VIC-20 restoration

Discussion in 'Retro & Arcade' started by callan, Apr 24, 2017.

  1. callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    I finally got around to having a decent look at the rescue VIC20 sent to me by badmofo. As I suspected after an intial look yesterday, I confirmed that there is no master clock signal. The likely cause is a failed 6561 VIC chip, or less likely the 74LS02 quad NOR chip. (reasoning below). Thort you guys might be interested (and badmofo did request a worklog :) )

    1Mhz clock signal from my working Rev1 VIC20 (Yellow line)
    [​IMG]

    Clock "signal" from the rescue rev II VIC.
    [​IMG]

    In the VIC20 the 1Mhz clock is generated by the VIC (Video) chip from a dual-phase dot clock. (14.318Mhz NTSC, 4.4Mhz PAL).


    The clock/video circuitry on the board itself:
    [​IMG]


    An annotated extract of the schematic shows the clock circuit.
    [​IMG]

    Crystal Y1 generates a 4.4Mhz signal, cleaned up and split into 2 signals 180° apart by the 74LS02. The dual output is fed into the 6561 VIC chip pins 38/39.
    The VIC chip divides down this signal to 1Mhz and feeds it out via pins 35 and 36 (also 180° apart). The signal from pin36 is not used.

    It is this signal that is missing: as per the above image, it is flatlined.

    Now the only thing that makes me suspect that the 74LS02 might be at fault is that the voltage of the 4.4Mhz signal from the rescue VIC20 is roughly HALF that of the WORKING VIC - so it's possible that the clock signal input is not sufficient to trigger the clock generator in the VIC. Whether that's due to exxcessive load pulling down the signal from the 72LS02 or VIC against the pullup resistors I'm not sure (the signal is pulled LOW from VCC (which I confirmed is correct), not HIGH from ground). NOTHING is socketed in this board which is a pain, but 16 pin 74LS02's are a damn sight easier to get and work with than 40 pin 6561 VIC chips.

    Rescue VIC:
    [​IMG]

    Working VIC:
    [​IMG]


    Next stop is to dig up a 74LS02 and drop it in. Being a basic 74LS series chip and only 14 pin it's simple enough to swap out, even it it's not likely to be the problem. I suspect the problem will persist, but then at least I'll know where next to start scrounging.

    Callan
     
    Last edited: Jun 9, 2017
  2. adz

    adz Member

    Joined:
    Oct 2, 2003
    Messages:
    1,151
    Location:
    Melbourne
    I'll get the popcorn :thumbup:
     
  3. badmofo

    badmofo Member

    Joined:
    Oct 10, 2011
    Messages:
    612
    Location:
    Australia
    Quite the turnaround of events for this little machine - abandoned in a pile of soon-to-be scrapped keyboards one day and world famous recipient of a Callan work log the next :thumbup:

    Looking forward to seeing this roar back into life.
     
  4. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    CO0uld take a bit of time if the 6561 is at fault: they're a bit tricky to source since CSG disappeared in a cloud of financial collapse and environmental pollution in the early 90's:Paranoid:.

    I'll socket both chips when I replace the 74lS02, though.

    Callan
     
  5. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,297
    Where are you getting the signal you’re showing?
    The messy one doesn’t look like the output of a clock buffered with Schmidt triggers.
    How does it look on the way in to the custom chip? Straight after the buffers.

    If the clocks on pins 38 & 39 are clean, can’t you then assume you’re screwed?
     
  6. JidaiGeki

    JidaiGeki Member

    Joined:
    Jun 28, 2001
    Messages:
    817
    Location:
    Sydney, NSW
  7. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    Buffers?? That's hilarious. The VIC20 design was cobbled together in about 8 weeks, and the first prototype was (literally) a sawnoff PET board with a breadboard glued on, jammed in a calculator case. The VIC chip itself was designed way back in 1977.
    No, non of the signals are cleaned up: the only buffers on the VIC20 board are a 7406 driving the disk drive port: hell even the INPUTS on that port are unbuffered:wired: We're in the wild west now. The C264 series (plus4/C16) even had the keyboard directly wired to the data pins of the main CPU BUSS... So were the joysticks, until that proved a step TOO far.

    The signals are taken direct from the VIC inputs/outputs and/or the quad NOR chip.
    The probe is pretty average, as is the oscilloscope, and early Commodore/MOS NMOS stuff was pretty dirty anyway. I wasn't that picky about the ground I used, either:)

    No, I can't assume I'm screwed as there is a clear difference in the 4Mhz signals between my control system (albeit different rev.) AND noticably LESS than 5V). It is HIGHLY LIKELY that the VIC chip is the problem, but sometimes when troubleshooting something like this if something involved can be quickly swapped out and eliminated then you might as well do it, even if it's not the most likely culprit. I'll socket both the VIC and 74ls02 as soon as I can, and see if I can scrounge at least an LS02 from somewhere. (might have one in my spares tubes.)

