SONY CDP-101 : CX20017 dac improvement

Discussion in 'Digital Sources' started by cizeta, Jan 18, 2018.

  1. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    To limit any current. The logical levels of the signals should be recognized virtually without any current passing. In reality a small current always flows ... So in theory there should not be, but in practice they are used to decouple the two integrated circuits. In CD players there are often such resistances between the filter and the DAC, even with values above 220ohm.
     
    FileFixer likes this.

     

    Please register to disable this ad.

  2. FileFixer

    FileFixer Well-Known Member

    Messages:
    545
    Location:
    Zagreb/Croatia
    Thanks for explanation... but you dont implement this in your mod?
     
  3. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    Later in time, right now I'm working on the release of the DAC. At rest it has about 3.5V DC to the output channels pins, according to the Nakamichi scheme. This continuous component is natural, since the TDA1311A has no negative feeds and can not supply the negative signal voltages at the output. It then operates in an intermediate zone of the power supply which it uses as its reference of zero. At the output, therefore, coupling capacitors must be provided for the signal, or a polarization of 3.5V DC on the input of the op amp which was previously referred to ground. I have just made this change, using the old zero adjustment trimmer already present in that area of the circuit and remained unused with the first modification.Centrating perfectly the working point of the op amp and eliminating the capacitors from the signal path, the audio quality is reaching very high levels ... I'm listening to old CDs that I know well and I'm discovering some details of the songs I had not noticed before ... IC508 and 509 are two NE5532 and IC514 is now an excellent LM49720. The modification proceeds well, but the wiring diagram has been heavily modified. At the end of the work (if it ever arrives), I will produce the wiring diagram with all the modifications made.
     
    Last edited: Jan 19, 2018
  4. FileFixer

    FileFixer Well-Known Member

    Messages:
    545
    Location:
    Zagreb/Croatia
    Nice job... enjoy in work. ;)
    BTW.. Did you know is that LM49720 are cross compatible with NJM4580 Op-amp? DV-300 have this op-amp on its output. I`m thinking to change this NJM4580 with something different like OPA2134, LM4562 etc...
    I`m do measurements on Pioneer DV-300 analog output and this is result...

    [​IMG]

    [​IMG]
     
  5. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    If this is your analog output with a 1KHz square wave signal you should be happy!
     
  6. FileFixer

    FileFixer Well-Known Member

    Messages:
    545
    Location:
    Zagreb/Croatia
    Yes, this is 1kHz square wave measured directly from RCA outputs of DV-300. It seems that lampizator have good nose for that things. :)
     

     

    Please register to disable this ad.

  7. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    LM4562 and LM49720 are the same op-amp. I am an admirer of the NE5532. It's a very good operational and used in almost all the professional audio equipment that I had the chance to open. It is so widespread that there is probably no known piece of music that has not passed through an NE5532 or his solitary companion NE5534 ...
     
    Last edited: Jan 19, 2018
    FileFixer likes this.
  8. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    However, if your outgoing signal is what you posted, then there is no low pass anti-image filtering on your DV-300. This thing can be dangerous for the amp to which you connect it ..
     
    Last edited: Jan 19, 2018
  9. FileFixer

    FileFixer Well-Known Member

    Messages:
    545
    Location:
    Zagreb/Croatia
    No problef for that... This is FR on my amp...
    Frequency Response:
    5 Hz–100 kHz/ +1 dB–3 dB

    DV-300 have 24-bit/96 kHz compatible DAC. It`s software upgradable and I try to find how to upgrade firmware to be able to play HD audio. Also I see that one seller near me seel DV-310 model with same DAC and all electronics configuration inside DVD but that DV-310 have additional USB port. Stronger model of this DV-xxx series have ability to play HD audio and have same DACs but all that players have different firmware.
     
  10. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    The problem is precisely the frequency response of your amplifier! It has a wider bandwidth than the audio and therefore also amplifies the ultrasonic components that are not filtered by the DV-300 .. It also amplifies the out-of-band signals! This involves power dissipation to amplify those unwanted components and tweeter breaking hazard due to overload of inaudible but very powerful signals ... this way you can break the tweeters even with normal listening audio levels ... be careful if the amplifier is very hot even at low volume ...
    it means that it is also amplifying something you do not hear ... a fixed note at high frequency, for example .... 96KHz is inside the 5-100KHz band!
     
    Last edited: Jan 19, 2018
    FileFixer likes this.
  11. restorer-john

    restorer-john Addicted Member

    Messages:
    5,344
    Location:
    Australia
    Measured on what piece of equipment?
     

