I was using this as a USB receiver for many months before I got the CM6631A. I used it with a 12Mhz OCXO connected through a ufl cable, and some ultra low noise power supplies. I upgraded the I2S output buffer to a faster VHC from the HC, and I used a potential divider and capacitance multiplier to create a low noise supply for the Tenor IC. Since I got the asynch devices, this has not been used, so now I’m using it to test DAC chips such as ESS Es9023 and PCM5102a.
The performance specifications for the ESS ( ES9023 datasheet ) and the Ti ( PCM5102a datasheet ) are both very similar so it will be interesting to try them both and see if ESS’ hyperstream technology really does make a difference. This is the circuit for the PCM5102a :
The parts count on this diagram is just 9 capacitors and 2 resistors. De-emphasis (pin 10) should be tied to ground with a small value resistor to avoid noise, let’s say 1.5K. Format (Pin 16) should be tied directly to ground for I2S input. Filter (pin 11) can be switched between ground and DVDD via a large resistor such as 45K R, and Mute (pin 17) can be tied directly to the voltage input. It would be wise to use 100nF X7R caps across all power pins, as close as possible to the pins, to decouple noise, and then add larger caps in parallel as power reserves to reduce distortion. This is done on pins 8 and 9 in the diagram above, but I will also do it on pins 1, 18, 20 and 5. Pin 5 is negative (-3.3V) so the cap positive lead goes to ground.
Here is the ES9023 circuit. It uses 7 capacitors and 2 resistors.
The first thing to try is the PCM5102a, so I soldered one onto a pcb I got on Ebay. I did 3 of these, so I can improve the component layout with each build.
Then I soldered on the necessary components. Of course, it looks a mess because it’s a first attempt on a non-specific pcb. I tried to add a ground plane of sorts with some copper sheet. I decided to use one 3.3V supply for the first attempt. In total, I used four electrolytic 10uF caps, two 2.2nF film caps, one 2.2uF X7R and six 100nF X7R caps; and four resistors; 47KR, 1.5KR and two 442R.
To test it is working and see ‘how bad’ it can sound, I connected it to a PCM2707 USB receiver that uses a humble 12Mhz crystal and a 35uV noise 3.3V regulator and USB power. I was very surprised to hear lots of detail and quite a smooth sound. I can switch between the IIR and FIR while songs are playing and the IIR filter has more natural sound, most noticeable on fine percussion and sibilant s. I much prefer the IIR filter. There’s a very light background ‘thock’ noise when plugging in the DAC, and no noise on disconnecting.
After testing the PCM, I tried an ES9023 DAC, following the datasheet recommendations. I used a similar 3.3V regulator as the PCM and a 140K resistor for R8 to ensure there wouldn’t be any clipping. I then hooked it up and did a listening comparison. The ES9023 sounds like there is less jitter than the PCM because the definition and 3D stereo image is better. Tonally, the PCM is a little warmer and smoother, with vocals more forward, and the treble is more natural. The differences are surprisingly small, and I can’t say one is better than the other. They’re both very good. One of the biggest differences is that the ES9023 has a noticeable sharp ‘shlick’ noise on connecting, but again, no noise disconnecting.
I built another Tenor Te7022+Es9023 usb dac in a tiny box as a mini DAC for quick tests. I found that the ES9023 can behave badly – when my laptop boots with the dac plugged in, I get multiple, rapid, loud pops. Not good, so I don’t plan to do any more ES9023 USB gear – maybe I’ll try spdif using a wm8805. I also plan to put the PCM into the Teralink but so far, haven’t found the time.