Printer port question

Discussion in 'Computer Information' started by Robert Baer, Jun 15, 2013.

  1. Robert Baer

    Robert Baer Guest

    According to the original IBM documentation viz IBM Technical
    Reference First Edition August 1981 page D-34 Parallel Printer Adapter
    schematic, the following pins of the DB-25 printer connector are
    exclusively inputs:
    pin 13 is +SLCT (select, high for "1")
    pin 11 is +BUSY (busy, high for "1")
    pin 12 is +PE (paper out, high for "1")
    pin 10 is -ACK (acknowledge, low for "1"; eg: inverted logic)

    So, how come with computer NOT CONNECTED to anything (NO power at
    all), one sees that pins 10, 11 and 12 are SHORTED TO GROUND?

    Please explain how a printer can possibly work with that condition.
     
    Robert Baer, Jun 15, 2013
    #1
    1. Advertisements



  2. Because they are held high by the printer that's not there? The printer
    pulls these lines high.
     
    Jeff Strickland, Jun 15, 2013
    #2
    1. Advertisements

  3. Robert Baer

    Robert Baer Guest

    On the surface,that _might_ sound reasonable..
    Except.
    The schematic shows INPUTS to a TTL gate for each one.
    On TTL logic, an input must be pulled down (below 1.4V) for a low to
    be asserted; a floating input on a gate acts like a high.
     
    Robert Baer, Jun 15, 2013
    #3
  4. Robert Baer

    Paul Drahn Guest

    Perhaps you have a computer with a version 2 parallel printer port. Not
    quite the same as the original IBM port.

    Paul
     
    Paul Drahn, Jun 15, 2013
    #4
  5. Robert Baer

    Paul Guest

    I would question how you determined "SHORTED TO GROUND".

    A short to ground, will read zero on any range on your ohmmeter.
    Including the diode range.

    If you use the diode range on the multimeter, and you're seeing
    a diode drop (0.5 to 0.7V), then, it's not a "SHORT TO GROUND",
    is it ?

    Remember that, when a hardware circuit is not powered, the
    VCC and GND appear shorted together. The protection network
    on the input of the gate, then becomes involved while you're
    using your multimeter. Some gates, have traditional protection
    diodes to VCC and GND (to prevent overshoot, to prevent
    undershoot). And a few gates are protected by what we used
    to call "up up up down", which is three diodes to one rail,
    and one diode to the other rail. Which allows larger
    excursions past the rail in one case than the other.

    So when you characterize a circuit, you have to know something
    about the internal protection structure, whether it's one
    stage or two, has a series resistor between stages, and so on.
    You need to know the protection structure, to interpret the
    results. In some cases (inputs "open to the world"), you'll
    actually see external protection diode networks, intended
    to provide a measure of ESD protection perhaps. The multimeter
    can end up probing the protection structure, instead of telling
    you something about actual circuit operation.

    You can use a variable power supply, and a resistor, to do
    your own crude characterization of the circuit while it
    is running. On a TTL gate, perhaps the knee is at 1.3V
    on the input. Applying a voltage below 1.3V, through a
    resistor, should sink a small current (400uA on a TTL
    gate of modern vintage, 10uA on CMOS maybe). Applying a
    voltage above 1.3V might source a small current, and the
    current will likely be a smaller flow than in the logic 0
    test case. It's a bit like running a curve tracer,
    in your head.

    So there are a few things you can try, with a bit of
    crude equipment. And your friends Thevenin and Ohm's Law
    ("It's The Law").

    *******

    If you see an actual shorting strap on the PCB, then
    that's another matter. I don't know anything about
    printer ports, so can't really comment on that.
    Status inputs usually need to remain functional,
    especially if they mediate the transfer rate in
    some way. Shorting all the status signals would
    not make sense. But if this is a question of
    characterization of an unpowered logic gate,
    go back and look at it again.

