Best way to route patch wire (for clock signals)?


I just got the schematic for my DAC (a Cambridge S700 isodac) and the clock is organised in a very strange way (I'll write more about it later, I got to go to a meeting in a moment) so I'm going to more or less remove the clock circuity and and re do it with patch wires.

I just wondered, what's the best way to route the patch wires on the PCB?? Shortest possible route or orthogonally across the PCB??

Many thanks,

Sounds better now...

Thanks - I did the work and the sound is much better now. I just routed the patch wires, more or less, the shortest route...

And this is what I did...

The S700 DAC uses a CS8412 reviever, PMD100 filter and SAA7350 DAC. However, things were quite strange in the clock department, partly because the thing has an optical clock output which enables it to sync to a Cambridge CD transport - There was a 16.??MHz crystal on board for this purpose. What happend was the clock from the 8412 was buffered and fed to the PMD100 but instead of sending the clock straight to the SAA7350, a W42C70 frequency multiplier was used to generate a 16MHz clock for the DAC. The Specs say that the jitter on output of this chip is 500ps!

The clock signals also went through a few gates which were used to select which clock source the SAA7350 would use. What I did was to remove the W42C70 and the gates for selecting the clock source and run the DAC straight off the buffered clock from the 8412. I also had to tell the SAA7350 to expect a 256Fs clock, rather than 384Fs.

After these changes the sound was more open and the stereo image better defined. A tweaker's job is never done, thought, so a couple of days later I modified the reciever's loop filter according to wildmonkeysects' ideas and this improved the sound further!

Now I'm totally happy - Almost - Now I gust want to Gen lock the DAC to my transport and replace the SAA7350 with a decent DAC...




2001-12-27 1:55 pm


I can`t find a datasheet for SAA7350 and I see, that You have a schematic of Your Cambrige S700. Could You send me a pic with part of Your schematic with SAA7350 pinout and post filter. I would like to modify my old Marantz CD72.
My opinion, that it`s not the least DAC, and in a good environment it sounds not bad. In my case I can compare it with TEAC VRDS-25(with AD1862, still not modified yet). Sounds are different, but I can`t say that one is absolutely better than other. It depends of feeling and daytime. Try before place a new DAC finally. Might be that something will appear, but something will miss.

my best regards,kvv


2004-01-18 10:18 pm
Cambridge Audio S700 circuit diagram

I have stumbled across one of these and before I rip the pcb out and start again I would like to see what can be done with this dac.

Provisionally, the psu looks rubbish the clock is mad and let’s not mention the output stage.

My question to you all is where can I lay my eyes on a circuit diagram. And does anyone have any suggestions regarding improving this unit. (apart from starting again).
Be very careful when judging Phase Noise performance simply from Jitter Specs. In realty due to the low clock multiplication ratio of 1.5, the RMS jitter is about 50pS. However, the real beauty of the jitter is the spectrum is now almost completely random, i.e. no signal correlated artefacts. If you’re going to have jitter – then have random phase noise.

The best is no jitter – so use the clock-lock function – you need to modify your transport / CD player to accept a 128Fs (5.6448MHz) clock input - and then use the optical SPDIF (for RF isolation between CD and DAC) for the Audio data, as jitter is no longer an issue.

I would not operate the SAA7350 at 256fs, as its noise shaper performance is much worse then in 384fs mode.

The ISOMAGIC design does not use the SAA7350 internal opamps, instead external AD712’s are used – these still perform well in the presence of Noise-Shaper RF. Once the external opamps are used, SAA7350 becomes a completely different beast.

The 5534's on the PCB are used for low noise voltage references – as such, they offer very good performance. A massive sound improvement was obtained when the AD711's differential summers and output buffering where changed to a simple discrete design – but Clock-Lock first.

Once the above have been performed, the next best gain in performance can be had by improving the PSU section. Supply regulated +/- 15V to the board.