Allo DigiOne S/PDIF HAT

First off, I’m ashamed to admit that I had this little gem in my possession for about 2 years before I finally got the chance to put it through its paces.

After all, it’s just a s/pdif output device for a Raspberry Pi, right? I mean, it’s just s/pdif, how good could it be?

It turns out it can be pretty damned good! But I’m getting ahead of myself.. Let’s start at the beginning.

The DigiOne is a HAT compatible with most if not all RPis and supported by most if not all audio distributions. It is intended to be plugged-in directly on top of the RPi, with no need for an isolator HAT. Plus, it is designed to be powered by the RPi via the GPIO header, so no need (or provision) for an external power source.

The DigiOne utilizes a WM8805 to convert the RPi’s I2S signal into s/pdif. The WM8805 is run in master mode, so as to minimize jitter due to the RPi’s problematic I2S clocking scheme. The WM8805 is clocked by the same oscillators that are used to reclock the s/pdif signal.

The WM8805’s s/pdif output goes through an Si8641 150MHz galvanic isolator and is passed to the “clean” side of the board.

There the signal is reclocked by a high quality flip-flop clocked by high quality NDK oscillators (housed inside a metal box, used for shielding against EMI/RFI). There exist two oscillators, one for the 44.1K family and one for the 48K family of sampling rates. The output of the oscillators is put through NB3L553s for buffering and isolation.

The entire isolated part of the board is powered by a DC-to-DC converter that offers galvanic isolation. Following this converter there exist a large number of LDO regulators and filter components. An LT3042 regulator is used to power one of the most critical parts of the circuit: the flip-flops that do the final reclocking.

So, very solid engineering all around. But how well does it sound?

The answer is, surprisingly well for the money.

My RPi stack included an RPi 3 with the DigiOne, powered by Salas’ new L-Adapter power supply and running Archphile. The music was coming from my NAS. No audiophile ethernet switches were employed. 😛
Pitted against that I had my Logitech Squeezebox Touch running the EDO plugin for up to 192K s/pdif from its coax output and my relatively pricey Pioneer DV-LX50 Universal Player (using its coax s/pdif output).

The music used was Dire Straits’ SACD album (having selected its CD (and not SACD) layer) which was also accurately ripped to my NAS.

Output from the s/pdif transports went into an AK4118-based s/pdif receiver of my own design which in turn feeds my dual mono AK4493 DAC. The DAC’s output goes through a Salas DCG3 preamp into my Hypex amp.

First up was the Pioneer. It had been a while since I had listened to it through its s/pdif output so I was in for a bit of a shock. Its output sounded coarse, strained, tiring. For a moment I thought that it was due to the SACD’s mastering (the CD layers of SACDs are rumored to be mastered intentionally bad so as to give the impression that the SACD layers sound even better than they actually do), but that changed when I switched to the Squeezebox. Things got noticeably better, actually listenable. Not exactly close to what I had been accustomed to using the Squeezebox’s USB port, but closer.
Then I switched to the DigiOne. Wow! All of the “coldness” of the music was gone, the stage gained depth and width, the music became more detailed and lifelike. This was definitely a step up.

I would dare say that this s/pdif setup came in fact close in SQ to my USB setup. This was a very pleasant surprise.

Now I need to do some A-B testing between the DigiOne and the USB output of the RPi. So to-be-continued..

Soekris DAM1021 s/pdif Inputs Board

I made a little s/pdif input board for my Soekris:

2015-12-05 16.41.34 (Large)

It has a coax input, two Toslink, and it includes a USB-to-serial adapter so as to facilitate easy update of the DAM’s firmware.

It also has an on-board low noise LDO for the Toslink modules and their switch, plus one more LDO for supplying the 1.2V necessary for the coax port.

More info to follow..

New page: Super Solid-state Sidecar

I built a solid-state alternative to the TPA’s Sidecar.

