Advices on 10 mhz Clocks

@sls883 The game changing moment came when you asked about using square wave outputs on the ock-2.

Please clarify, when you had that game changing moment, were you connecting Lhy to both U18 and X30 or just U18 alone?

@oddiofyl I do like the individual impedance switch on the OCK-2—it provides some flexibility, even though I don’t have that many devices to hook up. Thank you for introducing the Mutec MC3+. I’m currently deciding on a DDC (SU-2, SU-6, U18, etc.) as well.

According to GoldenSound’s measurements, the jitter output from the MC3+ via RCA (407 ps RMS / 927 ps peak) is higher than the SU-2 (86 ps RMS / 252 ps peak), while the U18 performs the worst (499 ps RMS / 1217 ps peak). So far, I’m leaning toward the SU-2. Electrical noise appears to be comparable among these units. While there are no measurements for optical output, the jitter performance from the MC3+’s AES output is better.

Thanks @pinwa also for the info on the 50-ohm BNC cable—I’ll need one if I go with these “audiophile” clocks, since what I have now is a Belden 1649A, which is rated at 75 ohms.

Beyond these technical considerations, the puzzle I’m trying to solve—raised by @baylinor —is whether connecting the clock to both the DDC and DAC results in better sound. I understand @baylinor’s point: that with an I²S connection, the DAC’s clock is effectively slaved to the DDC. However, I’ve received feedback from others such as @sls883 suggesting that dual connection (clock to both DDC and DAC) results in better perceived sound quality.  Also, @baylinor, I recall you have attested in a previous post a while ago that you trust your ears perceiving improved sound with both connected, isn’t it or I have mistakened?

*** I think the main reason is that Gustard components like the X30 or X26 III are among the few DACs that can be properly slaved to a 10MHz clock when the REF clock setting is switched to “External.” However, when timing information comes from the DDC via I²S, it’s unclear whether the X30 or X26 III will remain ’properly’ synchronized to the 10MHz clock feeding the DDC—especially if the DAC’s REF clock is set to the default “Internal”.  This could result in imperfect synchronization between the DDC and DAC.

@lordmelton The 10Mhz master clock can not be used to directly run the DAC because it is not divisible by 44.1khz or 48khz. Its frequency needs to be internally converted to 45.1584Mhz and 49.152Mhz word clocks that a DAC needs to process the audio data stream through the PLL (Phase-Locked Loop) system.  I think Denafrips / DCS use native clocks directly to avoid PLL-based conversion or internal reclocking.

Goldensound actually demonstrates the jitter (short-term timing) performance is actually worse when 10Mhz clock is brought in as a reference, as you have also touched on.

However, the intention discussed here is to use a more stable long-term 10 MHz master clock to synchronize multiple components—such as the DDC (or switch/streamer) and DAC—as a shared reference, rather than clocking just the DDC or DAC alone. The listening impressions shared by @sls883 support this approach. I’ve also explained the rationale in the post mentioned above.

@sls883  Thx. Great, it is shielded.  What is length you used?

Very interesting that both of you are hearing more detail, so I did some digging to explore why that might be. It turns out that a perfect square wave is the sum of an infinite series of sine waves at the fundamental frequency and its odd harmonics. See the reference below. https://www.prosoundtraining.com/2010/03/13/square-waves-and-dc-content/  Mathematically, it can be expressed as:

Square wave (10 MHz) = Sum( n⋅sin(2π⋅(2n-1) ⋅ 10MHz) / (2n-1) ), n=1,…,inf

The higher the order of the harmonic frequencies, the sharper the edges of the square wave. These sharp edges more precisely define the timing information in a digital audio stream, which can lead to the perception of greater detail in the sound.