     

    Please register to disable this ad.

  12. restorer-john

    restorer-john Addicted Member

    Messages:
    5,344
    Location:
    Australia
    Here's a few captures of DSO waveforms direct from a Sony CDP-101 (S/N 546495) This little DS1054Z scope is fantastic for grabbing images, using LXI over USB and much easier than fighting with my digital camera taking blurry CRO scope shots.

    Firstly, 100Hz square wave, (Denon Test Disc 38C39-7147- 1st generation) traces offset for visibility loaded with 10k ohms (important to load your outputs as scope input impedances are very high)

    RIGOL Print Screen20-01-2018 4_47_28 PM.794.jpeg

    Zoomed:

    The ringing is contained to the first 250uS and clearly demonstrates the non oversampling design of the Sony CDP-101.

    RIGOL Print Screen20-01-2018 5_35_19 PM.526.jpeg

    This is a single sample, 0dB. Note the pulsive signal leaves a trail of ringing, although it reduces rapidly and is all over in a few hundred uS.

    RIGOL Print Screen20-01-2018 4_54_57 PM.076.jpeg

    Zoomed for detail:

    RIGOL Print Screen20-01-2018 4_55_15 PM.829.jpeg

    Standard EIA toneburst, note the small wiggle at the burst start- this doesn't appear on my analogue CRO? I'll look into that.

    RIGOL Print Screen20-01-2018 4_58_26 PM.669.jpeg

    1KHz, sine, 0db, both channels. Levels are identical and inter channel group delay is minimal

    RIGOL Print Screen20-01-2018 5_00_48 PM.730.jpeg

    3.149KHz, 0dB. This is essentially the frequency at which the human ear is most sensitive to pitch variations (why it is used for W&F testing). The group delay at his point is an inconsequential 4-5uS I would estimate.

    RIGOL Print Screen20-01-2018 5_01_14 PM.846.jpeg

    Here's 10KHz:

    RIGOL Print Screen20-01-2018 5_02_07 PM.371.jpeg

    Here's 18KHz and I would estimate the delay to be 12uS. Spot on.

    RIGOL Print Screen20-01-2018 5_02_42 PM.104.jpeg

    Just or fun, here's 20KHz. We can't hear a 20KHz fundamental, but the interchannel delay is around 13.5uS on this shot. Does that correlate with the fS? Not quite.

    RIGOL Print Screen20-01-2018 5_03_26 PM.422.jpeg
     
    Last edited: Jan 20, 2018
  13. restorer-john

    restorer-john Addicted Member

    Messages:
    5,344
    Location:
    Australia
    The levels for both channels on my sample CDP-101 do not vary at all (+/- 0.01v) from 20Hz to 20KHz at the spot frequencies in each channel, both loaded with 10K ohms. See the scope shots.
     
    Last edited: Jan 20, 2018
  14. restorer-john

    restorer-john Addicted Member

    Messages:
    5,344
    Location:
    Australia
    OK, here's a Lissajous XY of the inter channel delay at 20KHz. The fact we don't see a perfect circle means Sony achieved part ways to their goal without needing a second DAC.

    Can I hear the channel delay? No. Not without resorting to test equipment or playing test tones that will highlight the delay.

    RIGOL Print Screen20-01-2018 5_15_37 PM.436.jpeg

    Here's a split screen:

    RIGOL Print Screen20-01-2018 5_16_11 PM.887.jpeg

    A better shot of the toneburst. Amazing. Try this with a vinyl front end and see what you get!

    RIGOL Print Screen20-01-2018 5_18_40 PM.887.jpeg

    This is an interesting discussion and perhaps I was a little harsh criticizing your modifications. Sorry. I see it is just you having fun and playing with NOS and perhaps even OS. For me however, I see the Sony CDP-101 as the culmination of decades of work, by ardent engineers all striving for one thing- perfection from their engineering perception. I hate to see their brilliant work dismantled but it's what makes anyone happy is what ultimately counts.

    Cheers. :)
     
    Last edited: Jan 20, 2018
  15. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    No excuse is due. I like to compare myself with other technicians, because opinions different from yours can enrich your personal baggage and can make you reflect on your technical choices. I appreciate your respect for the work of Sony designers and I assure you that I fully agree, but many detractors of this CD player complain about the slightly metallic sound of the first series of readers released on the market in the year 80. I have a CDP302ES to which I added the second DAC CX20152 that was missing, but fully respecting the Butterworth VCVS 7th existing filtering. Then I modified another of my 302ES with 4x oversampling and dual PCM56P-K DACs, redesigning the Sallen-Key filter for a new (higher) cutoff frequency. The design was done with Texas Instrument's Filter Pro software. The response curve is no longer Butterworth but Bessel, which guarantees no phase rotation on the whole useful bandwidth. There's my post about it. http://audiokarma.org/forums/index.php?threads/sony-cdp302es-digital-out-dac-mods.744104/
    Thank you for your kind technical comparison, which I appreciate professionally.
     
