(P)Re-Clocking the inputs
Quote from studiostevus on 2 December 2025, 3:15 PMHi Pedja - I'm fiddling around with various input sources for my S5 (sn #A01-2223, I *think* it should have the newer USB implementation?).
I was wondering whether the various inputs are reclocked in the S5, and to which degree they are dependent on input signal quality. My sources:
- Philips CD-880 unmodified via SPDIF/coax - solid transport, but old style clock implementation
- Sonos Connect modified with Tent Clock via SPDIF/coax - does the clock modification add anything, or is SPDIF reclocked in the S5?
- Macbook pro Audirvana via USB - my understanding is that USB is galvanically isolated and reclocked, hence signal quality/jitter matters less on USB?
- I2S I have no source but still curious whether there is reclocking applied or I would have to invest in reducing jitter at the source.
thanks!
Hi Pedja - I'm fiddling around with various input sources for my S5 (sn #A01-2223, I *think* it should have the newer USB implementation?).
I was wondering whether the various inputs are reclocked in the S5, and to which degree they are dependent on input signal quality. My sources:
- Philips CD-880 unmodified via SPDIF/coax - solid transport, but old style clock implementation
- Sonos Connect modified with Tent Clock via SPDIF/coax - does the clock modification add anything, or is SPDIF reclocked in the S5?
- Macbook pro Audirvana via USB - my understanding is that USB is galvanically isolated and reclocked, hence signal quality/jitter matters less on USB?
- I2S I have no source but still curious whether there is reclocking applied or I would have to invest in reducing jitter at the source.
thanks!
Quote from Pedja on 5 December 2025, 8:20 AMHello,
S/PDIF input - It features some jitter attenuation, but does not have reclocking in the strict sense. The S/PDIF receiver is CS8416, and the PLL filter used in the S5b is not fully the same as recommended by this chip manufacturer. The cutoff point here is a bit lower, but still similar. Generally, there is no jitter attenuation below 10 kHz, and therefore the final performance depends on the source quality. You will also notice the two modes of operation of PLL here (I have sent the performance difference between these two modes many years ago to the now ancient diyhifi.org forum). If set to Data ("D"), and with a good source, the jitter can be as low as 50 ps, and this is always added to the source jitter. Yes, the CS8416 jitter can go as low as 50 ps, and this is one of the rare cases where the part's performance is better than the one claimed by the manufacturer. So, for the S/PDIF input, consider that the jitter below 10 kHz will go through, and 50 ps will be added on top of it.
USB input - Yes, the same USB stage is still used in the S5b. It is galvanically isolated from the source, and the digital signal is fully (p)re-clocked "inside the DAC". You can still hear the differences between the different sources (for the reasons we still do not fully understand), but to a much lesser degree.
A-link input - This input provides a sort of direct path to the DAC stage. The only step between the A-link input connector and the DAC chip is the input switch. (The LVDS I2S input included in the later S5b additionally employs a differential line receiver.) There is no reclocking of any kind here, so the relation is simple: the better the source, the better the final result.
HTH
Hello,
S/PDIF input - It features some jitter attenuation, but does not have reclocking in the strict sense. The S/PDIF receiver is CS8416, and the PLL filter used in the S5b is not fully the same as recommended by this chip manufacturer. The cutoff point here is a bit lower, but still similar. Generally, there is no jitter attenuation below 10 kHz, and therefore the final performance depends on the source quality. You will also notice the two modes of operation of PLL here (I have sent the performance difference between these two modes many years ago to the now ancient diyhifi.org forum). If set to Data ("D"), and with a good source, the jitter can be as low as 50 ps, and this is always added to the source jitter. Yes, the CS8416 jitter can go as low as 50 ps, and this is one of the rare cases where the part's performance is better than the one claimed by the manufacturer. So, for the S/PDIF input, consider that the jitter below 10 kHz will go through, and 50 ps will be added on top of it.
USB input - Yes, the same USB stage is still used in the S5b. It is galvanically isolated from the source, and the digital signal is fully (p)re-clocked "inside the DAC". You can still hear the differences between the different sources (for the reasons we still do not fully understand), but to a much lesser degree.
A-link input - This input provides a sort of direct path to the DAC stage. The only step between the A-link input connector and the DAC chip is the input switch. (The LVDS I2S input included in the later S5b additionally employs a differential line receiver.) There is no reclocking of any kind here, so the relation is simple: the better the source, the better the final result.
HTH


