Connect with us

Datasheet 4116 memory chips; searching for Sinclair ZX Spectrum parts

Discussion in 'Electronic Repair' started by [email protected], Jun 21, 2006.

Scroll to continue with content
  1. Guest


    I'm suffering from the well known Blinking Coloured Blocks Problem
    (BCBP) on my 48K Sinclair ZX Spectrum. The 12V, 5V and -5V lines seem
    ok, 7805 regulator seems ok, ULA is 100% ok (put it in a different
    machine). Changed an electrolytic capacitor that had some slight

    Last one in the list (not looking forward to desolder these ones), are
    the 4116 16K base memory chips... But where to find replacements
    (peferrably in the Netherlands)? I also might want to change the
    transistors T4 and T5 (ZTX types), but I have not found these as

    One electronics shop offered me as a replacement for the STC4116-2N
    memory chips: the MCM4116BP15, but I can't find any information /
    DATASHEET on both, so I have no idea if this replacement is ok? The
    only thing I can find in ZX Spectrum 48K service manuals is that the
    type should be 4116 and 150ns and I know 8 of these make up 16K with
    multiplexed switching of an address line to address the memory

    Any help, pointers, datasheets and parts highly appreciated!

  2. Brane2

    Brane2 Guest


    I used to replace 4116 with 6164, ofcourse after disconnecting extra
    power lines and fixing them to appropriate spot, like GND. Not really
    high on aestetics and authenticity points, but it worked...
  3. n cook

    n cook Guest

    I've hundreds of them
    ET 4116N-3 datecoded 1983
    About 2 for 1 Euro plus p&p for about 20 off
  4. Guest

    You should be fine using 150nS chips (as you were offered by the shop)
    instead of 200nS chips (the ones in your computer). Someone else responded
    to your message that he has 300nS chips in stock, I think those would be
    too slow.
  5. Gerard Bok

    Gerard Bok Guest

    No :)

    That someone offered 4116-3 chips.
    4116-2 = 120 nsec access time
    4116-3 = 90 nsec
    4116-4 = 70 nsec

    (UMC 1986 Memory databook. I got suspicious.
    300 nsec was the slowest part, back in 1979. It would be odd that
    someone would still be stamping them out in 1983 :)

    A word of warning: I recall some early Sinclairs to be a real
    menace to desolder. Holes that are tight enough to serve as
    sockets, even without solder :)
  6. Brane2

    Brane2 Guest

    That's why god invented OLFA knife. Just press the knife vertically at
    the spot where pin enters the plastic housing and just a bit along the
    edge of plastic and cut it off.

    After cutting all the pins just heat them and pull them off one by
  7. n cook

    n cook Guest

    Also 14 pages of data for the Mostek 1980 version of 4116
  8. Guest

    Okay, weer wat geleerd... I was thinking of the codes on later DRAM
    chips, not knowing they used another system on these.
    A possible trick would be to cut off the legs of the old IC's, then
    unsolder them one-by-one.
  9. GdB

    GdB Guest


    So, these "ET 4116N-3 datecoded 1983" memory chips should be fine? So,
    if the number of nanoseconds is LOWER, it is never a problem; if it's
    higher then it is a problem??? Most important thing for me to know is,
    if these chips from N. Cook / Diverse Devices will work...

    I don't know exactly what is expressed in this timing/frequency... I
    thouht it had something to do with refresh rate, but then a refresh
    needed every 150ns sounds better than refreshing every 90ns...

    Maybe someone has an old datasheet, I'll search for the 2118 to see
    what it is.

    Any pointers to the ZTX213, 313 and 650 transistors?

    Kind regards,


  10. Hi,

    To speed up difficult repairs you can also let the pins that are cut off
    in the pcb if there's enough space, sometimes better to prevent heating
    damage when clearing the empty holes. This method used a few times when
    replacing (TXT) chips on the digital signal processing module of D16
    chassis television sets (repeatedly flashing wrong (txt) picture will
    mostly be the symptom then).

    succes, greetings Bart

    PS: if availability is a problem i have the chips possibly also, but not
    nearby (except the same type is used on an IBM 286 6Mhz AT system
    board.........., around 32 chip positions with 2 paralelled chips
    per position for increased fan-out.......)
  11. n cook

    n cook Guest

  12. Eric Rullens

    Eric Rullens Guest

    Usually refresh rates are in the ms range (and are not stamped on the chip),
    access times are in the ns range.

  13. ian field

    ian field Guest

  14. Guest

    = BC213, universal PNP, replaceable by BC557
    = BSV26, low power switching NPN, replaceable by BSX20 or 2N2369
    lo-sat 60V 2A NPN, replaceable by 2SC3328, 2SD1207 or 2SD1835
  15. DizTRacT

    DizTRacT Guest

    Those are static memory chips, which don't need or use refresh.
    That's why the type is called static (in contrast to dynamic, which do
    need refresh).
    The numbere of ns stand for access time for these chips.
    (the minimum time needed before a stable output signal is available.)
    So lower numbers are always allright in this case.

    - DizTRacT -

    Ik wil niet genieten, ook niet met mate :)

    Voor antwoord per e-mail, eerst ".DitNieT" uit het adres verwijderen !
  16. Eric Smith

    Eric Smith Guest

    No, the 4116 is NOT a static memory. It's dynamic.

    It's the access time. 90 ns is better than 150 ns.

    The refresh time is one complete refresh every 2 ms, regardless of the access time.
  17. Franc Zabkar

    Franc Zabkar Guest

    A colleague once showed me a clever trick. You can sometimes identify
    the faulty DRAM without desoldering it by piggybacking a known good
    DRAM on top of it. Despite my skepticism the technique worked.

    - Franc Zabkar
  18. Franc Zabkar

    Franc Zabkar Guest

    Don't you mean 4164? The 6164 is an 8Kx8 SRAM whereas the 4116 is

    - Franc Zabkar
  19. Mike Wynne

    Mike Wynne Guest

    That works. Also, sometimes you can find the faulty chip by sticking your
    finger on each DRAM in turn and seeing if any are noticably hotter than the

  20. Brane2

    Brane2 Guest

    True. My bad. I meant 4164. It would be kinda hard to squeeze 6164 as a
    pin compatible replacement without a hacksaw and some heavy magic ;o)
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day