LCM flicker

[Werner Almesberger]

Wolfgang Spraul wrote:
> So if GTA01 does not show the problem, it's either glamo or wiring?

It could still be a lot of things, e.g., timing incompatibility,
electrical incompatibility, configuration error, Glamo misbehaving,
etc.

> Werner (or raster), if you can e-mail some specific things you would  
> want Allen to look at that would help...

I think I'd start by looking for two things:

- the exact frequency of the flicker (1), and

- whether the RGB data is correct (2). With an entirely monochrome
  screen, all the bits should have a constant value. (Except perhaps
  during vertical and horizontal blanking.)

I think I'd start with looking at signal integrity of clock and RGB
data

Comparing the video signals (page 9 of TD028TTEC1*-V13-*.pdf) between
GTA01 and GTA02 may give a clue of what's happening.

(1) This could be done with an optical sensor, but I'm not sure if
    we have anything suitable. It may also be possible to get this
    information by monitoring the current drawn by the LCM. Maybe
    also a stroboscope effect could be used to find the frequency.

(2) Things I imagine could go wrong are:
    - analog signal fades
    - incorrect digital data coming from the Glamo
    - phase drift between PCLK and picture data

It also can't hurt to give all the voltage levels (backlight, LCM
supply, LCM boost converter) a second look, just in case.

But this is really fishing. I don't have a concrete idea of a failure
pattern that could cause this.

- Werner