intermittent logic

Discussion in 'General Electronics Chat' started by DexterMccoy, Apr 10, 2014.

Thread Status:
Not open for further replies.
  1. DexterMccoy

    Thread Starter New Member

    Feb 19, 2014
    429
    2
    When troubleshooting Intermittent logic problems , what are the best ways to isolate and find the root cause in logic intermittent problems?

    Can you list some logic intermittent problems you have seen or that is common?

    What I have seen or troubleshoot logic boards that had intermittent logic problems was

    1.) When turning the power of the circuit board OFF and ON would make certain Logic gates, Logic IC chips to output a glitch or output a voltage, output a spike which would trip or latch other IC components

    How would you guys approach this type of troubleshooting? any skills?
     
  2. profbuxton

    Member

    Feb 21, 2014
    233
    68
    (1) Use a logic probe with memory feature. Simplest but not always possible depending on how the problem is triggered or how often it shows up.

    (2) Use a digital logic analyser to look at all the signals of interest.

    There are no "lists" of intermittent logic problems or ones that are common. The vast array of electronic equipment can have any sort of intermittent problem you may care to imagine and some you may not see but once in a lifetime.

    If it is switching power on/off causing spurious glitches in the logic of interest I would suggest that it needs redesign by a competent electronics designer.

    As far as skills go you would need to have a adequate basic knowledge of the operation and faultfinding of electronics and correct use of suitable test equipment. From the level of this question (and others from you) I would suggest you enrol in a good quality training course if you want to be able to carry out these sort of tasks.
     
  3. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    Intermittent problems are the worst type of failures, for no other reason beyond you can never be sure they have been repaired.
     
  4. DexterMccoy

    Thread Starter New Member

    Feb 19, 2014
    429
    2
    How would a logic probe with memory feature help me find a logic glitch or logic intermittent problem? how would you use it?

    The Design is good, it just the Logic IC chips go bad after ageing

    When I switch the power on and off , it causes some Logic IC chips to Output a False logic state
     
  5. MrChips

    Moderator

    Oct 2, 2009
    12,410
    3,353
    Logic IC chips don't go bad with aging.

    The design is flawed.
     
    shortbus likes this.
  6. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    Deep memory Logic analyzer, somewhere on the Agligent site is a manual for a old 16500C logic analyzer (I have this machine), they did a learning board for B model and that manual details how to look for glitches.
    In short hook it up to all chips of interest and set it to trigger on the glitch, then get it to glitch, this will show you what was going go and where it was happening.
    Thats roughly the idea anyway
    What logic chips are they? Have you done the obvious and resoldered the chips in case of dry joints? Checked all caps (if any)
     
  7. profbuxton

    Member

    Feb 21, 2014
    233
    68
    If you don't know how to use a logic probe how do you expect to work on any electronic equipment.
    My suggestion you get some real training still stands.
     
  8. DexterMccoy

    Thread Starter New Member

    Feb 19, 2014
    429
    2
    No I mean how can a logic probe with memory get me any where to finding which Logic IC chip is causing the glitch or false triggering?
     
  9. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    Why not just use the equipment designed to do the job as I outlined above! They even give you a clue in the name of the machine. LOGIC.........ANALYZER.
    Or in case you have missed the connection, a device to ANALYZE LOGIC CIRCUITS.
    Have you never seen or used one?
    Did you check solder joints? A LOGIC ANALYZER will tell you what chip glitches and what happened before and after the glitch (if you use the trigger correctly of course and set it up).
    A logic probe with fast pulse detect will also spot a glitch (some times), but first you would need to know if the pin your testing is meant to pulse or not, there is your problem because in order to do that you would have to understand the circuit. And you dont, you are trying to trouble shoot problems on boards by asking how to test things.
    What you should be looking at is how things work, and understanding the circuits you are trouble shooting. The circuit and its type will to some extent dictate the best approach to trouble shoot it.
    Take say a Test technician or engineer, they are given a non working board and a schematic, because they know about the components they are dealing with they are able to determine the best way to approach the problem. They would have looked at the circuit to see how it is meant to function, then go through it and test each function. Your main problem is your not willing to learn basic electronics, you just want to know how to find X Y or Z fault, but you dont understand that before you can do that, you need to understand the material your dealing with.
    Start with the books I linked too, then build up from there. Master the multi meter first before trying the harder stuff.
     
    shortbus likes this.
  10. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    Logic analyzer is not a good tool for finding glitches. You need a high speed oscilloscope for that.
     
  11. DexterMccoy

    Thread Starter New Member

    Feb 19, 2014
    429
    2
    Why a high speed o-scope, what MHZ?

    What mode should I use? single sweep? normal mode?

    Why can't a low speed O-scope find glitches?
     
