--- In laser_design@yahoogroups.com, "Doug" <dx9s@...> wrote:
>
>
> Well I put 4x 1k ohm resistors inline with the datalines (as they
> change direction)... prolly can put 1k on the enable, RS (register
> select) and R/W (read/write) lines too..
>
> It works so my theory about 1k on data seems to be holding true.
> (insignificant in comparison to other relates things, driver current
> and target [load] resistance [being really high]).
>
> I'm going keep on tinkering...
>

FWIW: 1k on enable means the rise/fall times (most likely the fall
between two rises of a nibble in 4-bit mode) take too long. Suspect
it's a "filter" (RC) where the capacitance of whatever little is there
is enough to effect the rise/fall time.

Lower the Enable line to 100 ohms and it works fine (have some code
that sends continues stream of information to the LCD module for
testing several things). So can update the code to make the time
between the two rises in a nibble longer (timing correction in
software -- probably not a bad thing to code for worse case)...

Mind you all the rest of the lines are still 1k (including RS and R/W
which do NOT switch logic state as short/fast as Enable does in 4-bit
mode between nibbles).

In any case pushing the hardware to find limits! Learning & having
fun.... If I was issued a belt (like in some martial arts)... I would
no longer be a white belt (far from black as I don't know all the cool
things in the HD44780 protocol that I sometimes see)... More like blue
belt!

Wanna learn how to flip the text upside down or mirror or invert (on
pixel to off and reverse).. I know most/all of those things should be
achievable based on what I've seen with one LCD module with noise on
the lines.

--Doug