    Paul
     
    Paul, Jun 15, 2013
    #5


  6. Why do you care what the parallel port is doing? They have not sold a
    parallel printer in almost a decade. Surely if you have a parallel printer,
    you deserve a new one for Father's Day.
     
    Jeff Strickland, Jun 15, 2013
    #6

  7. If there is a gate array, then the gate is low until a printer comes along
    to pull the inputs high when the condition becomes true. If the printer is
    available, the select line goes high, if the printer is busy, the busy line
    goes high, if whatever is happening, the line goes high, or it goes low in
    the case of the not-logic.

    I don't understand the question. I really don't understand why you think
    something is broken on a port that has not been used in any current
    technology in the past 15 or 20 years. I guess that's not true, I'm using a
    Brother MFC that is connected to the parallel port. It could be connected to
    a USB port, but I'm running out and I had a parallel cable laying around.
     
    Jeff Strickland, Jun 15, 2013
    #7
  8. Robert Baer

    Paul Guest

    People use printer ports for GPIO functions. For example,
    I have a JTAG programmer cable, that plugs into a parallel
    port, and needs a driver installed to gain access to the
    appropriate pins on the interface. And I don't think my
    adapter, works on the end of a USB to printer dongle either - it
    has to be a real parallel port. I use a PCI Express card
    with parallel port connector, to move the JTAG cable to
    my current machine. That's because parallel ports are
    out of style now, so you pay extra to get them. This is
    the card I'm using.

    http://www.startech.com/Cards-Adapt...ile-Parallel-Adapter-Card-SPP-EPP-ECP~PEX1PLP

    That card uses OXPCIe952. Which apparently has more functions,
    than there are connectors on my card.

    http://www.plxtech.com/products/uart/oxpcie952#technicaldocumentation

    Paul
     
    Paul, Jun 15, 2013
    #8
  9. Robert Baer

    Robert Baer Guest

    NOT relevant!
     
    Robert Baer, Jun 16, 2013
    #9
  10. Robert Baer

    Robert Baer Guest

    * YEP! That is EXACTLY what i have; been in electronics for about 60
    years. I meant what i said and i said what i meant.
    * Your statement is correct, but i saw ZERO which is a SHORT; see above.
    * Again,what you say is correct and a good lesson for those that are
    weak in electronic theory.
    But (again) not relevant.
    * Powered or unpowered makes no difference; seeing ZERO means ZERO.
     
    Robert Baer, Jun 16, 2013
    #10
  11. Robert Baer

    Robert Baer Guest

    I want to use the four lines for nybble input data reads.
     
    Robert Baer, Jun 16, 2013
    #11
  12. Robert Baer

    Robert Baer Guest

    Get almost the same thing for about $17 (while they last):
    http://www.jameco.com/webapp/wcs/st...Id=10001&ddkey=http:StoreCatalogDrillDownView
     
    Robert Baer, Jun 16, 2013
    #12
  13. Robert Baer

    Paul Guest

    OK, as a confidence builder, do a lap of the pins, and tell
    me *all* of the pins that read zero ohms. For example, here,
    the D-34 has signal GND on 16, 19 thru 29, and 33.

    http://www.interfacebus.com/Design_Centronics_Connector_PinOuts.html

    Even if you'd managed to mirror image the thing, the 19 thru 29 are
    on the wrong row for that. And you're not likely to flip the thing the
    other way, and mix the rows up, because the shell shape gives a hint.

    Paul
     
    Paul, Jun 16, 2013
    #13
  14. Robert Baer

    Robert Baer Guest

    Finally figured it out.
    Had made a new cable with flat ribbon cable and an IDE connector or
    the PC using a hand vice.
    There is a goodly separation between the forked pin rows, and if the
    cable is not aligned exactly right,the edge of one can (also) cut into
    the next wire,giving a short when connector is plugged in.
     
    Robert Baer, Jun 17, 2013
    #14
    1. Advertisements

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.