  12. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    Glitches are very short duration signals. Low speed scopes will miss them. You need a very speed (GHZ) scope running in a mode that shows false colors determined by statistical signal characteristics. Single sweep and normal are irrelevant to these modes.
     
  13. Little Ghostman

    Member

    Jan 1, 2014
    294
    97

    I would disagree with that, I have a hp 16500c LA, with built in scope, mine came with a teaching board and course, in went into great depth into setting up the triggers etc for finding glitches.
    Its a great tool for finding a glitch, the complex triggering has many setting for glitch finding.
    And to be fair I did mention this earlier, I wasnt saying something like a pc LA will be upto it, but alot of the top end LA's are designed to be able to locate Glitches.
    One setting on mine allows you to record a sample (user determined lengh) of a set of chips doing there thing, there is then a simple command to tell the LA and scope to only trigger when the signal differs from the sample, you even get the option to specify how long the glitch has to be before triggering, Although a old obsolete machine now, it will run rings around a scope looking for occasional glitches. Then there is also the built in pattern generator that can be used to try and force a glitch by outputting a set sequence that was recorded when a glitch happened.
    So in this instance having owned and used the machine and done the course with it, I would respectfully disagree with you completely on this.
    Also keep in mind you get on average a max of 4-6 channels on a scope, on a decent analyzer you can monitor over 300 pins simultaneously. So I would say a LA in this situation will out perform a Oscope, especialy if your unsure whats causing the glitch
     
  14. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    The sample rate is too low for glitch finding. I've used that logic analyzer for decades, but never to find glitches.
     
  15. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    Also your not taking into account the kind of logic chips he is using! I bet most are over 30 years old
     
  16. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    How do you know I'm not taking the kind of logic chips into account? Do you read minds? Sample rate for your LA is too low for glitch finding.
     
  17. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    http://sigma.ucsd.edu/Facilities/manuals/16500.pdf

    page 7 near the bottom of section 8,
    and this one is for models upto the 16500B the one I have 16500c is better. But the traing book clearly details how to trouble shoot glitches with inter module triggering
     
  18. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    this is taken from the manual of the lowest in the family range the model A, I have used the traing board to do these exercises.
    "Configure the Intermodule Bus
    In this exercise, you will configure the Intermodule Bus so that the timing
    analyzer arms the oscilloscope. You have seen in the "Run the Oscilloscope
    in the Repetitive Mode" exercise that the oscilloscope will not reliably trigger
    on the glitch.
    When you complete this and the following exercises, the timing analyzer and
    the oscilloscope will be configured so that the timing analyzer’s glitch
    triggering capability will capture the glitch. The timing analyzer will then use
    the Intermodule Bus to arm the oscilloscope so that it, too, captures the
    glitch."
     
  19. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    That's for the oscilloscope, not the logic analyzer. And, that's for a phony glitch generated from a training board. In real life, it would be much more difficult to find a glitch in repetitive mode, just as the warning on the page tells you.

    The glitch generated by the training board occurs more regularly than it would
    in a malfunctioning system. Therefore, you will see the glitch on consecutive
    traces.
     
  20. Little Ghostman

    Member

    Jan 1, 2014
    294
    97
    No read it correctly.
    In this exercise, you will configure the Intermodule Bus so that the timing
    analyzer arms the oscilloscope.


    in the Repetitive Mode" exercise that the oscilloscope will not reliably trigger
    on the glitch.

    Capture the Glitch with the Timing Analyzer and
    Oscilloscope
    1
    Touch the Group Run field then, without lifting your finger, drag it to
    Single in the pop-up and lift your finger.
    The timing analyzer will run, and once it triggers, it will arm the oscilloscope
    via the Intermodule Bus. Since the oscilloscope will not always trigger on the
    glitch, the glitch trigger of the timing analyzer and the Intermodule Bus will
    ensure the glitch is captured by the oscilloscope.
    2
    Touch the s/Div field to change its color to light blue.
    3
    Rotate the knob to change the s/Div to 20 ns.
    You will now see the glitch on both the oscilloscope waveform, and
    TCOUNT7 in the timing analyzer section of the display.
    If the oscilloscope module was not calibrated in its present slot, then you
    might not see the glitch unless you change the Delay setting.

    Page 100 of the manual I linked to gives you the screen shot.
    This is a training board By HP to teach people how to find glitches etc and use there equipment! How is it a phoney glitch? What it shows is the method used when you want to find a glitch, or are you sgesting HP dont know what they are talking about? Why do they sugest using the timing card of the LA to trigger the scope? Why do they say the Scope wont aways capturethe glitch properly souse the LA timing instead to capture it?
    Maybe you do know more than HP, but having actualy used the top of range LA in that series of LA for over 2 years I know it works and works well! I also have a HP scope 1GHz 6 channel, and I would use the LA over that any day for this sort of thing
     
Loading...
Thread Status:
Not open for further replies.