    FileFixer likes this.
  16. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    .[/QUOTE]
     
    Last edited: Jan 20, 2018

     

    Please register to disable this ad.

  17. FileFixer

    FileFixer Well-Known Member

    Messages:
    545
    Location:
    Zagreb/Croatia
    Measured on not to much professional equipment. It is totaly amateur way and equipment for professional use but before I choose this way to measure audio equipment for measurement between 20Hz-20kHz I checked and compare result to see how much is precise and does it worth to measure audio signals with this procedure at home and result show that for audio measurement on audible range, that way is realy nice and precise. The only requierment is to have good sound card. IMHO, C-media CMI6206 is tootaly unprecise for that purpose...
    On my workpalace I have few HPs, few Rohde & Schwarz and few Hantek DSO osciloscopes what I use to compare results and precision with this measurement method. I know that this method is not professional but maybe you dont believe, this is more then good method to see what happened in this micro electric world. As I mention before, maybe more then good sound card is needed if you want precise result and powerfull PC computer also. This is my configuration... All is free, try and see is this worth or not for home use.

    Soundcard Osciloscope by Christian Zeltnitz:
    https://www.zeitnitz.eu/scope_en
    [​IMG]

    Creative Soundblaster Audigy2 Platinum eX (kX drivers)
    http://www.guru3d.com/articles-pages/creative-sound-blaster-audigy2-platinum-ex,3.html

    [​IMG]
     
  18. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    When there is interest and passion for a sector of science like electronics, I am always respectful and I feel a natural empathy for the person who engages with enthusiasm in this world. We all dream of owning something we like and that we can not buy, even if we already have what we need for our experiments or for our profession. Using the sound card as a low-frequency oscilloscope is certainly fun and stimulating, while not guaranteeing absolute precision readings. I would like to mention the free "VISUAL ANALYSER" software at http://www.sillanumsoft.org/ which contains, in addition to the oscilloscope, also a useful audio spectrum analyzer and other useful tools. I fully support your enthusiasm, FileFixer ..
    Greetings!
     
    Last edited: Jan 20, 2018
  19. cizeta

    cizeta Member

    Messages:
    78
    Location:
    Milan - ITALY
    There should not be a circle, but a straight line inclined to the right at 45 degrees. Only in that
    case the two signals are in phase.
    The group delay must be, by definition, constant at all frequencies.
    In "linear phase" filters have a group delay (group delay, defined as a derivative of the phase with respect to the pulsation), in which all the frequency components have an equal time delay.
    Conversely, a filter with a "non-linear phase" has a group delay that varies with frequency, resulting in phase distortion.
    The anti-image filtering is the part that most degrades the audio signal of the CD players, so we try to make it less intrusive, moving the spectral components of the sampling rate as high as possible, so as to have a filtering of the lowest possible order and with lower phase rotations.[/QUOTE]
     
  20. restorer-john

    restorer-john Addicted Member

    Messages:
    5,344
    Location:
    Australia
    A full circle was what pretty much all first generation machines gave at 20KHz, whereas the Sony gave a slightly squashed circle- hence my comment a about 'part ways'. The perfect 45 degree line only came with two D/As. Even some twin D/A machines have some unintended delay at 20KHz.

    PC soundcard based oscilloscopes will give deceptively much nicer looking square waves due to the fact that they cannot 'see' anything over half of the sampling frequency due to the decimation filter on the A/D of the soundcard. Square waves will appear to be flat, with no ripple when in fact they are nothing like that. Consider even if a 24/192 sampling soundcard is used, the response of the oscilloscope is really only 96KHz bandwidth. A square wave needs at least 10 times that that to display even 3 harmonics.

    Agree with the use of the free software, VA. Fantastic piece of software. Italian too. :) There is/was a hardware front end that was designed in an Italian electronics magazine but I'm not sure if it still available. The FFT is superb for low frequency audio on a budget. I use it on laptops for a portable test suite, but you really need a line-in (not mic) and they are rare on laptops.
     
    Last edited: Jan 20, 2018
    FileFixer likes this.

Share This Page