IC Cable termination

[DF96]

Yes.

Don't worry, worrying about a non-problem is normal for a newbie; we have all done it, well I have certainly done it!

 

[HiFiNutNut]

Thanks, Andrew and DF.

You are right that I didn't know what I was talking about. Your notes prompted me to do a bit of reading and for the first time I found articles "lumped vs distributed". I certainly don't have a electronic background.

DF,

You suggested that the lumped elements analysis did not yield the truth at 42MHz and a distributed system must be used to calculate with a coaxial cable at RF.

I don't dispute that. But I found it interesting that most people discard using a distributed system to calculate anything with audio, including audio connection cables - interconnect and speaker cables. Well, audio frequencies are defined as 20Hz to 20kHz but audio equipment still must do things to suppress RF!

Now I am learning. The theory is that when tr/td > 6 --> lumped element and tr/td < 2.5 --> distributed element where tr is the rise time and td = length/velocity. While velocity is a constant, td is in proportion with length of the line. I often heard arguments that the line length is too short for an audio cable to be considered as a transmission line, and that fits in the above formulas.

In other words, transmission line theory doesn't apply, so I guess lumped elements should apply. In that case, should the simple LCR model work?

Regards,
Bill

 

[DF96]

In other words, transmission line theory doesn't apply, so I guess lumped elements should apply. In that case, should the simple LCR model work?

Transmission line theory, correctly used, always applies. However, often it is not correctly used for audio and almost always it is unnecessarily complicated when the simpler lumped elements approximation is good enough for practical purposes. Use lumped elements when you can (sufficiently low frequency); use distributed elements when you must (higher frequencies).

If audio people had a better understanding of RF then you would see a lot less nonsense written about transmission lines, but some have just enough RF knowledge to confuse themselves and impress others.

 

[HiFiNutNut]

What I got from the formula is that there are two factors that determine if we should use lumped elements or transmission line - (1) the rise time which relates to frequency and (2) the length of the wire. While at RF above certain frequency the rise time is fast enough, there is a second factor. Because audio cables are relatively very short, transmission theory does not apply.

If that is the case, my previous posted picture of the simulation of the coaxial cable using lumped elements should be correct.

 

[HiFiNutNut]

But now I still have not solved the original problem - why an audiophile cable my friend bought sounded substantially better (more details, more air, more sense of space, more texture of musical instrument, more transparent, more textures, etc) comparing to my Belden 1694F DIY interconnect cable?

So far my progress is that I should ditch the idea of "transmission line termination". As suggested, some simple R should provide damping to the LC of the interconnect cable. The R can be 100R or so.

When I listened to the cables, the source component (Marantz UD7007) has output impedance of about 150R. That should be very sufficient to damp any possible LC resonance of the Belden coaxial cable.

So what was it? Why the Belden cable didn't work as well?

I guess I have to go and measure the two cables for their LCR parameters before I try anything else.

 

[DF96]

If that is the case, my previous posted picture of the simulation of the coaxial cable using lumped elements should be correct.

Correct for audio frequencies, maybe. Not for calculating a 'resonance' at 42MHz!

But now I still have not solved the original problem - why an audiophile cable my friend bought sounded substantially better (more details, more air, more sense of space, more texture of musical instrument, more transparent, more textures, etc) comparing to my Belden 1694F DIY interconnect cable?

If two audio interconnect cables sound genuinely substantially different then at least one of them is faulty (possibly by design) or the equipment is faulty (possibly by design). Common problems are:
1. high output impedance leading to tone control by cable swapping (due to shunt capacitance)
2. poor screening leading to RF interference creating added 'sparkle' and 'detail'
3. instability, which is not always heard as such but its effects are usually audible
You should assume, in the absence of evidence to the contrary, that the more expensive the cable the more likely it is to be poor in an electronic engineering sense: unusually high or low inductance or capacitance, poor screening (or no screening!), microphonic. The same can apply to equipment.

The simple Belden cable probably works fine. What you need to investigate is what damage the other one is doing to the audio signal, and then ponder why you prefer the damaged signal.