Advertisement



  1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Modifying NCC200 boards

Discussion in 'd.i.y.' started by a.palfreyman, Jun 11, 2022.

  1. a.palfreyman

    a.palfreyman pfm Member

    Well, just over 3 years ago now (how time flies!) I bought an almost fully populated spare set of NCC200 boards from @glenn jarrett, so maybe I ought to do something with them!
    I have been reading the stuff on the Doug Self 'blameless' amp and inspired by this plus some other recent threads on here, I got the box of goodies out. In light of the Self stuff, it struck me that the earth arrangement at the input, with the incoming screen connection and the adjacent earth return to the psu being 30mm apart might be a source or ripple injection as the input loading and the F/B cap go to the psu return point and are thus not common with the incoming screen. I decided to measure the resistance of this trace out of curiosity by using my bench psu and a dummy load of 2x15R // to give 7R5. Adjusting the psu to give 7.5Vdc across the dummy load gives 1A current and passing that through these two connections and the trace, with my DVM clipped to the board so it makes contact with the solder pads beneath showed 14.7mVdc ie roughly 15mR:
    [​IMG]
    I then soldered a short 'u' shaped piece of wire to the back to bridge these two 'earth' points:
    [​IMG]
    and then reconnected as before:
    [​IMG]
    This now gives 0.4mVdc, equivalent to 0.4mR, which is a 30dB improvement! Thinking about this, at 100Hz a 100uF cap has an impedance of circa 15R. Therefore any ripple voltage across the capacitor will induce ripple between these two points at -60dB. The modification should now reduce that to approaching -90dB. The screen connection is at the +ve supply side compared the the psu return and this probably explains why the board responds well to regulation of the +ve supply as this removes any ripple between these two points.
    I now have to decide how much of the 'blameless' circuit to incorporate... Initially I'm inclined to do the current mirror and the VAS mods on one board, then 'jiggle' the circuit boards so the drivers work as emitter-followers and feed it with my 3.1kHz low distortion signal source and scope the output.
     
    ramona and Mike Hanson like this.
  2. bravem01

    bravem01 Active Member

    Similar to you I had concerns re the earthing on my NCC200 and Qudos boards after reading Self's book so I lifted the 0v side of the two electrolytics and linked them together and then ran a separate wire from there to the amplifier star earth. According to Self this reduces the decoupled supply noise feeding back into the input circuitry.
     
    Mynamemynaim and a.palfreyman like this.
  3. andrewsutton

    andrewsutton pfm Member

    The NCC220s appear to have a wider track at this location.
    Is it wide enough?
    Cheers Andy.
     
  4. a.palfreyman

    a.palfreyman pfm Member

    Probably not...
    @bravem01 has the correct approach as it separates the incoming signal ref back to psu from the asymmetrical earth return currents through the front end 100uF caps. Just trying to work out a neat solution...(mybrains are slow) :eek::D
     
  5. Mynamemynaim

    Mynamemynaim 38yrs a Naim owner

    Reminds me of a pfm member (cannot remember who!)
    They fly leaded every negative on a naim preamp back to the star earth
    It looked an arse....but they claimed it lifted the sound to a totally different level ...
    Was the thread called something like "a Christmas present to everyone "???
     
  6. a.palfreyman

    a.palfreyman pfm Member

    Split the front end earth returns.
    Three cut tracks:
    [​IMG]
    Input load resistors and F/B points linked back to 'screen' tab (silver coloured wire) and return of +ve caps linked to '0V' tab (lacquered copper coloured wire) now called 'SIG0' & 'FE0' respectively:
    [​IMG]
    'Screen' tab just needs a stackable female spade (to take incoming screen wire) linked back to main psu board.
     
  7. Avon

    Avon pfm Member

    At least the power ground and the signal ground are independent. But I always notice the lack of attention to the earthing arrangements used in power amp PCBs. I’d suppose the signal currents are so small that they may not influence each other. Unless you make a before and after comparison, then you aren’t going to know if splitting the signal grounds made any difference. It can’t do any harm obviously and it seems like a no brainer.
     
