[Nerull]

Hey again!

Yes, I'm back with a schematic again, but I've learned a lot since then (redoing the board as well.) I just want to make sure everything checks out before I do the board again (hopefully no more revisions...). And here she is (meant to be used with the L1210 CD module and it's clock, which is the same frequency of the clock on the DAC board.) My only concern is the clock's ability to drive the SCLK of the PCM1794 and the CD servo (from the DAC board.) Guido's site said that the clock is able to drive 3 HCMOS gates, so I don't think it'll be a problem (will it)? All the power supplies will be driven by JSR-02 Regulators by Per-Anders.

All I2S cables and the SCLK cable will be shielded coaxial, with shield attached to the chassis.

Also, where should I ground the transformers to? They are Lundahl LL1674s and need a ground attached to somewhere, but where?
~Tom

 

[Garbz]

If you can spare a few extra parts then take Guido's indication and actually run the clock the HCMOS buffers like the Philips 74HCU04. This has got to be better then running the clock through resistors. Resistors just don't behave nicely with frequencies that high going through them, but I guess that clock despite being a precision clock would still be a sine output. In any event this would also solve some of your issues of running the clock over a cable and how many devices it can power as the 74HCU04 can put out more then enough current.

What transformers are you talking about? One side should be attached to the previous stage's ground, the other to the following stages ground. So if you bring coax cables in, do not tie them to the case. Rather tie the shields to the transformer and tie the other end of the transformer to the case.

 

[Nerull]

Well, I emailed Guido Tent to see what he thinks of the multiple output problem (on his boards, he uses only resistors for the outputs, and he has three outputs on the boards). So perhaps it's not needed, as long as it's not a very long distance (plus the logic gates will induce jitter, even if it's just a slight bit of it.) I would probably use high frequency resistors, so it may not be a problem.

The transformers in question are output transformers (here: http://www.lundahl.se/pdfs/datash/1674.pdf
) and I do not know where to connect that ground. Perhaps the PCB's (analog side) ground?

The coaxial cables wouldn't be used for the transformer actually, just simple shielded Belden cables. The coaxial are strictly for the I2S and SCLK lines.

~Tom

 

[Nerull]

Well, Guido responded by saying that I should use the resistors, so apparently it shouldn't be a problem (47 ohm high frequency resistors).

 

[Garbz]

Ahhhh Transformers for I/V why did you not say so

One side of the transformers should connect to the DAC's analogue ground. The other side assuming you are keeping the entire output stage passive should have its own ground not connected to anything but the output, and not the case either. If it is all hooked up correctly there should be no continuity between the output ground and the DAC's analogue ground.

 

[Nerull]

It's done! After much deliberating and pondering, I have finished it.

I've tried for minimal loop area, and minimal decoupling distance. Signal paths are also as short as possible...

Mounting screw sizes are #4-40.


~Tom

 

[Nerull]

And the board....