8088 Follow-Up

Discussion in 'The Projects Forum' started by nDever, Jan 24, 2011.

  1. nDever

    Thread Starter Active Member

    Jan 13, 2011
    154
    4
    Hey Guys,

    I have a relatively complete 8088 system built and running but the issue with this system is that it runs sometimes. What I mean is that when I connect the circuit to power, sometimes the system runs and sometimes, it does not.

    When the circuit finally does work, if I trigger the reset button on my 8284A clock chip connected to the 8088, sometimes it resets the processor as it should, however most of the times, it stalls the processor.

    I think it may be undercurrent since the 8088 can draw as much as 350mA max. If the circuit was not receiving adequate current, should it be working at all?
     
  2. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    How many filter caps do you have on the chips? This sounds like it could be the classic problem, spikes propagating on the power supply feeding other chips.

    General Electronics Chat Sticky Decoupling or Bypass Capacitors, Why?
     
  3. nDever

    Thread Starter Active Member

    Jan 13, 2011
    154
    4
    For every one chip I have one .1uF cap to absorb some voltage.
     
  4. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Do you have an oscilloscope to look at your power supplies?

    A schematic would help, as would pictures of the layout.
     
  5. nDever

    Thread Starter Active Member

    Jan 13, 2011
    154
    4
    I apologize for the shoddy quality. There should be the image of the entire system and one of the peripheral interface board.

    I do not have a schematic available yet but to describe the layout, there are a total of 6 breadboards, the mainboard is made of 4 of them; they are connected physically and electrically. The other two is the peripheral interface board.

    I read the suggestion of the amount of bypass caps to place on the board, but I do not have 30 capacitors. Since the mainboard is all connected to the same bus, I could put just 6 caps on the main source.
     
  6. GetDeviceInfo

    Senior Member

    Jun 7, 2009
    1,571
    230
    one thing that I found most useful for troubleshooting similar circuits was a decade counter used to continously reset your system, along with an oscilloscope. Depending on what you use to drive the counter would set you scopes reference via the reset line. For a system that won't get going, I'd clock right to the system clock and watch for the device to cleanly fetch the first opcode.

    But now I've seen your images. My experience would tell me to not go any further, get a good perf board and point to point solder each operational section at a time. Your current attempt with breadboarding will be nothing but frustration.
     
  7. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    What frequency are you clocking this system? GDI is referring to the fact that protoboards have very lousy performance at high frequencies (such as 1Mhz). You might try clocking your system way down, say 100Khz or less.

    I don't know how you are lighting up the LEDs, but if it is with CMOS chips you will need buffers. Something like this will work well...

    [​IMG]
     
  8. nDever

    Thread Starter Active Member

    Jan 13, 2011
    154
    4
    I thought that the clock rate was a bit high for initial performance. I have an 8088-2 which is running at its maximum rate of 8MHz. At first, I thought that since the processor was clocked at its maximum, it may be running unstably.

    At the moment, all that I have available to me is a 24Mhz crystal which is fed through the 8284A clock generator which divides the crystal frequency by 3 and at a 33% duty cycle. I do not have any other crystals but I have been trying to generate lower frequencies with flip-flops.

    The LEDs are attached to a 8255 PPI which is TTL compatible according to the datasheet.
     
  9. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Anything over 1Mhz (OK, 100Khz) probably should be protoboarded, or have a custom PCB made. A protoboard is a series of capacitors sitting next to each other for each row of pins. This is why they don't handle high frequencies well.

    Like I said, you can compensate by lowering the clock rate dramatically. I don't believe the chip will care.
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Long interconnecting wires + high clock speeds on a breadboard = big trouble due to the parasitic inductance and capacitance. Your wires will "ring" like mad when they change logic states, sort of like whacking a bell with a hammer.
     
  11. nDever

    Thread Starter Active Member

    Jan 13, 2011
    154
    4
    Will a PCB eliminate a significant amount of noise? How can I supply my circuit will a smooth, clean stream of DC voltage so that I no longer have to connect bypass capacitors to the power bus?
     
    Last edited: Jan 24, 2011
  12. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Sorry, bypass capacitors are with us always. My Bread and Butter in CRG (Customer Returned Goods) for Collins Radio was bypass caps that had shorted out. But they perform a very necessary function, absorbing the digital spikes the various chips create as they go about their business. Otherwise the spikes will go down to the next chip, creating havoc everywhere.

    The long wires are a different issue, though the symptoms look the same. They are like antennas, radiating and receiving signals, most of which are generated by other parts of the board your are working on. The shorter the trace, the better it works.
     
Loading...