DJW said:
I want to make up a cable to go between my sound card (sound blaster
live PCI)and my CD-ROM drive. I know the right and left channels wires
are the furthest out to both sides of the plug. And I am fairly
certain that the ground (common) is the center two wires on the
connector. I did a quick look to see what was available on ebay and
from the photos of cables it looks like they only use three wires.
My question is can the stereo hook up shared common leads? So only one
needs be wired to ground to be able to do stereo? And if so does it
matter which of the two grounds is used. Also if I were to make up the
connection with four wires would it matter anyway?
I have two optical drives here, that came with CD-audio cables.
Each cable has three wires.
*******
This is from the ALC650 datasheet. (That is a motherboard audio chip,
one that interfaces to a four pin CD audio input, amongst other things.)
http://img62.imageshack.us/img62/1354/cdaudio.gif
The chip (an AC97 codec used on motherboards), actually has a three
pin interface on it. It has CD_L, CD_R, CD_GND. The ground is only
on one end (I think, the drive end). Where CD_GND is on the AC97 chip,
the connection is AC coupled.
The idea behind this, has to do with noise. The CDROM audio cable
functions as an antenna. All the wires pick up noise. The noise is
"in-phase" and has identical amplitude on each wire. Even the ground
wire, has the noise imposed on it.
When the three wires arrive at the CODEC chip, the differential amplifiers
are present. (CD_L minus CD_GND) is the math the left channel diff amp
provides. Say there was a 1 volt noise sine wave on both of those wires
(i.e. on the ground wire and on the left channel wire). When the differential
amplifier subtracts the two signals, the noise cancels out. And it cancels
out, as long as the noise coupled into each wire is equal. It the noise
in one wire, has a higher amplitude than the noise in the other wire,
then the noise will "leak through".
CD_Player CD_L --------------||------- AC coupled into
CD_R --------------||------- the dual differential
solid earth ---> CD_GND -------------||------- amplifier circuit
on this end
Now, if you had four wires total to work with, and twisted together
CD_L with a CD_GROUND_L wire, then a more precise differencing could be
done. That may have been the original intent. But the three wire thing
does much the same thing, as long as all three wires are co-located
(close to each other), for best matching of noise signal levels. And
using only three pins, on the chip, saves a pin. So even if you used
a four wire cable, the three pin interface on the chip, may remove
any benefit it might have provided.
As the other posters note, using DAE (digital audio extraction), bypasses
the need for audio cables. But this doesn't work on older optical drives.
A newer drive should support this tick box. The advantage of DAE is,
in theory there is no noise. But I'd still want some way to verify,
that I'm making an exact copy. It isn't guaranteed, that all drives
rip the same way.
http://h10025.www1.hp.com/ewfrf-JAVA/Doc/images/c00220931.jpg
Paul
