The impact of network cables: seven cables analyzed

We couldn’t resist another – perhaps now for the last time? – tinker around with the network again. This time, we asked Audioquest to send us all the network cables for analysis. And the results are quite interesting!

Let’s start by being clear about the expected impact of a network cable. Anyone who follows Alpha Audio knows that we 100% “believe” in the impact of cabling. Every link in a solid, transparent audio system does something with the sound. Whether it’s a mains cable, interlink, speaker cable, network switch or network cable…. it does something to the sound.

And that makes sense, because a cable transports not only an audio signal (digital or analog), but also noise. And the better a cable manages to keep noise out, the “quieter” and more transparent the presentation becomes. This is true not only for cables operating in the analog part of the system but also for cables doing their work in the digital part of the chain. And so this is also true for network cables.

The measurements

To see how much the impact of a network cable is, we took a few measurements. We put the Volumio Primo back in the measurement box with the active probe on the 100 MHz clock that drives the dac chip.

This measured clock signal goes into either the Aeroflex PN9000 or the Wavecrest SIA-3000. These devices both measure jitter, but in a different way. The SIA-3000 analyzes jitter in the time domain (and also measures high and low frequency modulation, as well as clock stability (cycle to cycle), the Aeroflex PN9000 measures phase noise. This is in the frequency domain.

We have explained in this article what exactly is what. So if you want to know more about the measurements and the meaning of all those numbers…. :-).

The third measurement is noise. We connected the various cables via the ‘Ethernet’ CDN (coupling / decoupling network) to the spectrum analyzer to see how much the cable now filters. Basically, you are measuring the noise from the switch via this setup (and thus also the noise from the adapter feeding the switch). But perhaps the cable is also filtering out some noise; that should show up.

The candidates

We received from Audioquest all the cables from their range: the Diamond, Vodka, Cinnamon, Pearl and Forest. All cables are three meters long. We added another homebrew (double shielded CAT7A with Hirschmann plugs) and a standard UTP cable. That makes for a nice comparison.

Noise

Let’s start with the noise measurements. There we see very subtle differences. But if you look closely, you can see that the Diamond and Vodka in particular show less noise. It’s subtle, but consistent. DBS on gives slightly more attenuation. The “Alpha” cable has more noise. This is odd, because in the other measurements it is not really worse.

Wavecrest measurements

Although the differences are small, they are again reproducible AND consistent. The Vodka and Diamond score well. So does the Alpha cable here, but not really in the phase noise test. We still have to investigate that. We estimate that the SIA-3000 measures differently. And frankly, we trust the phase noise measurement more in this case, because the noise is also high on the spectrum analyzer.

The Vodka and Diamond score quite structurally better than the Pearl, Cinnamon and Forest. DBS sometimes works, sometimes it doesn’t. It just depends on what you are measuring. In the phase noise test, the Diamond measures better with DBS disconnected. Vodka and Diamond are close in both cases. In every test actually.

The stock cable just scores poorly in all cases. Cinnamon here too, but not in the phase noise test. That’s remarkable. As mentioned, the SIA-3000 measures differently (broadband). So it could be that there is measurable interference somewhere, making the Cinnamon come out worse in this test than the test on the Aeroflex.

Jitter and Allan Variance

As in the Wavecrest test, Diamond and Vodka are very close. Cinnamon scores well here. Alpha cable just scores mediocre here.

What is consistent is that Vodka and Diamond score well again. It’s a difference of almost 2ps from the standard cable. And yes: you’re just going to hear that. Is that worth an investment in a Vodka or Diamond? We’ll leave that up to you. After all, that depends entirely on the system, your expectations and the budget you have to spend.

Conclusion

Does a network cable do anything? Yes… 100%. We have demonstrated this several times now measurably and have also been able to establish it auditorily.

Should you replace all your network cables with expensive cables first? No, not that either. These are the finishing touches. But if you have already invested in acoustics, equipped your network with decent switches and are now thinking, “gosh, what else can be done?”… borrow some nice network cables. You will certainly be amazed at the impact!