Cable differences are real

[mansr]

1 minute ago, Arpiben said:

Just by curiosity did you measure both directions of AQ  'pseudo' directional  RCA cable ?

We'll get there in a bit.

 

1 minute ago, Arpiben said:

Not defending AQ,but analogue RCA cables are not supposed to work at Mega Hertz frequencies or have same bandwidth has coaxial.

They don't need to work at high frequencies, yet most of them do.

[mansr]

Now that we've familiarised ourselves with the cable impedances and signal reflections, we are ready to take on the AudioQuest cable. We begin by sending pulses into the cable with far end open, observing "correct" orientation according to the arrows.

 

 

What the hell? That's not how it's supposed to look. The first reflection is strongly attenuated, and then the second one is stronger again. Clearly, this is no normal cable. Turning it around, feeding it in the "reverse" direction yields this result:

 

Well, that's more like what we expected, a strong first reflection after about 10 ns followed by some minor wiggle. Apparently there is something asymmetrical about this cable. The only way to find out more is to crack open the connectors. Doing so reveals two insulated conductors connected to the pin and sleeve of each plug. Additionally, a shield is connected to the sleeve at the "output" end of the cable only. My guess is that the dangling shield causes additional signal reflections not seen with a properly constructed cable. In the reverse configuration, the shield and the main conductors are open in the same end, giving it a more normal behaviour.

 

Configuring a second scope channel to 50 Ω termination (same as the spectrum analyser) and connect the other end of the cable there, we obtain this image:

 

We note that the propagation delay of the pulse is 6 ns. The signal speed is apparently a bit slower in this cable than is usual. Furthermore, the received pulse amplitude is much lower than that at the source end, the "missing" energy having been reflected back. Indeed, 6 ns later, a negative pulse shows up at the input, and then a positive pulse at the output after another 6 ns.

 

For comparison, the same measurement of a 2 m RG-59:

 

Apart from the longer delays due to the extra cable length, the reflections are much smaller than with the AQ cable. From this we can conclude that the characteristic impedance of the latter is higher than 75 Ω.

 

A longer cable would have been preferable for this test as that would have provided better separation between the various pulses. However, this is what I had (didn't pay for it), and I'm not curious enough to buy another one.

[mansr]

16 minutes ago, d_elm said:

@mansr, it would be interesting if you could measure some popular configurations of the Lush^2 USB cable.

That would be even more pointless than this exercise.

[mansr]

9 hours ago, d_elm said:

What do you mean ?  

Measuring analogue audio cables in the 100 MHz range is already rather ridiculous. A USB cable either works or it doesn't, no matter what characteristics it possesses.

[mansr]

I forgot to mention, while measuring the AQ cable, I noticed it is sensitive to mechanical manipulation. Merely nudging it alters the waveforms quite a bit. The coax cables are, of course, impervious.