    Thanks. I'll certainly be looking around: they've become a lot scarcer over the past few years. Date code won't matter: Unlike the move from NMOS to PMOS with the C64 VIC/SID chips, the two-odd revisions of VIC20's have interchangable core chips.

    To our collective dismay a cottage industry hgas sprung up in the last 10 years: remarking, counterfeiting and outright fraud trading in rare Commodore chips: Due to the proliferation of retro synth scene (and standalone/MIDI SID-based devices) the SID 6581/8580 is the most common victim.
    The most common chips to fail in C64's: the SID and PLA chips have serviceable modern-day replacements/recreations, but not so the VIC20 VIC chip or, more concerningly the C264 series TED chip. (which is both rare AND terrifyingly fragile.)

    If I DO need to go down the path of getting a replacement I'll seek contacts through the startlingly active Commodore hacking community.

    There is a period of time after the peak of popularity in an artefact that they become worthless. Society finds something shiney and the old is tossed aside. Those wise and patient enough hold onto them, and appreciate them for their native, not monetary worth is. Then there's a bit of a mad burst of short-lived retro-nostalga, then they fade away to a hardened few. People whisper of "Barn Finds", Fleabay fills with opportunists and scammers, and then it all fades away into an almost underground industry.
    Some stay that way (Old pocket watches, for example), some crystallize into something more long-lasting (think old cars), and then - erm - there's Magic the Gathering cards: They'll be worth a mint SOON!!!!!! You'll see.:p
    Where we're at with Vic20's - not sure.


    Callan
     
    Last edited: Apr 24, 2017
  8. shane41

    shane41 Member

    Joined:
    Nov 10, 2008
    Messages:
    5,353
    Location:
    on the edge
    Who are you man :wired: sometimes I wish I studied more on electronics
    Want some more projects Callan, I have my fair share & then some
     
  9. @rt

    @rt Member

    Joined:
    Nov 30, 2005
    Messages:
    2,297
    Ok, not Schmidt trigger inputs, but the NOR gates are buffers for the crystal,
    and the overall output is a logic clock yes?
    With all of their individual inputs tied together that makes them all inverting buffers.
    180 degrees out of phase is another term for one output being NOT the first,
    since you are in the digital domain by that stage.

    I was asking what the difference is in the appearance of what should be a square wave from pins 38 & 39,
    to pin 35 (the clock output of the custom chip).

    If both of the outputs from the chip you don’t like to call buffers are square,
    and the output of the custom chip is a mess, that’s when I was suggesting you’re screwed.
    If anything in the top part of the circuit is a mess, then no problem that is difficult to fix (at least on paper).
     
    Last edited: Apr 25, 2017
  10. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    The main difference is that the 38/39 input is at 4.4Mhz, and the pin 35 output is at 1Mhz. There is obviously some other interference - most likely an artefact of my dodgy oscilloscope setup that's superimposing itself on the 4Mhz signal, but I'm not too concerned about it at this stage as the WORKING vic had the same intereference. The total absence of clock OUTPUT is the clear difference.

    Got a bit of time this afternoon...

    [​IMG]

    [​IMG]

    I seemed fresh out of 14 pin sockets, so I pushed out the last 2 pins of a 16 pin socket and fitted that. Not ideal I know - but there was the room in the cannister, and with the pins removed it's very clear where the chip should go.

    Got a pretty full dance card over the next few days, so I suspect there won't be much progress till the weekend or next week.

    Callan
     
  11. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    BAD news, and GOOD news.

    I popped out and picked up a 74LS02 today, and had a play. (Never sure about replacing 74 parts with 74LS when working with clocks, but usually seems to work).
    edit: my earlier post had the original as 74LS - that's incorrect: it was a non-schottky plain jane 74 part. The schematic has it as the LS variant, though.

    The BAD news is: The VIC is stuffed.

    The Good news:

    That's ALL that's stuffed. (or at least stopping it from running - haven't been able to test the 6522's. That's less of an issue: they can still be bought NEW today: WD make them - with working shift-registers and'all :)

    I dropped in the 74LS02 and tried it - sure enough - clock in, no clock out, no video.

    Reluctantly I dropped the faulty VIC in my working PET keyboard VIC20 - same symptom (no clock out)

    Then equally reluctantly I dropped my only GOOD 6561 VIC in the old board, AND...

    [​IMG]

    It fired up just fine.

    [​IMG]

    So now it's off to get a replacement VIC.

    Progress:thumbup:
    Callan
     
    Last edited: May 9, 2017
  12. badmofo

    badmofo Member

    Joined:
    Oct 10, 2011
    Messages:
    612
    Location:
    Australia
    Great to see some progress, and very pleased to see you didn't kill your healthy hardware in the process - always a nerve-racking moment that!
     
