Msg: 6437 *Conference*

04-15-96 08:53:06

From: RON WIESEN

To : PAUL WHITING

Subj: REPLY TO MSG #6428 (STRANGE PIXELS ON LCD AT 102 POWER UP)

I assumed you shut down from the Main menu Paul.  Even so, prior applications
leave "fingerprints" in regions of RAM.  The op system uses certain regions of
RAM for general purposes rather than commit them to specific purposes.  For
example, what is called the "previous LCD buffer" is a 320-byte general purpose
region in RAM.  It's name reflects its most extensive usage: to hold the 8
lines (8 x 40 =320) of the LCD which relate to the "Prev" (F1) function of
TELCOM under Terminal mode.  But the op system uses smaller parts of this
region under other circumstances as well.  Given its large size, this 320-byte
region is very popular and its a favorite place to ORIGinate machine language
code so you end up with a .CO file that you can invoke without having any HIMEM
requirement.

Another 320-byte region of RAM is for the current (last displayed) 8-line LCD
characters.  Although this is nearly a "specific purpose" region, even it can
be used in a general way.  The specific purpose of this region is for "line
scroll" by the op system.  The LCD circuit needs no character buffer, in fact
it's the 6 by 8 pixel raster that the Op system sends to the LCD circuit to
display a "character".  But the actual character is also saved in this 320-byte
region so that "scrolling" can be subsequently done by the Op system.  By
reading the 2nd line of 40 characters from this region, the Op system can paint
40 pixel rasters (characters) for the 1st line and store the 40 characters in
the 1st line part of this region to effect a scroll of the 2nd line to the 1st
line.  Repeating this for another 6 lines, a "screen scroll" is produced and
the character codes (bytes) of 7 lines (280 bytes) are "shuffled" by 1 line
position.  But the truth is that once pixels paint character symbols on the
LCD, the 320-byte content which "may" be needed for subsequent screen scroll
"might" not be used.  So a program that "limits" or otherwise carefully
controls screen scroll, can use this 320-byte region for a purpose unrelated to
screen activity.  I've made and use such programs.

As far as a "warning signal" about impending LCD failure goes, it might be more
like a simple symptom that something minor "has" failed.  I don't know for sure
and perhaps someone else does or can find it in an LCD circuit description of
the M100 Tech manual, but I bet big money that a "power-on" signal is applied
to the LCD circuit and where activated it clears all pixels.  If this is true
and your LCD no longer responds to this signal or receives the signal (broken
wire or source circuit) then application of power produces a normal "random
pixel print" that persists long enough to be seen.  Again I bet big money that
the LCD circuit on power application doesn't, of and by itself, reset all
pixels.  If I'm right, you have some circuit (LCD or power) or conductor (wire
or connector) fault that can't get worse and is of minor consequence.  If this
is the case, the transient "random pixel print" looks identical or nearly so
each time you see it.  Of course this same fault on another LCD circuit will
present an entirely different "random pixel print" than does your LCD circuit
but it will be repeated identically each time.  Think of it as a harware
"fingerprint" which doesn't vary for a given LCD circuit but which is "random"
with respect to exactly what pixel pattern is inbred from one LCD to the next
LCD.  Your's looks one way, and if mine persisted long enough for me to see it
would look different than yours.