2015-11-03 20.19.20_resize

Its main features are:

  • Switching between I2S and S/PDIF with bus switches (solid state devices).
  • Support for two I2S inputs, with source selection.
  • On-board LDO regulator with 4.9V output for the Buffalo III’s 4-input S/PDIF board.
  • Drop-in replacement for the Sidecar.

Schematic, PCB, etc. in its dedicated page: Super Solid-state Sidecar for Buffalo III

The Texas Instruments PCM4222 Evaluation Board

I had been looking for a good ADC board for lab use as a measurement device.

At first I considered making my own, using something like the TI PCM4222 ADC chip, due to its excellent performance, good availability and relative ease of build.

But then I got lazy and purchased the PCM4222 EVM instead:

PCM4222EVM top

It is a 4-layer, well designed and implemented board.

Besides the necessary 2 clocks, it has a full set of digital outputs:
– Dual AES3 compatible outputs (both coaxial at 75Ω and XLR at 110Ω)
– I2S
– DSD
– Raw, directly from the modulators

Its analog inputs are balanced and very low noise.

Make no mistake, this is an evaluation module. As such, there are no mounting holes. Just four rubber “feet”:

PCM4222EVM bottom

Since I was in a hurry to get it to work, I set it up for operation with the least amount of power supplies (+/-15VDC and +5VDC) and set the DIP switches for I2S operation at 24bit and 96KHz.

I wanted to use one of its built-in AES3 outputs, but my sound card only had consumer level s/pdif inputs. After a short Google search, this came up: http://www.rane.com/note149.html

In other words, I had to make a 75Ω cable, with an XLR at one end, an RCA plug at the other and this circuit embedded in the XLR plug:

AES3_to_SPDIF

Surprisingly enough, it fit.

IMG_9400_resize

I hooked it up to the ‘scope to see if everything looked OK, and after being satisfied with what I saw I hooked it up to my X-Meridian’s coax s/pdif input. The sound mixer in Windows showed a full level signal coming out of it but it was heavily distorted. What was going on?

While troubleshooting, I tried the serial output port by connecting it to a WM8804 board that I had lying around. Sure enough, I got proper undistorted s/pdif output, so the PCM4222 was working as it should. So the problem had to be in the AES3 drivers.

After some more Googling I came across a post in some forum by someone who had a similar problem with me. He said that for some reason he had to set the ADC’s output to Left Justified (instead of I2S) in order to get the AES3 transmitters to output a proper signal. I made the necessary changes to the microswitches and lo and behold, it worked!

This is the configuration that I ended up with:

config_2496_LJ_2

config_2496_LJ_1

config_2496_LJ_3

Next up is a proper input stage, since the ADC has balanced inputs but most of my equipment-under-test is single ended.

The Soekris R-2R DAC

The UPS guy just dropped off my brand new Soekris R-2R DAC:

2015-01-30 13.55.04_resize

Also known by the very bland designation “DAM1021”.

It is a sign-magnitude R-2R DAC (a.k.a. “ladder” DAC), meaning that it is quite different in operation than the regular run-of-the-mill DACs.
It is more like a PCM1704-based DAC but with 192KHz+ support plus a bunch of high tech goodies, such as a built-in FIFO buffer.

It is available in three versions, with resistors of different tolerances (0,01% (high grade), 0,02% (mid), 0,05% (basic)). I got my hands on the 0,02% version.

It has three inputs:
1) I2S (electrically isolated)
2) Coax s/pdif
3) TTL for a Toslink receiver

Board-diagram

It is powered directly by a 2 x 7-8V transformer, but may be powered by a bipolar DC power supply.

It outputs a single ended signal at 1.4V RMS and also has a buffer for balanced output at 4V RMS.

It has a serial port for firmware upgrades as well as control.

I have already began work on a Soekris R-2R version of my TFT HiFiDuino Arduino code, tailored to controlling this particular DAC via its serial port.

The board will of course get its own page pretty soon.. Edit: the board now has a page.

To do: hook the board up and actually listen to it play. Stay tuned.