› ›

ZX8401 / PCF1306P / Amstrad 40058 for 48K and 128K

bversteebverstee
edited March 2013 in Hardware
ZX8401 = PCF1306P = Amstrad 40058, let's get started with that.
These chips are used as surrogates.

Most of us might think those chips were only used by 48K ZX Spectrums, but in fact these are also used in 128K models (toastracks and grey +2)!

For the 48K ZX Spectrum these are:
- PCF1306P 519450U HfH8436 (checked on iss6 board)
- PCF1306P 521870U HfH8440 (checked on iss6 board)
- PCF1306P 521860U HfH8442 (checked on iss6 board)
- PCF1306P 522290U HfH8444 (checked on iss6 board)
- PCF1306P 522300U HfH8444 (checked on iss6 board)

For the 128K ZX Spectrum (toastrack / grey +2) these are:
- PCF1306P 522880U HfH8505 (checked on two toastrack boards and one +2)
- ZX8401 CF70001 (checked on toastrack)
- Amstrad 40058 8618 CF70001 (checked on grey +2)


So in general it seems:
Any ZX8401 / PCF1306P HfH84xx = 48K
Any ZX8401 / PCF1306P / Amstrad 10058 HfH85xx / CF70001 = 128K


I thought this is important info to register at WoS.
Post edited by bverstee on

Comments

  • jamorskijamorski
    edited March 2013
    Ben, anyone would think you had 'several' spectrums to test all these hypothesis on! I feel the need to insert a 'lol' here, but will resist.
  • smogitsmogit
    edited March 2013
    On a toastrack (issue 6U) 128k I have,
    PCF1306P 539620U HfH8505 1 X

    On a 48k issue 6A I have,
    PCF1306P 522660U HfH8501 1X

    I'm just wondering, but are these U or HfH numbers serial in someway? if so maybe its a good way to judge the build order of the machines?
  • nuggetreggaenuggetreggae
    edited March 2013
    well the 8501 means it was made in the 1st week of 1985
    and the 8505 means the 5th week of 1985

    so it certainly gives a good idea of when it was manufactured
  • smogitsmogit
    edited March 2013
    Ahh, makes sense, is this the same coding used on the ULAs?
  • bversteebverstee
    edited March 2013
    smogit wrote: »
    On a toastrack (issue 6U) 128k I have,
    PCF1306P 539620U HfH8505 1 X

    On a 48k issue 6A I have,
    PCF1306P 522660U HfH8501 1X

    I'm just wondering, but are these U or HfH numbers serial in someway? if so maybe its a good way to judge the build order of the machines?

    So HfH85xx is not 128K only?
    Hmmm, it's a pain in the * to figure out which chip is for which machine then...
  • bversteebverstee
    edited March 2013
    OMG, I was completely wrong...

    Well, these ZX8401, PCF1306P and Amstrad 10058 are used in 48K (issue 6 only), toastrack and grey +2, but...

    They are totally equal!
    No differences or whatsoever between the 48K and 128K models!

    To prove it, at least concerning the pinout:
    ZX8401_PCF1306P.jpg

    I just tried it on a toastrack, works flawlessly.

    Tha clears it all up!

    So that chip is for upper memory adressing, check out the 48K issue 6 and 128K (can't find online) / +2 schematics.

    But I always thought that because of the ram paging at the 128K, the logic would be totally different.
    But obviously tha PAL chip inside the 128K models copes with paging, leaving the upper memory logic intact, like the 48K.

    [edit]
    OMG again: now I understand.
    At the 48K pins 24 and 25 are used to select the 'upper or lower half' of the 4164/4532 upper ram chips, so Sinclair could buy cheap ones that only worked for half.
    Sinclair used that selection logic at the 128K for paging ;)
    "Now that all upper ram chips we order fully work, why not use them fully?" he must have asked himself. Haha!
    So the PAL adds even more paging logic, but the original memory adressing has been kept intact.

    What a wonderful solution!
    I love that man! ;)

    [edit]
    Ow, and one last edit: issue 6 must have been manufactured after the toastrack was born.
    They probably invented the ZX8401 / PCF1306P to save space and probably costs, and after ordering a million of them, they must have thought: "let's make one final batch of poor mens rubber Speccy's", using the brand new ZX8401/PCF1306P instead of the 6 OLD chips for adressing logic. Man....
    I wish I was an engineer back then.
    Or maybe in 30 years someone get's their hand on some software I am writing for my employer right now, or someone 3D-x-ray-scans one of my (and Miguels) PC keyboard interfaces / PlusD mini's / joystick interfaces and writes in the 2043's internet-succesor on something we now call a forum, but probably then will call something we haven't named yet: "man, that Ben Versteeg was brilliant!" :) ;) :) :) :)
  • KnikKnik
    edited March 2013
    Interesting find Ben !
  • csmithcsmith
    edited March 2013
    Check out the multiplexor schematic (equivalence circuit) on www.yarek.pl, it'll explain what's going on.
  • Pluto63Pluto63
    edited March 2013
    You can also check here
    http://trastero.speccy.org/cosas/JL/pcf1306p/PCF1306P.html
    And a brilliant idea here
    http://www.zxprojects.com/index.php/the-fix-a-spectrum-blog/21-noiseandaddresslinemess
    but I could not find if he completed it
Sign In or Register to comment.