  8. a.palfreyman

    a.palfreyman pfm Member

    Problem with class b amps is the asymmetrical psu currents. This can put asymmetrical (albeit small) voltages on the front end psu return wire that modulates the front end, muddying the sound. At 1W output, you have 2.8Vrms on the output, that is equivalent to 0.1Vrms on the input. Therefore even a tiny voltage of say 0.1mVrms modulation is 60dB below the input level and this also get amplified.
    I'm going to jiggle the resistors feeding the drivers and on the driver outputs so it can drive the feedback circuit and some headphones. That way I can listen to each in mono, one modified and one un-modified.
     
    Mynamemynaim and Mike Hanson like this.
  9. Richard Lines

    Richard Lines pfm Member

    Good Morning All,

    Wouldn't a more elegant solution be to just use a piggy back connector to the signal earth and run a separate earth wire from the piggy back to an earth location?

    Regards

    Richard
     
  10. a.palfreyman

    a.palfreyman pfm Member

    Biscuit tin engineering:
    [​IMG]
    I have fitted the VAS mod and the current mirror.
    Questions:
    1. How accurately do the transistors need to be matched in the current mirror? These measured about 6mV difference Vbe on my tester and I'm getting about 3% difference in the voltage across the 100R input tail resistors.
    2. Also, can I tweak the 68R current mirror degen resistors to balance the mirror?
     
  11. bugbear

    bugbear pfm Member

    Some googling led me to "Linnik's XMas mod", which I think is this thread:

    https://pinkfishmedia.net/forum/threads/fantastic-mod-for-nac-found-a-little-piece-of-gold.18261/

    The actual mod is post #42

    BugBear
     
    Mynamemynaim likes this.
  12. a.palfreyman

    a.palfreyman pfm Member

    Arf. When I read the above I initially misread it as 'MORON'....
     
  13. a.palfreyman

    a.palfreyman pfm Member

    On the side with the lower tail current, I put a 2k7 across the 68R degen resistor and this has reduced the difference in the voltages across the 100R tail resistors to better that 0.5% where it was about 2.5%. :)
     
  14. bugbear

    bugbear pfm Member

    I may have found the "Linnik's Xmas mod", but I'm failing on the "VAS mod" and current mirror.

    Any details or old posts you can point to?

    BugBear
     
  15. a.palfreyman

    a.palfreyman pfm Member

    Hi @bugbear,
    The mods I'm doing aren't related to the Xmas mod. This is about the Self 'blameless' amp and taking inspiration from that in molding a set of NCC200 boards. :)
    @Mynamemynaim dragged it off topic :eek::D
     
  16. bugbear

    bugbear pfm Member

    Understood.

    I was asking about the mods you ARE doing, the VAS mod and the current mirror (I'm assuming the latter is associated with the long tail pair, as per this discussion https://sound-au.com/amp_design.htm#s1)
     
  17. Mynamemynaim

    Mynamemynaim 38yrs a Naim owner

    Not so much dragged it off topic....more of opened a door marked "do not open":rolleyes:
     
    Mike Hanson and Richard Lines like this.
  18. S-Man

    S-Man Kinkless Tetrode Admirer

    Yes, that door should remain closed. Unless you like the effects of loads of extra impedance in the tracks and the vastly increased susceptibility to interference and pickup.

    AP is trying to improve the grounding using engineering principles, not by applying the "if a little is good, more is better" fallacy.
     
  19. a.palfreyman

    a.palfreyman pfm Member

    Sorry @bugbear I misunderstood you.
    The full article is here:
    http://www.douglas-self.com/ampins/dipa/dipa.htm
    I've read through it about a dozen times now and SOME of it is starting to sink in...
    It really is worth taking the time to read as everything is explained in detail even if some (most!) takes some getting into your head. :)
    @S-Man I have @bravem01 to thank for pointing out the split earth arrangement which makes great sense.
    @Mynamemynaim I dont mind going o/t into interesting areas. ;)
     
  20. jpk

    jpk pfm Member

    The preamp with separate GND lines was called Starfish...
     

Share This Page





Advertisement


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice