Search

Technical Discussion Group Forum

This forum is provided for user discussion. While Beacon EmbeddedWorks support staff and engineers participate, Beacon EmbeddedWorks does not guarantee the accuracy of all information within in the Technical Discussion Group (TDG).

The "Articles" forums provide brief Articles written by Beacon EmbeddedWorks engineers that address the most frequently asked technical questions.

To receive email notifications when updates are posted for a Beacon EmbeddedWorks product download, please subscribe to the TDG Forum of interest.

TDG Forum

PrevPrev Go to previous topic
NextNext Go to next topic
Last Post 16 Aug 2007 06:47 AM by  sygi@canbus.pl
Touchscreen for a 6.4" DisplayKit and LH79520
 11 Replies
Sort:
You are not authorized to post a reply.
Author Messages

peter_vc



New Member


Posts:
New Member


--
02 Mar 2006 03:21 PM
    I am trying to get the 6.4" LCD display kits to work with the LH79520 DevKit and I'm having a issues with the touch component. The LCD displays fine at any resolution but the touchscreen fails the sanity check because it seems to Trapizoidal in shape. The right side is almost half the size of the left. Is this a issue with the hardware itself (ie. a bad panel or cabling) or is this a possible driver or other software issue we are missing?

    kurtl@logicpd.com



    New Member


    Posts:
    New Member


    --
    02 Mar 2006 04:59 PM
    The display kit's touch panel accuracy is functionally tested in the unit before it is shipped so the hardware should be working ok pending no damage during shipping.

    If you want to manually check the linearity of the touch panel, apply a small bias voltage across the X+ and X- lines, like 3 volts or so, then measure the voltage seen on Y+ or Y- when you touch the screen using a volt meter. As you move the pointer across the screen you can get a feel for how linear the touch screen response is. Swap the bias to the Y pair and monitor X to check the orthogonal direction.

    -Kurt

    peter_vc



    New Member


    Posts:
    New Member


    --
    03 Mar 2006 05:25 PM
    We happened to have a second touch panel direct from 3M thats the same as the one shipped with the LCD kit. We hooked it up and got the same results. Watching the co-ordinates from inside the driver (ads784x.cpp I believe) prove that the hardware is telling us the panel is not square. Am I correct in that ads784x.cpp is the first point of contact between the panel and Linux? If so it doesn't appear the controller needs any sort of initialization.

    I really dont know where to start looking for the problem. As far as I can tell the controller doesn't need to be told what resolution or dot pitch (for lack of better terms) the panel should scale to.

    The BDK patch mentions in its readme that the 6.4" touch panel isn't accurate and it could be fixed in the driver but I fail to understand how. Do they mean the ads784x.cpp driver or the mouse driver (qwsmouse_qws.cpp?) or elsewhere?

    I guess we'll start by testing the panel to check for linearity and then feed the controller some known voltages to see if its translating properly.

    This is my first adventure into the depths of the kernel and would appreciate any tips you can provide.

    peter_vc



    New Member


    Posts:
    New Member


    --
    03 Mar 2006 06:01 PM
    I've just discovered the touch-cal program inside of LoLo and it works much better but not perfect, so we're definately dealing with a driver issue. Is the source code for LoLo available?

    kurtl@logicpd.com



    New Member


    Posts:
    New Member


    --
    06 Mar 2006 08:24 AM
    peter_vc,

    The source for Lolo is not freely available, however, you can contact sales@logicpd.com for a quote on what it would cost to obtain the source code.

    -Kurt

    peter_vc



    New Member


    Posts:
    New Member


    --
    11 Jul 2006 01:31 PM
    I'm back at it and using a logic analyser we have determined that lolo does not use interrupts. It polls the ads chip. Has anyone got a 6.4 display kit up with linux. I think we may have defective hardware.

    kurtl@logicpd.com



    New Member


    Posts:
    New Member


    --
    11 Jul 2006 01:53 PM
    Hey Peter,
    I need to correct your comments. Lolo DOES use interrupts for the touch interface. The intial touch of the screen creates an interrupt back to the cpu, the cpu services the interrupt, reads multiple ADC samples, and averages them to get the result.

    -Kurt

    peter_vc



    New Member


    Posts:
    New Member


    --
    11 Jul 2006 02:11 PM
    Thanks the tip might prove useful. The logic analyser determined that the power down bits of the ads chip data request were set which disables the interupt pin. Lolo must wait for an int and then switch modes. I'll try that.

    peter_vc



    New Member


    Posts:
    New Member


    --
    01 Feb 2007 01:10 PM
    Just an update to this continuing problem. We are still not able to use the touch screen from linux as the system will still not pass the sanity check. After extensive experimentation including building identical circuits with identical parts which work fine, we have determined that there is too much capacitance on the touchscreen lines themselves. We are searching for the source of this unwanted capacitance but so far have been unable to find it (it seems to exist on the circuit board). The result is a switch at the midpoint from capacitive to inductive behavior on the four lines giving very non-linear behavior. When scoped, the top of the wave form is convex, flat at the center and concave below center. There are four very small unpopulated spots for unknown parts connected to the four touchscreen lines on a small board that is part of the interface kit and acts as the glue between the dev board and the touchscreen/display panel (not included on any schematics). Our hardware engineer suspects that adding more capacitance to these four lines will swamp out the inductive effect and we will still see the rounding but it will be consistent both above and below the center of the waveform. Any thoughts/suggestions?

    sygi@canbus.pl



    New Member


    Posts:
    New Member


    --
    13 Aug 2007 07:16 AM
    I have the same problem with my LCD touchscreen (AMPIRE AM320240). Characteristic of my TS is much like triangle. Did You solve this problem ???.

    kurtl@logicpd.com



    New Member


    Posts:
    New Member


    --
    13 Aug 2007 07:19 AM
    Hello,
    If you would like to test with less capacitance on the touch lines, you can remove or replace C87-C92 with lower value capacitors. These caps are present to help filter any noise that is coupled to the touch screen inputs.

    -K

    sygi@canbus.pl



    New Member


    Posts:
    New Member


    --
    16 Aug 2007 06:47 AM
    I've decreased capacitance on the TS lines to 10pF, but without any results. . In the other side bigger capacitance cause much worse behavior (cursor is hanging).

    Thank for advaice

    Regards
    You are not authorized to post a reply.