AndyC_772;238566 Wrote: > I think you're over-analysing the behaviour of the TOSLINK connection, > and making comparisons that don't really apply. > > Of course if you're considering long-distance high speed > communications, then pulse spreading due to optical line width and > multiple propagation paths along the fibre are significant. But over a > few feet, running at 5.6 MHz? > > (As an aside - try looking up a data sheet for the type of high speed > comparator used as a line receiver for coax. I bet you'll find a skew, > which translates into data-dependent jitter, which is orders of > magnitude greater than any spreading due to optical effects in the > cable). > > What I do find surprising is that anybody designs a DAC that uses the > SPDIF input as a timing reference rather that merely a source of bits. > I've spent some of my spare time this year designing a DAC - based > around the AK4396 as it happens - which makes no attempt to directly > recover a clock from the SPDIF input. Incoming edges are used merely to > identify where bits start and finish so they can be sampled correctly, > nothing more. So, it's an inherent property of the design that input > jitter makes no difference at all. > > Doing this is not expensive, and I don't regard the use of a crystal > and an FPGA as "fancy". I do, however, regard the topology as "correct" > - and, fortunately for those of us with a working design with commercial > potential, "unusual". One day, all DACs will be made this way.
Ah.....but you are assuming that a few feet and 5-6 MHz aren't all that hard to do. Actually, the laser diodes used in the glass fibre setups do not like a few feet of fibre. Light bouncing back from the RX end does all sorts of odd things to it, at those lengths. It is not a problem of spreading, but that the spectral lines move around way too much. I call them "fancy", because most D/A boxes do not do that. They take the usual Crystal RX chip........hook it up as shown in the data sheet (which was written by people as wrong as the engineers who designed it), and figure it will work. Yes, it obviously works. But how well is another matter. Since you seem to have some technical background, here is something to try: Put a audio listening device (I'll let you decide the best way) to PLL loop filter point on the RX chip. Listen to it. Then play some music and then listen to it. You may be surprised. Or sickened, possibly both. Then report to us, and let the naysayers tell you that jitter is a lot of codswallop. -- ar-t http://www.analogresearch-technology.net ------------------------------------------------------------------------ ar-t's Profile: http://forums.slimdevices.com/member.php?userid=13619 View this thread: http://forums.slimdevices.com/showthread.php?t=33146 _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/lists/listinfo/audiophiles
