AtoD converters - how to troubleshoot?!!

Discussion in 'Embedded Systems and Microcontrollers' started by peskywinnets, Feb 25, 2009.

  1. peskywinnets

    Thread Starter Active Member

    Jan 19, 2009
    47
    0
    I have a suspect A to D chip, that samples about 5 different analogue voltage levels & converts them to digital (under the control of a Motorola DSP)

    I'm an analogue man through & through...I've never had to troubleshoot AtoD (or any other digital) chips...I need a bit of help where to start!! (links, tips, books, general advice etc)

    Tools wise, I have a 2 channel USB oscilloscope and a multimeter..but that's it!

    Here's the actual chip...

    http://focus.ti.com/lit/ds/symlink/tlv0838.pdf

    (even though the datasheet shows that the IC has 8 analogue channels in, the circuit doesn't use them all).

    And the circuit I'm working on is attached to this post

    I've checked that VCC is there...it is.

    I've checked for clocks - all present.

    I've checked the DC levels on the on analogue inputs change when their associated pots are turned - they do.

    I suspect this chip as being perhaps faulty, because no changes are heard from the mainboard's audio ouput when I move these associated pots (eg tone, volume etc) - this particular chip got very hot when a cable loom was connected to the mainboard erroneously....but it'd be nice to prove it.

    Otherwise do folks just order a replacement & slap it in on the off chance it was the IC? (it's SMT, so this isn't a great option for me!)
     
    • atod.jpg
      atod.jpg
      File size:
      126.2 KB
      Views:
      33
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Read page three of the datasheet carefully. I think there is a problem in the program.
     
  3. peskywinnets

    Thread Starter Active Member

    Jan 19, 2009
    47
    0
    Thanks for your input...I did read page 3 - pretty much gobbledegook! (at this stage)

    How can you establish that 'there's a problem with the program' from the fairly minimal info I've provided?!

    Is there anyway I can prove this - say in a standalone mode etc (eg a breadboard construction?)
     
  4. AlexR

    Well-Known Member

    Jan 16, 2008
    735
    54
    If the circuit has worked previously or if its a commercial unit then its most unlikely to be a software problem.

    You seem to have checked quite a lot of circuit but there are still a couple of things to check before you start ripping out the old chip.

    Make sure the CS pin is low.

    Check for 3.3Volt on the REF pin.

    Stick the scope on the DI pin and see that you are getting a serial data stream on it. This is the command input to the DtoA which tells it to start the AtoD process and what channel to digitize. If there is no data here the the fault is elsewhere in the circuit.

    Finally stick the scope on DO pin and look for a serial stream of digitized data. If everything else looks correct and you have no data on the DO pin then its time to break out the crowbar and change the chip.
     
  5. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    I don't know if it is a software problem, I am just guessing because it is easy to make a wrong instruction in the code. If you read the datasheet carefully of the ADC and apply the proper signals then it will work if it is not burned.
     
  6. peskywinnets

    Thread Starter Active Member

    Jan 19, 2009
    47
    0
    This is a previously good working commercial system (ie previously the code worked on this board - it's not under development.)
     
  7. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    If it was working then the ADC is probably burned.

    Apply a voltage on each of its analog inputs and check the digital output value if it is correct (if there is one).

    Also, check if the control signals are good and if the uP clocks the chip.
     
  8. peskywinnets

    Thread Starter Active Member

    Jan 19, 2009
    47
    0
    That's my original point - how do I do that?! (I already know there are voltages at the analogue inputs that change when I move the associated pots)
     
  9. AlexR

    Well-Known Member

    Jan 16, 2008
    735
    54
    The simple answer is you can't. At least you can't without some fairly sophisticated test gear such as a SPI protocol analyzer.
    You can however check that there is data (any data at all) being sent to the control pin of your AtoD by looking for transitions on the DI pin. Without an input on the DI the AtoD will not start a conversion.
    You can then look for data coming out of the AtoD by looking for transitions on the DO pin. Its not a 100% foolproof method of fault-finding, you could have data coming out of the AtoD that is garbage but its the best that you can do with your limited test gear and 9 times out of ten it will correctly isolate the fault.
     
  10. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Isolate the uP from the ADC and apply the control signals yourself, even the clock signal, to see if the ADC works.
     
Loading...