Need -60dB 1kHz sine wave for CD

[Pedja]

Hi Bernhard,

OT: Why don't you scale your analyzer to 0dB @ 1kHz ?
That would be much easier to read the graph and see @ what level the harmonics are.

I measured different things with different levels and because I used the different levels of the attenuation (to use available resolution as much as possible) it would took some time to scale it properly every time.

I know that the distortion of the I/V dominates when a 0dB signal is used.

With a classic opamp I/V this can be done to be below the DAC chip’s distortion. However, I have found the things are sounding better if feedback is not used here. Hence the higher distortion figures. Btw, that graph above is pretty good figure for the common base I/V, ask Jocko.

Also even if the I/V would be totally clean, with 0dB signal the distortion of the TDA would be invisible on an analyzer unless you have a 1 million $ machine.
The lowest signal my analyzer can detect is -75dB, with averaging. ***
So this is why I asked for the -60dB signal.
(…)
*** The analyzer has only 12bit ADC but the repetitive accuracy is abolutely perfect.

Ok, this explains the things.

The reason why my posted graph looked so bad, could be:
I/V distortion dominates, like in the graph on your website.
Or the -0,5dB sine from test CDis a bad signal with high embedded distortion, maybe even recorded from an analog oszillator.
The -10dB sine from another test CD was terrible.
And on one test CD, you can even hear 50Hz noise in the pink noise.
I will be very happy to get the dithered 0dB signal, not for TDA testing, but for I/V stage testing.

It is uploaded, the 6N purity (THD=0.00003%) sinewave. Dither makes the structure of the file more complicated so the zipping is not as effective as with undithered sinewaves (where more than 5MB is packed into the 27kB file), hence because of the traffic limit on the Geocities (about 4MB/hour) I decided to cut both 0dBFS and -60dBFS dithered files to 30 seconds. Otherwise the single download of these two one minute files would block the traffic for the hour. I think 30 sec is usually enough. If you however need one minute, drop me an email address, I will send you.

So, why are some results from the -60dB signal so good ?
It is obvious that some plain 1541 chips beat some S1 chips.
Every chip got a different spectrum.

Yes, interesting, and the differences are surprisingly huge.

It seems that 1541A chips have always been selected, so if it is not S1 then it is crap.
Plain 1541 has never been selected, so there might be some with S2 performance among them.
What is also important:
Philips maybe did selection only per total THD vs. frequency, if you look @ datasheet.
They did not look @ single harmonics.

In fact, what was the way the S versions were made? Are they really selected? Or are they from the certain production run? For me it is hard to believe that someone took the 1000 chips and measured them to select, say, 10. And then 990 chips went back for the normal price. For the price of S chips as they are sold now it could be imaginable, but these chips were much cheaper while they were in production. But again it is not impossible, of course, just hard to believe. We need here an insider from Philips to get the answer, anything else is probably not more than a guess.

Pedja

 

[Bernhard]

Hi Pedja,

1541 datasheet say selection grade.

Also they are hand stamped...

Just there are the rumors around that S2 has been produced for Marantz, but who knows, maybe it was produced the normal way and selected after.

It is no real problem to select the chips, just put them into a socket and press a button...


Results from your undithered 0dB sine:

K2 -72
K3 -72,2
K4 below noise
K5 -73,9

but:

1800 Hz -66,9
2200 Hz -66,4
2600 Hz - 69
4200Hz - 66,5

There are two more OP27 in the signal path because I had to use my attenuator for best resolution of analyzer.

Oversampling was SAA7220B.


Bernhard

 

[Pedja]

1541 datasheet say selection grade.

Yes, it says so… But, how much money can be earned with measurement of 100 chips? How much time does it take the measurement of one chip? How much time does it take the measurement of two chips? Ten chips? Fifty? ... Ninety eight, ninety nine, hundred! How many S are among these? What was the price of S1 chip when it was in production? Discount the money that goes to the vendors. Do not forget, none semiconductor manufacturer would do this for the love of the audiophiles (be they subjectivists or technical freaks). Just thoughts.

Results from your undithered 0dB sine:
K2 -72
K3 -72,2
K4 below noise
K5 -73,9
but:
1800 Hz -66,9
2200 Hz -66,4
2600 Hz - 69
4200Hz - 66,5
There are two more OP27 in the signal path because I had to use my attenuator for best resolution of analyzer.

According to the content it is the intermodulation but the intermodulation results are higher than the basic freqs?

One notice: for harmonic distortion only measurements, dithered signal is certainly preferred but when it comes to the comparison with the data provided by Philips I overlooked the fact Philips’s data are for the THD plus noise and not only the THD, so for this comparison the undithered signals are better suited.

Pedja

 

[Bernhard]

damn it...

Today I built a preamp with one OP27 and the 50Hz is gone.

Still I can not get the noise floor below -66dB with that amplification.

Tested is 1541 from CD650 with SAA7220B with 512 averages.

The bad news is that first I wired the whole thing to the other channel of the CD880 and I got a nice surprise...

The result was very different

Left ? K2 O K3 -59,5 K4 -63,8 K5 -49,3

Right ? K2 O K3 -56 K4 -63,3 K5 -64,2

K3 is better but K5 is worse.


So both channels need to be tested

I have still another 6 S1 silvercrown candidates...


What will happen if chips are paralleled ?

 

[Bernhard]

One channel.

Bright points mark harmonics.

 

[Bernhard]

Other channel.

 

[studiostevus]

Sorry for bumping this... as we all know, Pedja's site is offline...

Where can I find these test tones, or can someone please send them?
Thanks!

 

[SoNic_real_one]

I made one with Arta. It can save wave files on HDD af specified lenght and frequency (I used 991Hz so the harmonics are not on the "grid lines"), then you can burn those on CD.

 

[studiostevus]

mmm thanks.... only on the registered version though...

By the way, do you use 32bits and then dither (as explained by Pedja in this thread) ?

 

[SoNic_real_one]

No, I don't dither, I think is cheating (lower bits will be noise, not data). Straight 16 bit.
I don't remember exactly, I did got somehow the Wav file on HDD... I need to get back home to check.

 

[SoNic_real_one]

This is how you get the wav file in Arta (unregistered).