I'd like to address those - point them out please. I've noticed you made the assumption that the designers of your DAC knew enough to avoid ground noise issues.
The ADuM isolators have relatively low capacitance in-out (I think around 2pF) but there is likely a route between groundplanes which has higher capacitance than that in your DAC. For example if the analog and digital supplies are both fed by separate toroidal trafos (quite a common occurrence) there will be much more coupling between them via trafo parasitics than through the ADuM.
It's often said that optical SPDIF is prone to more jitter than coax so any advantage of the galvanic isolation may be negated.
PC's are notoriously noisy for audio so it wouldn't surprise me if this is the source of the noise.
Very much the case that Toslink is subject to substantially more jitter than a coaxial connection, typically. How much difference the extra jitter makes does depend rather on the DAC.
Of course optical connection is very jittery but isn't synchronous reclocking within the DAC supposed to rectify that?
And one more thing. Let;s forget about the optical connection because I made these experiments at coaxial connection between PC and converter.
What is 'synchronous reclocking' in your particular case? Does the manufacturer describe in detail the technical aspects?
None of the 3 connections in OP have proper terminations so signal integrity may be an issue. Scope should help to find out if there are severe over/undershoots or reflections which may cause data errors that are perceived as noise. Another possible issue may be propagation delays which may cause data errors in reclocking (e.g. due to metastability if flip-flops are used).
I have full schematics of my DAC. Synchronous reclocking is done by 1 IC: SN74ABT574AN and then resynchronised signals go to PCM-63.
e
Well, looking at the schematics of my DAC and converter I see terminating 100 Ohm residtors at both ends (n the DAC and converter)
That tells me precisely nothing useful. Where does the clock signal on the ABT574 come from? That's pin11 on that chip.
There is one more thing that came to surface when I was looking at the schematics of my DAC and converter. My DAC uses ADM1485 for RS485 whereas the converter uses SN75LBC176D for the same purpose.
That does not mean that terminations are ok. You can't assess signal integrity without measuring.
OK. Thanks. I got your point. This might explain slightly worse quality of RS-485 connection.
Is that oscillator a VCXO? If not then the noise could indeed be caused by metastability in the ABT574 as @bohrok2610 has suggested. If the Golledge osc is not locked to the incoming stream then what you have is asynchronous reclocking. not 'synchronous'.
The sound card is synchronised by this clock in the DAC. So this is syncrhonous reclocking.
This one : https://www.golledge.com/products/g...d-oscillator-high-drive-capability/c-26/p-819
Seeing as your reclocking is for sure asynchronous, your issues are quite likely to be metastability related.
How does the 33MHz signal from the Golledge oscillator reach your soundcard?
Seeing as your reclocking is for sure asynchronous, your issues are quite likely to be metastability related.
How does the 33MHz signal from the Golledge oscillator reach your soundcard?
The clock goes to a flip flop and at the same time it is divided by 2 and then goes to converter where it is converted into SPDIF signal by DIT4192 and then it gets sent to the souncard as SPDIF signal and the card is synchronised to that incoming SPDIF signal. That's how it's done.
Last edited:
There's plenty of 'slop' in that locking arrangement so whilst I take back my 'asynchronous' reclocking remarks, there's still plenty of possibility for metastabilty as there's nothing ensuring the ABT574's set-up and hold times will be met.
Is the Golledge osc actually running at an audio multiple frequency like 33.8688MHz?
Is the Golledge osc actually running at an audio multiple frequency like 33.8688MHz?
How would you describe the noise? Is it similar to hum, hiss or random taps (like specs of dirt on LP or scratch on CD)?