• State
  • Locked Locked
  • Replies 8 replies
  • Subscribers 46 subscribers
  • Views 363 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Original question:

[FAQ] Getting Started on TI E2E™︎ design support forums

TSC2004: TSC2004 Touch controller problems

David Taylor
Prodigy 10 points
Part Number: TSC2004

The TSC2004 touch controller is connected to a four wire resistive touch screen and to a uProcessor via I2C.  The product has just enetered production and some units stop behaving correctly, the touch position reported is not correct.  When the touch signals are checked with a scope probe some noise can be seen on the signals and usually the action of probing clears the issue temporarily.  The filters in the touch controller are on.  The problem currently seems to affect about 10% of the units.  From the testing we have done it seems to be a hardware issue but so far we have not found what causes it.  Once in the 'bad' mode a quick power cycle does not usually resolve the issue.  If the unit is powered down for several hours it behaves properly on start-up for some random time period after which the issue starts again.  The touch controller an uProcessor are on a SOM with the SOM mounted on a carrier board which has the main power supplies and interface circuitry, the touch screen signals are on a flex cable with the display video signals and are conencted directly to a connector on the SOM, they do not go to the carrier board.  Reseating the SOM also clears the issue and seems to resolve it for a long time / permanently. Therre are only a couple of caps and pull-up resistors connected to the touch controller and the touch signals land very close to the controller.  Any thoughs on what could casue this behaviour?

  • 0
    TI__Mastermind 23030 points

    Sometimes noise pickup on the X Y  pins can create issues. Have you tried 47p Capacitors on all X,Y pins close to the ChIP? 

  • 0 in reply to Sanjay Dixit
    Prodigy 10 points

    The touch works correctly for around two hours or so and then the reported values are incorrect.  A power cycle does not resolve the issue.  If powered down for a few hours then when powered up again the touch works correctly for another while.  The display behind the touch is a static display during the test so there isn't any noise from that.  Touching the X lines with a scope probe resolves the issue while the probe is in contact with the signal, as soon as it's removed the position is incorrect again.  Add 10nF caps to the signal but this made no difference, same effect.  If the GND side of the caps is Earthed with the earth lead of the scope probe the position reported is correct.  Again, when removed the position is incorrect.  The 47pF caps gave a similar result.

  • 0 in reply to David Taylor
    Prodigy 10 points

    Attached is a picture of the XL signal when it is not working properly.  The noise is not present when the system is working properly.  The noise appears randomly with no change to the system and once present stays until a long power down (hours) occurs.  Tocuhing the lines with a scope probe also fixes the problem but only for a few seconds/minutes.

  • 0 in reply to David Taylor
    TI__Mastermind 23030 points

    Does your equipment have a 3 pin Power cord? or does any computer that is connected to the equiment have a 3 pin socket.

    In general oscilloscope ground pins are tied to the earth connection of the Oscilloscope(The 3rd Pin of the Plug).

  • 0 in reply to Sanjay Dixit
    Prodigy 10 points

    Sanjay, no, the product is suplied from 24Vdc supplies from the mains.  In general there is a an Earth terminal but it is not usually connected.  Some power supplies have the 0V connected to Earth but usually the 0V is floating.  During these tests the power supply is a bench power supply with the outputs floating.

  • 0 in reply to David Taylor
    TI__Mastermind 23030 points

    You could try earthing the ground pin to see if the problem happens

  • 0 in reply to Sanjay Dixit
    Prodigy 10 points

    Hi Sanjay, connecting 0V to earth made no difference.  However, some progress has been made.  The hardware reset is tied high so is not used as it's tied to the same rail as the supply to the controller.  SBAA193 the POR document indicates that for a realiable POR the supply to the controller should rise at a minimum of 12Kv/Sec.  The supply to the controller is rising at about 5KV/sec which according to the document could put the controller into an internal TI test mode.  The only way to get out of this mode is to do a soft reset.  Currently the firmware uses the standard TI driver which does NOT use the soft reset as it's a generic driver for the full TSC2xxx range some of which don't have a soft reset.  We are modifying the driver now to use the soft reset.  This may explain why some units don't detect the position correctly after a power cycle but would not explain why units powered up for hours and working correctly from power up would suddenly stop working correctly and start giving incorrect position information.  Could the unpredicatble nature of the failure be related to the ic not being powered up correctly and not being soft reset?

  • 0 in reply to David Taylor
    TI__Mastermind 23030 points

    The only way this can happen several hours later is a Glitch on the power line.

    I would propose a Strong Decoupling Capacitor on power line ana a check to see if fault happens. A Hardware RESET can be tried at the moment to see what happens