NAD S500, AudioNote Dac1.2 upgrades

[Guido Tent]

Originally posted by Bryan
Guido,

What are the "best" available replacement's, and where does one source them? Is there ample support literature to guide one throught the implementation of the new clock?

Lets deal with the opamps later.....
I assume the above are your favorites.

How much money do clock replacements cost, and how long of a project is it to finish, 1 hour, 1 day, few days???

Also, does the new power supply come from tapping into an existing supply in the CDP, or is it built/added alongside during the clock swap? I assume this PS needs to be very clean?

Always learning more!!!

Thanks,

Bryan

Bryan,

I propose you do a search on DIYaudio. There are many clocks available:

Trichord
Audiocom
Elso Kwak
LCaudio
TENTlabs

and probably more that I am not aware of

I haven't been able to compare them all soundwise, but by measurement I found differences. Pricewise they differ a lot, as well as the DIY level of what is delivered and required to mount them.

Power supply in general should be very clean, and adding seperate transformer / rectifier is recomended.

As I am one of the manufacturers, I will not comment any more, but I am sure other members may be of help.

[Bryan]

Guido,

Understood

I'll do a detailed search, and ask around for feedback.
I appreciate you getting me started.

If anyone else wants to add any more input, please do so.

Thanks All,

Bryan

[Ulas]

Originally posted by Guido Tent

The timing of taking subsequent samples however depends on the short term stability of the clock. When converting from digital back to analog, a timing error automatically leads to an amplitude error.

Well said. But the timing error and subsequent amplitude error is only significant if the amplitude of the digital sample can be reconstructed to exactly equal the amplitude of the original analog signal. It cannot.

First, the digital sampling, with a finite number of bits, cannot accurately represent an analog signal. There will be a sampling error in the range of +/- ½ LSB.

Second, most modern-day digital audio production adds dither to the signal, which amounts to another ½ to 1 LSB of random noise.

Third, DACs are rarely linear down to the LSB.

Fourth, the typical DAC settling time is usually far greater than any XO jitter.

Fifth, given the inherent jitter in any digital circuit, especially one implemented in slow HC logic I fail to see how a little XO jitter can be a bad thing.

In fact, the effects of jitter are really no different than the effects of dither; they both introduce a small random error to the reconstructed audio signal.

[Guido Tent]

Originally posted by Ulas



Well said. But the timing error and subsequent amplitude error is only significant if the amplitude of the digital sample can be reconstructed to exactly equal the amplitude of the original analog signal. It cannot.

First, the digital sampling, with a finite number of bits, cannot accurately represent an analog signal. There will be a sampling error in the range of +/- ½ LSB.

Second, most modern-day digital audio production adds dither to the signal, which amounts to another ½ to 1 LSB of random noise.

Third, DACs are rarely linear down to the LSB.

Fourth, the typical DAC settling time is usually far greater than any XO jitter.

Fifth, given the inherent jitter in any digital circuit, especially one implemented in slow HC logic I fail to see how a little XO jitter can be a bad thing.

In fact, the effects of jitter are really no different than the effects of dither; they both introduce a small random error to the reconstructed audio signal.

Hi

My ears and measurements tell me that clock oscillator jitter still counts, at least improving the average oscillator in a CDP improves sound in my ears.

I am not refering to the lsb only, modern DAC are linear to and beyond 16 bits, which is on the disc, I do not expect more to come of.

Settling time is rather constant, though depends on (local) power supply variations, so I do not see the correlation with jitter you include here.

Oh, and the well know everywhere to find 121 ps jitter requirement is not valid. I dd some simple tests using a VCXO and well defined signals on the control port to show that.

best regards