B³II DAC: Bare Bones Buffalo II DAC
I have decided to first use Buffalo II without any external tweaks. I appreciate the engineering effort that Russ and Brian at TPA have invested in this latest implementation of the ESS Sabre32 DAC and believe they have selected the best components for a balanced approach (meaning that all components work at similar levels of performance -also meaning that for any meaningful performance increment you would have to change everything to the next level of performance).
This is what I plan to use for this initial listening phase:
Power supply: single “wall wart” modified to output 5.4V (the regulator is a LM317 class, so it was a matter of changing the adjusting resistor). I also salvaged a DC jack which has been implemented with ferrites and capacitors for EMI filtering.
Output connectors: balanced cat-5 through RJ45 Ethernet jacks
Output stage: none. I’ll swallow the 12db THD penalty. But will connect directly into a balanced amp. Besides the amp best specification for THD is -82db whereas the Sabre32 in voltage mode has a THD specification of -108db. I doubt I will miss those 12 dbs .
Control: AC1 (later I will work on an Arduino Implementation)
AVCC shunt is installed on the back of the board, bare bones doesn’t mean I eliminated the AVCC regulator
(IVYIII: What I really want to do is to use IVY as the input stage for Hypex amps, replacing the input buffer and connecting directly to the UCD module)
“In spite of that, SDIF3 is 75 ohms. Not that I’ve ever minded. That aside, most equipment terminates at 75 ohms (if they terminate at all, big if in word clock connections) but the connectors are 50 ohms. This is because the pin of a male 50 ohm BNC is thicker than that of a 75ohm one. If you stick a 75 ohm cable part into a 50 ohm chassis part that usually works. OTOH if you do the reverse you have to jam it in after which the 75 ohm chassis part is damaged and will no longer work with 75 ohm cable parts. Because of that, most boxes have 50 ohm connectors installed. At least nothing will break. At frequencies encountered in word clock and SDIF2/3 connections, the impedance mismatch problem at the connector is academical.” [link]
If you own a Buffalo II DAC, you might have noticed that in the backside, there are a few ferrite beads:
- L1-L5 are ferrite beads at the outputs of the local regulators into the DAC. TPA has been doing this for their designs. Take a look at the circuit diagram for the OPUS DAC.
- L6-L8 are ferrite beads at the inputs of the LT1763 regulators. Removing these will disable these local regulators (so you can use your favorite regulator instead)
- Ferrite beads are used for EMI noise suppression. This application note from Murata describe the properties of ferrite beads
Why Musiland 01-Mini?
Because it is the cheapest model (I paid US$ 60), and easiest to get to the I2C lines.
SUMMARY OF MODS (all easy to do)
- Enable SPDIF: link (The circuitry is already in the board, but requires installing a resistor to enable the spdif output. When you load the driver and the Musiland Mini is connected, you will see a SPDIF option)
- Tap I2S lines: link (avoid a conversion step)
- Disable DC-DC power supplies: link (cleaner data/clock lines). Before and after clock trace.
- Increase linear regulator capacitors: link (better power regulation)
- Added a (5V) power indicator LED (There are better ways to install the LED. For example, remove C60, install a surface mount resistor (500-1K ohm) on the power line leading to C60, and cut the power to the DC switching supplies near C12; then install LED in C60) Read more…
It has been a year since I started HIFIDUINO, primarily as a notebook for my own learning. Many users have found it useful and I plan to continue chronicling my own adventures with Arduino and DIY audio. But like anything else in the web, it is good to try something new. So I am moving from Blogger to WordPress. The old blog will continue to be available here: www.hifiduino.blogspot.com