Musiland MINI I2S to Buffalo II DAC

Here is how I connected the Musiland I2S outputs to the Buffalo II DAC:

The twisted wire pairing to each signal wire are just ground wires to provide some noise immunity.

As it turned out, I had forgotten to solder the non-spdif pins and therefore I had no sound in my previous attempt. And even though I had modified the s/w for 24 bit I2S input and dissabled the automatic SPDIF detection feature, the current version of the s/w (v.05) should work.

The current software uses many startup default register configurations including auto-detect of SPDIF and 32 bit I2S; as such, the Buffalo II DAC can:

• Automatically detect whether you have applied I2S, DSD or TTL-SPDIF signals in the pins (by looking at the bitclock and data rate) [link].
• Automatically determine the bit-depth of the I2S signals

You must have a TTL-level SPDIF and bypass the comparator (SPDIF switch set to “OFF” or “1”) for this to work. If you pass the SPDIF signal through the comparator (SPDIF switch “ON” in the DAC), then I2S will not work because the comparator is also feeding pin D1, and even if there is no SPDIF signal, the comparator will hold Pin D1 low.

COMPARISON WITH SPDIF INPUT

I did not do any critical listening, but to me both interfaces sound the same. Surprisingly, however, the SPDIF interface apparently has less jitter; yes, LESS jitter.

With the SPDIF input, I had no problems using the lowest bandwidth for the DAC’s DPLL. With the I2S input, when using lowest bandwidth, I exprienced dropouts (the DAC loosing lock for a fraction of a sec and then regaining lock). I had to increase the bandwidth to “medium-low” in order to eliminate the problem. The Sabre DAC has adjustable DPLL bandwidth with 8 settings. So even though both SPDIF and I2S signals are generated in the same FPGA, the SPDIF signal is cleaner. This sort of contradicts the general notion of preferring I2S over SPDIF. There maybe other factors affecting the quality I2S in the Musiland implementation, but right now I can only point at increased jitter.

Update #1:

Coincidentally, two other users have experienced the same behavior in their USB->SPDIF/I2S converter boards: one board is the TPA USB board with the PCM 2707 chip and the other is the Teradak USB-SPDIF-I2S converter with Tenor 7022 chip (Teralink-X2)

Update #2:

There is something I don’t understand about how the Sabre DAC handles the I2S data. There could also be more to the way SPDIF is handled that ESS is not saying. For now, SPDIF allows setting the DPLL bandwidth to “lowest” which is the default setting of the Buffalo II DAC.

Because the lowest setting provides maximum jitter rejection, SPDIF is  the preferred interface with the Sabre32 DAC. For now don’t bother with I2S.

Below is a chart of transmitter jitter vs PLL bandwidth

Update #3:

Reader Rosendorfer posted a comment about using the HiFace USB-SPDIF converter.

Update #4:

There are reports about a problem with the I2S interface to the point that “a manufacturer takes the I2S input and converts it to SPDIF before sending the data is transmitted to the Sabre DAC“.

Reports of having problems with I2S/DSD interface into Sabre DAC:

1. Hiface Evo [Link]
2. TPA USB board based on the PCM2707 chip [Link]
3. Musiland MINI, based on Spartan FPGA [This post]
4. Teralink-X2 based on TENOR TE7022L and 1ppm TCXO [Link]
5. AudioGD ESS DAC [Link]
6. “Manufacturer X” converting I2S into SPDIF before feeding the DAC [Link]
7. SDTrans192 [Link]
8. ElectrArt USB Interface and SACD players: best DPLL bandwidth is medium-low [Link]

Reports of NOT having problems with I2S interface into Sabre DAC

1. Hiface USB interface [Comment section in this post]
2. SDTrans192 [Link]