Reply
 
Thread Tools Display Modes
  #1   Report Post  
 
Posts: n/a
Default Digital Interconnect to DAC, but once inside...

Hi, all,

I have a question about digital interconnect design/application.

I have a Musical Fidelity Tri-Vista 21 DAC and I'm using a Magnavox
CDB-560 CD player as a transport. I am using a .5M length of Monster
M1000D digital cable as an interconnect.

I have always believed that any properly designed and well-constructed
cable should do a good job of transferring a digital signal, and the
Monster cable (about $45) seemed to me to be a pretty good compromise
between spending big bucks for a "premium" cable and still calming the
"audiophile" angst that I get whenever I read the audio mags and
newsgroups.

I have been surfing the 'net reading all kinds of info about digital
interconnects, and the general drift is that a digital interconnect
ideally needs to be constructed so as to maintain a nice, constant 75
ohm characteristic impedance, including the connectors (the connections
and connectors theoretically being the source of any transmission
problems).

I understand that it is difficult to properly manufacture an RCA
connector for this application. Some sources (proponents of expensive
cables mainly) warn of dire consequenses if there are tiny
irregularities in the insulation/shield diameter, quality of the solder
joints, etc.

NOW... I reallize that there is controversy as to the real importance
and audiblity of all this stuff, but my question has to do with what
happens to the signal once the signal physically enters the DAC!

The Tri-Vista 21 uses the usual RCA coaxial connection at the input,
but the connection between the connector and the circuit board is a 5
inch long twisted pair of 22-gauge wires.

What effect (theoretically) does the change to the twisted pair have on
the characteristic impedance?

Wouldn't the change from coaxial cable to twisted pair have more of an
effect on signal quality than the short length of "improper" shield
diameter going through an RCA connector?
  #2   Report Post  
Stewart Pinkerton
 
Posts: n/a
Default Digital Interconnect to DAC, but once inside...

On 21 Oct 2005 05:45:38 GMT, wrote:

I have been surfing the 'net reading all kinds of info about digital
interconnects, and the general drift is that a digital interconnect
ideally needs to be constructed so as to maintain a nice, constant 75
ohm characteristic impedance, including the connectors (the connections
and connectors theoretically being the source of any transmission
problems).


That's correct in a qualitative sense, although seldom a quantitative
problem.

I understand that it is difficult to properly manufacture an RCA
connector for this application. Some sources (proponents of expensive
cables mainly) warn of dire consequenses if there are tiny
irregularities in the insulation/shield diameter, quality of the solder
joints, etc.


There will certainly be an impedance discontinuity, which will
introduce reflections. Whether that will disturb the datastream enough
to cause reading problems is much more arguable.

NOW... I reallize that there is controversy as to the real importance
and audiblity of all this stuff, but my question has to do with what
happens to the signal once the signal physically enters the DAC!

The Tri-Vista 21 uses the usual RCA coaxial connection at the input,
but the connection between the connector and the circuit board is a 5
inch long twisted pair of 22-gauge wires.

What effect (theoretically) does the change to the twisted pair have on
the characteristic impedance?


That depends on the impedance of that twisted pair - it may readily be
made to be 75 ohms by choosing wire with insulation of suitable
thickness.

Wouldn't the change from coaxial cable to twisted pair have more of an
effect on signal quality than the short length of "improper" shield
diameter going through an RCA connector?


Not necessarily, for the reason given.

--

Stewart Pinkerton | Music is Art - Audio is Engineering
  #3   Report Post  
Timothy A. Seufert
 
Posts: n/a
Default Digital Interconnect to DAC, but once inside...

In article , wrote:

Hi, all,

I have a question about digital interconnect design/application.

I have a Musical Fidelity Tri-Vista 21 DAC and I'm using a Magnavox
CDB-560 CD player as a transport. I am using a .5M length of Monster
M1000D digital cable as an interconnect.

I have always believed that any properly designed and well-constructed
cable should do a good job of transferring a digital signal, and the
Monster cable (about $45) seemed to me to be a pretty good compromise
between spending big bucks for a "premium" cable and still calming the
"audiophile" angst that I get whenever I read the audio mags and
newsgroups.


To an engineer, $45 for a cable already sounds ridiculously expensive
for this application. Please rest assured that your angst is just
angst.

I have been surfing the 'net reading all kinds of info about digital
interconnects, and the general drift is that a digital interconnect
ideally needs to be constructed so as to maintain a nice, constant 75
ohm characteristic impedance, including the connectors (the connections
and connectors theoretically being the source of any transmission
problems).


Designing in the digital domain is done, in part, to gain high immunity
to imperfections such as noise, impedance mismatch, and so forth. The
whole idea is to reduce dependence on exact analog levels and instead
allow for some slop. Well, more than "some", a lot of slop. Never
forget this when reading material which suggests that absolute
perfection is required for digital circuits, particularly at relatively
low data rates like S/PDIF.

Speaking of that low data rate... at 0.5 meters of cable, you truly have
nothing to worry about. Changes in impedance (and the signal
reflections they cause) really only matter when the cable (or printed
circuit board trace, or whatever) is long enough to operate in the
so-called 'transmission line' domain. A general rule of thumb is that a
cable (or printed circuit board trace, etc.) is a transmission line if
it's more than 1/4th the wavelength of the signal it's carrying. I just
did a rough estimate assuming a 5.64 MHz signal (S/PDIF transmits 32
bits per 16-bit sample, and uses an encoding technique which doubles the
signal bandwidth), and assuming that signals propagate at 1/3 the speed
of light (this is going on memory, but I think it's roughly correct).
Given these assumptions, 1/4 wavelength is over 4 meters. Issues like
impedance changes at connectors are very, very unimportant for a cable
so much shorter than the critical length.

I understand that it is difficult to properly manufacture an RCA
connector for this application. Some sources (proponents of expensive
cables mainly) warn of dire consequenses if there are tiny
irregularities in the insulation/shield diameter, quality of the solder
joints, etc.


The proponents of expensive cables are basically lying when they say
that these things are important to S/PDIF. S/PDIF was designed to use
common 75 ohm coaxial cables normally used for composite video. I
guarantee you that none of the standard's designers anticipated somebody
spending $45 on a 0.5m 75-ohm coax cable and thinking that they had
spent just barely enough to be comfortable.

NOW... I reallize that there is controversy as to the real importance
and audiblity of all this stuff, but my question has to do with what
happens to the signal once the signal physically enters the DAC!

The Tri-Vista 21 uses the usual RCA coaxial connection at the input,
but the connection between the connector and the circuit board is a 5
inch long twisted pair of 22-gauge wires.

What effect (theoretically) does the change to the twisted pair have on
the characteristic impedance?


The twisted pair is probably mismatched to the characteristic impedance
of the main cable, but this is not likely to have a significant effect,
as per above.

Wouldn't the change from coaxial cable to twisted pair have more of an
effect on signal quality than the short length of "improper" shield
diameter going through an RCA connector?


Most likely, but none of this is anything you should be worried about.

--
Tim
Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
common mode rejection vs. crosstalk xy Pro Audio 385 December 29th 04 01:00 AM
Artists cut out the record biz [email protected] Pro Audio 64 July 9th 04 10:02 PM


All times are GMT +1. The time now is 01:07 PM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"