HMI LCD Not displaying properly

[CAPFX]

I've got a HMI panel (EZP-S6M-R from EZAutomation) and have been trying to repair it. I bought it cheaply from ebay as a broken unit and also purchased a main board as well that was said to be working. With the original board, the LCD was blank and probing the 4 data lines showed there was no data being sent. I've since scraped this board. With the purchased board, I am able to get some data on screen albeit not the data I am wanting. I've attached photos of the screen in operation and some rendered images that I created by probing the data connection from the main board to the LCD (captured with the LCD disconnected). As can be seen by the rendered images, the fault is present on the data connection and not a fault of the LCD itself (plus I've driven the LCD directly with another project and it works fine). I have gotten the display show up a little bit of the text of the main setup screen but that happened once and I haven't been able to do it again.

The things I am able to test on the HMI do work (I cannot test the two RS232 interfaces as I don't yet have the software for that). I can put the HMI in it's configuration mode by pressing two corners of the screen and I get a beep. I can then press the screen where the buttons for the different options should be appearing and I can go into each setting menu and change things. It seems the processor in it is working fine. So I think the issue is somewhere else. When I tap the invisible screen buttons, the pattern on the screen changes but is still a bunch of garbage. You can see this in the two rendered images. The pattern isn't static either, it's sorta wavy or scanning or scrolling slowly down the screen. I can get a video if necessary.

The display is driven using a CHIPS F65545 display driver paired with an Alliance 265K x 16 DRAM for the display memory. There is a XILINX Spartan FPGA which goes to the CHIPS driver and the main processor and probably to the touch panel circuitry and the two RS232 ports (though other driver ICs) but it's a bit hard to tell with the density of the traces. The main processor is a Motorolo Coldfire MCF5206 and has two Cypress 64K x 16 SRAM chips as well as a SST 512K x 16 Flash chip.

Four images of the display are shown here: http://imgur.com/a/0JI4e

EDIT: Here is a video of the scrolling I'm talking about (sorry about the audio):

The FPGA seems fine as I believe the touch interface is controlled thru that and the processor seems fine as I am able to enter the setup screen and go thru the menus although it's all invisible to me. The CHIPS driver seems to be doing something as there is data on the bus. The data has a pattern to it and changes as I enter different menus so it doesn't appear to be noise. Is this a common issue that someone can just go "Yea, that is caused by X"?

I've tested the voltages at the various test points and have visually inspected the board. I'm not easily able to probe most of the connections as the pitch is so fine.

[Rasz]

model of the lcd panel?
have you went over every single connection between panel and controller and checked for continuity?
if its scrolling it might be missing clock, the fact you got it to almost display correctly would suggest intermittent connection

[CAPFX]

model of the lcd panel?
have you went over every single connection between panel and controller and checked for continuity?
if its scrolling it might be missing clock, the fact you got it to almost display correctly would suggest intermittent connection

The display is an EDT EW32F10NCW from Emerging Display Technologies.

The panel has a flat flex connection as well as through hole connection. They are both 14 pin interface and have the same pinout (so you can use either). This product uses the FFC. I have checked continuity between the FFC on the LCD and the FFC of the main board and it's good for each of the pins. I also checked continuity between the FFC and the CHIPS pin and it was all good. I've also tried to manually wire in the through hole connect to the pins holding the FFC to the main board and I have the same issue. There is a second unpopulated connector on the main board that is not used on this model but is for a larger screen. It uses the same pins on the CHIPS driver chip and when I wire the display up to that connector, the issue is still there. It doesn't appear to be a connector issue.

I have probed the lines with a logic analyser (which is how I generated the two rendered images) and all the signals were there and they looked fine to me. I can share the Saleae Logic file if necessary. I've added a link to the original post to a video showing the scrolling I'm talking about.

When I got the partial display, I was able to shake things around and it held the display. It only glitched out again when I pressed the touchscreen to move to another menu. I have not ever been able to get this to happen again. With the glitched noise stuff, it's the same where I can shake things and the pattern stays the same (just scrolling) but the patter will change when I press a button.

[Rasz]

disregard my previous message, I was expecting ordinary panel + driver, this is a 4-bit wide serial with sanyo shift registers, now I get last 2 pictures
its not really scrolling, more like cycling? most of the garbage is static and looks like bad panel, moving parts remind me of bad video ram (or resolution/graphic mode)
might be bad connection to one of the ICs or decoupling on the driver board
discharge yourself into solid earth (plumbing piping?), turn it on, go over every chip on driver board with a finger touching pins

btw I didnt know F65545 supported this type of interface, I though it only did rgb and parallel