Finally, an affordable CD Transport: the Shigaclone story

[hayenc]

Another view on the issue of connectors, terminations, and cabling the digital output. This is one of the better summary articles I have seen.

http://www.positive-feedback.com/Issue14/spdif.htm

[johnm]

Interesting article!

I'm no engineer so haven't really got a clue what I'm talking about (!), but couldn't a diode be implemented at either the transmission or receiving end, which would stop any chance of a reflection when using shorter lengths of cable? I'm currently using 2 meters so it's not really a problem, but even so it would be nice to implement something which would stop any chance of these reflections taking place.

[chakija]

Now all i need is PS

Another newb question...if i remove choke (R4) ,will it be unable to use stock JVC-s PS then ?

[johnm]

Yes you can still use the stock power supply - you just need to follow Okapi's PDF to see where to solder the PSU wires

[chakija]

Well ,what i actually want to ask is will it work without needing to solder anything-just plug in PS socket and play ?
.
I suppose it wont then

[johnm]

Afraid not

[TriLithium]

If you place 2 jumper wires in place of the choke you can use the normal way of connecting things up. Its just soldering 2 wires instead of soldering 2 wires

[kevinkr]

Originally posted by chakija
Well ,what i actually want to ask is will it work without needing to solder anything-just plug in PS socket and play ?
.
I suppose it wont then

I'm assuming that the reason you are asking this question is you don't have an iron or can't solder?

Unfortunately the stock wiring is pretty marginal, and bypassing all of those long traces and eliminating the choke and making a direct connection back to a low impedance power supply is part of the process in unlocking the "magic" in this hardware.

Additionally the other mods listed in the pdf make a large difference in the performance of the mechanism/board - without them the full potential of this thing as a transport is not realized. The stock RCEZ32 is just ok unmodified, and really improves substantially with a few simple mods, and I assume the RCEZ31 based player most people here are messing with is much the same in that regard.

[kevinkr]

Originally posted by johnm
Interesting article!

I'm no engineer so haven't really got a clue what I'm talking about (!), but couldn't a diode be implemented at either the transmission or receiving end, which would stop any chance of a reflection when using shorter lengths of cable? I'm currently using 2 meters so it's not really a problem, but even so it would be nice to implement something which would stop any chance of these reflections taking place.

No... The amplitudes present in this transmission protocol are low enough that the diode may not conduct, plus even if it did most diodes take a finite amount of time to turn on and off and would introduce additional distortion in the transition time. (Very fast diodes are available) I suspect that when the reflection hit the diode as it was turning off it would just be reflected right back where it came from and at the receiver end this would make the problem even worse.

Forward drop of most diodes is a bit of a problem as well, gobbling up most of the available amplitude. (Not that this really relevant as it would not work for the reasons cited above anyway.)

The best approach is to try and minimize reflections where possible and when not to place them in a part of the time domain where their mischief causes no degradation. The previously cited PFO article provides some excellent information on this issue. (Disclosure: I am a technical editor at PFO and contribute sporadically.)

Since spdif is biphase ac coupling is generally used somewhere in the signal path. I thinking of trying ac coupling on the output and see what if any effect that has on the performance driving typical cables, currently the net dc value is about 1/2 the pp swing on the output. I'm not sure it has any effect at all. I have noted some transports have ac coupled outputs and others don't. (By definition anytime you see a transformer that is AC coupling.)

[kevinkr]

Last night I changed C950 (C952 to those of you with RCEZ31 based transports) from the stock 1000uF/10V cap to a 100uF/25V silmic, and there was a small but audible improvement in the performance.

The BG in this mechanism had been previously installed in the other mechanism and had been run in to some extent. Installing them in the newer mechanism did not result in exactly the same sound, however like the first one the sound seems to evolve with use.

I am planning a fairly wholesale replacement of most of the remaining electrolytic caps, and the deletion of a few as appropriate, but will do this in a piecemeal fashion when I can replace them without disassembling the mechanism. All that are inaccessible without mechanism disassembly will be replaced at one time to minimize handling.

When reading data sheets for normal use ignore the "Test" circuit values - these are intended for use on ATE DIBs for automatic testing and are values intended to work with typical high speed testers like the ones made by my employer. They aren't optimum for use in application circuits.

In many cases you can fine tune with excellent results using Peter's approach - most of the original component values were chosen so that until absolute end of product life minimum capacitance, and esr values are maintained to allow at least adequate circuit function. When using better parts it is often the case that they do not deteriorate as rapidly or at all (film caps) and a consistent level of performance is maintained with much lower and possibly more optimum values. Minimum acceptable values will differ significantly from device to device as well so Peter's optimum value may or may not be optimum for the chip in your device - tweaking is required. I think this is also where a lot of the fun of working with this stuff comes from, that and the idea of making a giant killer out of a $2 piece of hardware..