Noise in DAC

[Bernhard]

Hi,

Story:

Still my CD304 seems to be a very little bit smoother and darker, compared to the CD880.
Unfortunately the CD304 refused to play the test CD rom.
Today I cleaned the laser and it worked, so I was expecting different harmonic spectrum.

I plugged the TDA chip from 880 in 304. Test signal -60dB.

On one channel ( lower graph ) it is really a little different, 5 dB worse K2.

But other channel ( upper graph ) has noise that hides all harmonics.

Usually the noise is - 68dB.

This noise in the faulty channel is -48dB close to 1 kHz and goes down to -58dB @ 10kHz.

I have no idea where such noise may come from.

It is not noticeable when playing normal music, only visible with the analyzer.

If this noise makes sound nicer because of hiding harmonics, I would like to know how to achieve it

Is noise preferred over distortion by human hearing ?

[johnferrier]

Guess my thinking is that while there is disagreement about what the distortion profile should be like, there is agreement that noise should be low. However, everything is relative and it's really is up to your ears and mind.


JF

[Bernhard]

Originally posted by johnferrier
Guess my thinking is that while there is disagreement about what the distortion profile should be like, there is agreement that noise should be low. However, everything is relative and it's really is up to your ears and mind.


JF

This noise shows up with the -60dB test signal, with 0dB signal of course it will be much smaller.

[jean-paul]

Maybe a simple explanation is that the TDA is EOL ? Also measuring with -60dB is asking much from second generation chips that have had a life of several years of service and much power on/off sequences. As I told you before TDA's can fail ( the non A versions more than A versions ) which ends in one channel distorting or even no signal at all.

If it is only one of your collection I would not think much about it and try another one.

It is not noticeable when playing normal music, only visible with the analyzer.

It will be noticeable in time. The results you've measured definitely have influence on the sonical results, the human ear is sensitive to low level signals. Try listening with different material.
I think results will differ at power up and after 15/20 minutes warm up.

[Bernhard]

Originally posted by jean-paul
Maybe a simple explanation is that the TDA is EOL ? Also measuring with -60dB is asking much from second generation chips that have had a life of several years of service and much power on/off sequences. As I told you before TDA's can fail ( the non A versions more than A versions ) which ends in one channel distorting or even no signal at all.

I plugged the TDA chip from 880 in 304. Test signal -60dB.

The chip is 100% ok.

I have re-checked in CD880.

Maybe one of the 14 decoupling caps on the CD30 board is bad soldered or dead.

[cabbagerat]

Is noise preferred over distortion by human hearing ?

Yes, it is.
That is why dithering and noise-shaping techniques on digital signals can sound so good. They reduce harmonic distortion at the cost of increased noise. The noise masks the harmonics and is generally inaudible under the signal. Basically our brains are very used to hearing a level of white noise, but harmonic distortions sounds unnatural.
You might find this interesting:
http://recpit.prosoundweb.com/viewtopic.php?t=1284