.

Tom S. - further comments interspersed...

Cowboy wrote:
>  Wanting to be sure, so after doing a little fact-checking,
>  by consultation with Mr. O.......
>
> Wed Jan  2 22:43:50 2008
>    
> In recent years the industry has neglected that which
> used to be standard. Properly terminated audio lines.
> Recall we used to have 600 ohm matched loads.
> Today, we employ low impedance outputs and high
> impedance (bridging) inputs. While possibly convenient,
> the after effects cannot be ignored.
>
> High impedance inputs allow for greater crosstalk and
> ingress of foreign signals. This would ESPECIALLY be
> true of left and right channels being routed through a
> common connector. Disbelieve this? Remember old
> fashioned sinks with separate hot and cold water faucets?
> Today with a common spigot we can have any temperature
> of water from the cold to the hot. Read:Mixed signals!
>   
Poor analogy - at the sink, you WANT the mixed "signals" - you're 
DELIBERATELY summing them.  L+R!  Y-adapter!
> Most zip cord constructed of stranded wire employ twist
> of the same direction for both channels. This causes an
> issue with the velocity of propagation. Ideally the left channel
> should have a left handed CCW twist and the opposite for the
> right channel. This creates an equality where common mode
> rejection is greatly enhanced due to the inherent cancellation
> of the common mode induced signals. Of course, this does
> increase the manufacturing cost of the cable. 
True; the solution, of course, is to use 100% shields between pairs so 
the crosstalk can't readily occur.
> We must also
> consider the connector assemblies. The suggestion of using a
> marker is Ludacris. This can adversely impact its shielding
> properties and capacitance, thus affecting the audio.
>   
I've never seen this in real-world situations, at least not as a 
measurable effect on the audio - again, we're talking 100% shields - 
XLR-type - connectors or DIN plugs with metal shells that bond to the 
drain (or the chassis).  You might as well add in the capacitances 
caused by the residual oils left on the shell by your fingers; there may 
be some effect, but it's NOT going to measurably affect the audio.
> The negligence in these matters no doubt arises from the
> fact that audio engineers and designers have had no experience
> with RF. Properly terminated transmission lines are essential
> for efficient and clean transmission. If we used the low-Z
> out (a tx) feeding a high-Z load directly (a tower) we would have
> high VSWR, poor bandwidth, etc.
>   
Agreed; the lines must be properly terminated.  I suspect the trend 
started when consumer gear started appearing in pro studio situations; 
they hook it up, it works, so it must be OK.
> Whether it be audio or RF they key word here is "transmission".
> Attention to line length, composition and connectors is very
> important to ensure square wave performance, symmetry and 
> quality reproduction.
>   
Semi sidebar - the only time I've had real problems with unequal lines - 
SIGNIFICANT problems, that is - is in dealing with a certain telco - 
four stereo pairs, no two lines of which were even in the same CABLE.  I 
could NOT find one pair that would maintain stereo separation... until I 
broke the stereo linkages at the transmitter, so I could find two pairs 
in the same cable.  Which proved impossible; they were routed through 
different cables between the studio to the CO, and different between the 
COs, then different again for the last mile connection.  Example; pair 1 
might be in an aerial cable between the studio and the CO, with pair 2 
underground (repeat using second and third aerials and undergrounds) At 
the CO, pair 1 would switch to an underground cable, while pair 2 might 
be underground - in a different cable.  At the midpoint, the 
relationships changed again, and again at the final CO.  Hairiest proof 
I ever did...

Tom S

--