Connect with us

Using existing pinball switch to fire external circuitry ...

Discussion in 'General Electronics' started by Wes, Jul 30, 2003.

Scroll to continue with content
  1. Wes

    Wes Guest

    I would like to utilize the closure of an existing switch in my
    pinball machine (Bally KISS) to trigge
    r external circuitry. My initial
    idea is to have a spinner switch create a pulse that would serve as a
    clock for a decade counter (4017
    ).

    So I researched the fundamentals of the switch matrix and how it is
    scanned by U10 on a -35 MPU. The
    manual shows the 2 wires (poller/send + switched return) on the MPU J2
    connector to be at pins 4 and 14, respectively.

    So initially, I figured that I could simply tap off these 2 wires and
    feed them into an AND gate, usin
    g the AND output for the clock pulse into the 4017 counter. BUT, here
    is what I have found, and I'm h
    oping someone can help me accomplish the goal. I'm thinking a
    comparator or some sort of A/D converte
    r is needed, but I'm inexperienced in the application/use of linear
    ICs.

    When I measure the voltage of the strobe/send wire (J2 pin 4), I do
    get the expected pulse and I was a
    ble to clock a 4017 directly with this send. Obviously, this doesn't
    isolate the closure of the spinn
    er switch, though.

    The voltage at the RETURN (J2 pin 14) is where I start to lose it. I
    expected it to be near 0V when t
    he switch is open, and near +5V when the switch is closed. But what I
    find is that this pin seems to
    be strobing, and when I measure the voltage, I get around +3.2V when
    the switch is open, and around +3
    ..4V when closed. I WAS able to see that this .2V delta is consistent
    with other switches in the same
    column of the matrix, so ....

    I do seem to have almost what I need - I'm thinking I can determine
    (theoretically) that the switch mu
    st be closed if:
    1) the send wire is "high"
    2) the return wire is relatively "highER" (the .2V delta)

    So, I THINK if I could just translate the +.2V delta into a logic
    "high", then I could feed the AND ga
    te as I originally thought I could do. Any insight into how I could
    make this work would be greatly a
    ppreciated.

    I do realize that there are other (possibly better) ways to accomplish
    this end goal - i.e. an isolate
    d parallel switch, an opto-decoupler - BUT I'm intrigued by the idea
    of tapping directly into the swit
    ch matrix, and if I could get a handle on it, it would open up a world
    of ideas for modifications with
    easy installation (piggy back off the MPU J2 connector).

    Thanks! -Wes
    ~
     
  2. Rich Grise

    Rich Grise Guest

    (Wes) wrote in message
    Well, you've figured out that it's strobed (or scanned) - this is a
    Good Thing. :) So what you want to do is sample that particular
    switch's return line _while_ it's being strobed. And since the
    spinner is so much slower than the scanner clock, you don't want
    to gate the clock pulses with the spinner switch - clock a D
    flip-flop with the strobe line, and get its input from the other
    side of the switch. (possibly conditioned). That should give you
    a reasonable approximation of a rectangular wave which shows
    the current state of the spinner switch.

    Good Luck!
    Rich
     
  3. Wes

    Wes Guest

    Thanks Rich! This has helped me out tremendously, although I still
    can't get the desired result.

    Your mentioning that the switch contact is much slower than the strobe
    frequency made me see that I was on a horribly wrong track. Since
    closure of the switch will only "make a copy of" the strobe signal on
    the return line, by feeding an AND gate with these 2 lines, all I'll
    get is the same strobe at the out of the AND - not at all what I want
    to clock the 4017.

    But I just don't understand what I'm seeing - the return line is
    pulsing, whether or not the switch is closed. I would have thought
    that as long as no switch on that return line was closed, its voltage
    would be "low". And I would have thought that if ANY switch on the
    return line was closed, it would pulse at the rate of the send for the
    closed switch. Not the case - that return line just pulses away
    regardless of the switch's state...

    When I hooked up the D FF as you suggested (clock with send, use
    return line of switch for data), the Q output just pulses away whether
    or not the switch is closed (it's a nice steady pulse). That seems
    really weird to me - since we're tracking the return line ONLY when
    the send line is going high, it seems like a steady pulse wouldn't
    happen.

    Here's one thing that occurred to me: since the FF will clock on the
    rising EDGE of the send pulse (thus transferring the D input to Q
    output), and the D input on the other side of the CLOSED switch is
    just a "copy of" the send/clock, won't the D input then be in sort of
    a transient state (going from low to high just as the clock is), thus
    making it a bit unpredictable as to what will get transferred to Q?

    Why would that return line be consistently pulsing? Could it be just
    be the effect of the pin being READ? I would have thought since it's
    an INPUT pin to the PIA, its being read wouldn't affect its voltage.

    See, this is how I understand the switch matrix works:

    There are 8 sends and 8 returns. Each send drives a column of 8
    switches, and each return "reads" a row of 8 switches. Then in a
    round robin fashion, this algorithm happens:

    make only first send high
    read each return, sequentially
    (1 closed switch on send line will cause
    high to appear on 1 return line)
    make only second send high
    read each return, sequentially
    (1 closed switch on send line will cause
    high to appear on 1 return line)
    etc....

    Could it be the sequential READING of the return lines that makes it
    pulse?

    - Wes
     
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

-