  13. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    Ya. All good, though - I'm very protective of my baby. My working VIC is quite an exemplar: immaculate condition (albeit slightly yellowed), so early it needed an FCC exemption (Vic is dated first week 1982), and completely unmolested. Probably one of the better items in my collection.

    [​IMG]

    [​IMG]

    [​IMG]

    Got a few leads on where to get a replacement 6561 on CBM hackers forum last night: I'll be chasing them up.

    Callan
     
  14. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    Just a small update.

    I seem to have secured a source of some tested 6561 VIC chips from a fellow in Germany. Mite take a while to get here, but at least the ball is rolling :)

    Callan
     
  15. badmofo

    badmofo Member

    Joined:
    Oct 10, 2011
    Messages:
    612
    Location:
    Australia
    Excellent.

    All this Commodore talk is getting me in the mood for some C64 action - I've been meaning to try out 'Kings of the Beach', so I think tonight's the night.
     
  16. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    The VIC 6561's Teutonic dispatch is now confirmed:thumbup: Once it's arrived and I have the mainboard nominally working I'll restore it a bit - replace dodgy caps, clean it, tidy up some of the repair solder from a previous replacement of a 6522VIA etc. etc.


    So, with the chip some weeks away my attention now turned to the keyboard.
    Aside from being ABSOLUTELY filthy there was also a missing MINUS (-) key and the spline for the key snapped off.

    So I split the keyboard, providing access to the switch contacts, allow removal of the keyboard movement and replacement of the broken spline.

    [​IMG]

    I tied the keyboard cable to the backboard, as unsupported it's individual wires are prone to fracture.

    [​IMG]

    Then I turned my attention to replacing the broken spline.

    It was at THIS point that I said.... "Bother". For like the donor C64 keyboard this Vic20 keyboard is made by Mitsumi, but the movement is significantly DIFFERENT from the donor keyswitch:sick:
    The shaft is shorter and fatter, the spline dimensions different, the distance between the base of the key and the end of throw slightly greater - and finally the spline recess is less deep into the 2-shot cast key. They will not fit.

    [​IMG]

    [​IMG]

    So now I need to precisely measure up both keys and attach a sawnoff portion of the REPLACEMENT spline and shaft onto the EXISTING shaft. Attaching it won't be difficult: I'll just drill the existing and replacement shaft , run a steel pin through both and infill with a bit of 2-part polystyrene casting goop and it'll be quite strong. Determining the dimensions of said replacement shaft, however will be quite 'interesting' - looks like I'll be digging up my vernier gauge and doing some maths tomorrow.

    At least the key colours match up - gotta be happy about SOMETHING!

    Callan
     
    Last edited: May 16, 2017
  17. underskore

    underskore Member

    Joined:
    Nov 5, 2002
    Messages:
    4,148
    Location:
    3147
    it's great to see such a detailed and technical worklog ticking along. :thumbup::thumbup:
     
  18. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    It is tempting just to replace it with a C64 keyboard outright (they're a dropin replacement), but I wouldn't be able to get the function key colour correct.
    This keyboard has actually seen a lot of action - the splines have significant scoring (which I guess is a good thing as it shows it was loved and used). The whole thing will need to be disassembled, cleaned and serviced. The painted on graphics key images are still in good nick, however.

    Neat tip: Mr Sheen furniture polish is simply silicone oil in solution. Sparingly and strategically applied it lubricates plastics brilliantly - the carrier liquid evaporates off leaving a powdery silicone film which doesn't wick, spread or attract dust, lubricates well without attacking the plastic.

    I'll probably do the keyboard restoration before repairing this one key, as I don't want to disturb the repaired keyswitch anymore than is necessary.

    Callan
     
    Last edited: May 16, 2017
  19. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    Keyboard Day!!

    [​IMG]

    Blergh!..

    [​IMG]
    That dust is really gritty and abrasive, too - it's scored the keyboard shafts and bound up the shift-lock mechanism. Nasty stuff. Could be worse I suppose: at least the former owner wasn't a smoker.


    Callan
     
  20. OP
    OP
    callan

    callan Member

    Joined:
    Aug 16, 2001
    Messages:
    4,435
    Location:
    melbourne
    Late last night I ransacked my scrap pile and came across a very dead Rev1 Vic20 (from Girton College, Bendigo) that I'd forgotten I had. Bit of a tragedy as it's a VERY early example (16th week 1982). It's a PET keyboard with bizarre reworking in wirewrap under the backboard, but quite unsalvagable. The case is in good condition however, but more importantly it has a seemingly working VIC6561 :thumbup: So for now I've dropped it into the project VIC, and in the next few days I'll work on the board. Having a tooth implant and some minor face surgery tomorrow, so I might be out of commission for a bit.

    When the German VICs arrive I might still have a go at repairing the Girton machine, but it seems pretty far gone. Schools are not kind to computers.

    I had mixed success with the keyboard repair: still a work in progress.

    Pics later.

    Callan
     

Share This Page