clock feed to transport board question


2002-02-21 5:30 pm
Bath UK
I have a dpa Cdp which has separate Dac and transport boards, the clock is on the Dac board. It is a standard 74hc00 clock, the output connects via a 240 ohm resistor and 9-10 inches of cable to a sony CXD1167Q dsp on the transport board. the waveform on the transport side of the resistor is almost a sine wave presumably because of the cable capacitance (12pF) and the input capacitance forming a lp filter with the resistor.
What I would like to know is this likely to have sonic consequences, and any suggestions on how to correct it. I can't find a data sheet for the obselete dsp.
Although this was a £950 player and the dac board is a well made sm unit the transport board is actually an obviously cheap and badly made standalone cdp which has been adapted to interface with the dac. I have found that tweaking this side of the player to be much more rewarding than the dac.
thanx in advance for any assistance

Hi Khush,
If you use a normal probe to check the waveform of the clock you will see a distorted sine wave as the large capacitance of the probe lead is the cause. What you need is a piece of double terminated cable that is for 50 Ohm cable a resistor of 50 Ohm in series with the core of the cable at the measuring end and 50 Ohm from the core to ground (shield) at the scope. Now you will see a squarewave!
The CDP will work with a sinewave; I tried that with my KWAK-CLOCK but result is much better sonically with a squarewave. That's why I added the comparator in the KWAK-CLOCK. (BTW schematic available by sending me a email) Sound will be <B>much</B> better than with the 74HC04!
I also found out that some players do not seem to like the very steep waveform my clock produces. As in your case I would retain the 240 Ohm resistor. I consider it better to use a very short cable say 1 inch. Maybe you can arrange it that way if you install the KWAK-CLOCK.:)
I agree with Elso that a 74HC00 is probably not optimal as a master oscillator, although it is already better than using an on-chip oscillator somewhere in the decoder circuit.

However, I don't think you need to worry about jitter injected on the way from the DAC to the transport as long as the master clock is fed directly into the DAC or digital filter chip. Even jittery data will be latched into the filter/DAC synchronously with the master clock. The prerequisite being that on the average, the data rate is the same. This is fulfilled by driving both from the same oscillator.

So playing with the transport should be much less rewarding than optimizing the master oscillator. What chipset does your DAC use?




2002-02-21 5:30 pm
Bath UK
Thanks for the replies,
I was planning to build Elso's Kwak-Clock and was looking at the feed to the transport board as a way of optimising the benefits of this. The Dac is DPA's own circuit it is implemented in standard 74 logic, I don't know how it works and probably wouldn't understand if I was told. There are feeds from the clock to parts fo the dac so I don't know if I can move it but it would be worth a try, presumably the dac should operate from the extracted clock signal from the transport board.
The mods I have done to the the transport board such as replacing the caps with Os-cons and putting ferrite beads and rf chokes on the power lines have made a big improvement to the sound quality, but there still seems to be a lot of crap on the signals and power lines. would it be a worthwhile removing the redundant Dac on this board?
Removal of outputcoupling cap in KWAK-CLOCK

Hi fmak,
In my CDplayer the pin of the digital signal processor to which the clock is connected is biased to half the supply voltage (2.5V) , by a circuit internal to the IC. To avoid damage to the comparator in the KWAK-CLOCK I included the coupling cap. BTW is sounds better with the cap in place too.
Hope this helps,:)