XMOS-based Asynchronous USB to I2S interface

Maybe someone will point me out to a product (hopefully not expensive :rolleyes:) so I can include it in all my future board assembly? Obviously I will not add capacitors floating around near oscillators like Walter because those pads will not last due to mechanical forces but this Blu-Tack tweak could be in my list if it proves to be worth to be taken into account!
Kind regards,
L
 
Maybe someone will point me out to a product (hopefully not expensive :rolleyes:) so I can include it in all my future board assembly? Obviously I will not add capacitors floating around near oscillators like Walter because those pads will not last due to mechanical forces but this Blu-Tack tweak could be in my list if it proves to be worth to be taken into account!
Kind regards,
L

Hello Lorien,

Imho, better & more usefull to add a connector with 5 twisted wires (5 pairs because close return gnd path needed- coaxial the best) to your kit for the isolated output like Ian does with him own kits (to avoid shorts as there is as well some voltage on one of the pins !). The result should be better than adding blue tack which anybody can find himself without special skill !

Possible to have some sharing between Ian & you to make a same I2S plug order to plug your boards together with no crossed non coaxial wires of different lengths (I talk about I2S links) ? (don't know if him (Ian) own boards have all the same order rank for plugin with I2S to the Wave I/O I2S J6 plug: FIFO, I2StoPCM, etc !)!

PS : does the isolator chip needs its own supply for better results ? Or does a shared supply on the dac board (or any shared PS after the isolator) is enough for sound quality ? I hesitate to give it an entire PS from the diodes bridge to its own standalone regulator just for it (the isolator chip) ?

regards
 
Last edited:
What surprise me is the mounting legs to tied the pcb to the case seems to have an influence ??? I mean hearable !

So what the best : plastic legs with double adhesive 3M scotch (very easy to find everywhere in specialised electronic shops)
I use myself corks, because it's easy to glue and screw and is a good damping material (I mean the cork, not the wine)

But read too much damping seems to be bad as well (rubber isolation no good at all : Wlowes experiments and input above) !

or a sheet of wood like some makes with traffos to isolate from its own vibrations ?

Firstly before playing with XO damping what is according to you the best way to tied the pcb to the case please ? (I already putted yellow "Blue Tack" on the 3 XOs !) !
 
No longer a curiosity. To satisfy the itch, I soldered a 4.7u Blackgate N to pin 4 of X1 clock. (actually to the + side of the OSON feeding X1) [...].
Thanks for sharing this. Could you show the two connexion points ?

Black Gate are scarce these days and if you find them their price is... too high :D. However it would be interesting to find a replacement part for this BG capacitor. Lorien and others, do you have some idea ?
 
what is according to you the best way to tie the pcb to the case please ?

Sorry but I just don't know. All I can say is that, if I use brass pillars to mount PCBs with vibration-sensitive components (as many do), I get an improvement in the sound if I damp the pillars with PTFE tape, Blu-Tak or whatever but that decoupling the PCB with rubber O-rings makes things worse.

My guess is that brass pillars drain vibration away from the PCB but it IS a guess, no more. Nor can I say whether other types of PCB mount would be better or worse or whether I'd get the same effect with a different case material. (I usually use plywood but plan to try slate tiles. Well, some day.)

IOW, you need to experiment for yourself. If you do, be sure to let us know how you get on even if it's only to say that you can't hear any change and that you think I'm insane . . .

HTH

Dave
 
Last edited:
I think damping components is not the way.Acoustic feedback it's to be considered on whole equipment . A good material to absorb vibrations is a synthetic stone the Corian placed under on spikes.

As I see it, acoustic feedback (I take it you mean that induced by a loudspeaker) has nothing to do with what I'm discussing.

Xtals, being electro-mechanical devices, vibrate. In any case, my WaveIO board rests on MSA Blackpods standing on a 25mm thick granite slab mounted on a 30mm thick seasoned oak shelf which, in turn, is mounted on 100mm thick cork blocks. The system as a whole sits on a 200mm concrete floor.

Despite all that, the difference in SQ achieved by damping the WaveIO's clocks with Blu-Tak as described is readily apparent.

D
 
I believe that a nice way to mount the WaveIO to help stop the vibration is the following

- find 4 very strong rubber elastics (or something similar) and mount them in each screw holes
- install 4 standoffs beside those holes but further away so you can attach the rubber elastics with enough tension and that it pulls in four opposite directions for all 4 corners

This will leave the module floating dampening vibrations. I've seen this type of arrangement for some XO modules and some CD transports. Very effective

Ciao!
Do
 
Mine bought one month ago has it yet ! There is no a new layout design for what I know

Lucian needs at least 3 weeks minimum after you order it.

If you I would try to hack the I2S in the Subbu if you can use USB ! With a short 6 wires flat cables where 3 wires (I2S) are connected after the removed résistors on the traces sided near the dac chip (between the w8804 and the ES9023). the three others wires which are between the 3 I2S wires (gnd return path) are tied between them near the subbu pcb and soldered near a ground pin close of the digital ground pin of the dac chip : scratch the painting of the pcb near it to see copper and use flux to solder on it the 3 tied gnd wires, or near the resistor traces on the gnd plane on the side which is closer to the dgnd pin of the chip !

But this is theory :D
 
Last edited:
@ zibra: there are no changes at the board's level but I did switch the NDK oscillators couple of months ago from 30 ppm to 20 ppm overall freq. tolerance (smallest number I could get from NDK). I guess this is the biggest change for the whole board. There are others to passives but small and hard to notice.
In addition, drivers aren't available for free download anymore (only on request) since I have my own strong reasons to do it but this is not tied to the hardware side, as you asked!
So, there's still SPDIF connection present on WaveIO, as it always was!
Kind regards,
L
 
Thats good news. I remember I saw some user Subbu dac paired with WaveIO through SPDiF connection. Dont remember name but I wonder if there was improvement because it also depends on SPDiF implementation in DAC.
In future Id probably try to connect it to Subbu through I2S but Im not super convinced if such DIY wiring for this type of connection is the best idea. Its very sensitive.