On Sat, 14 Apr 2001, Wojciech Zabolotny wrote:

> On Fri, Apr 13, 2001 at 04:15:19PM -0500, Scott Dattalo wrote:
> > > Eg. the YSCL timing -
> > > there is only the "typ" value (eq. 220 usec.) provided, and it is unclear
> > > if it is safe to use other (longer) times.
> >
> > YSCL? You sure you don't mean XSCL?
>
> Yes, it is YSCL. Both
>  http://www.controlanything.com/manuals/lcdpdf/EG2401.PDF
>  http://www.controlanything.com/manuals/lcdpdf/EG4801.PDF
> contain that signal (table 1-4, pin no 7 - "Row scan shift clock input")
> And the problem is that I'm not sure if "undeclocking" of that pin may lead
> to LCD destruction, due to "DC driving of LCD".

Well, I reckon I don't know the answer. The graphics LCD's I've seen only have
XSCL for clockin in the column data. I suppose YD on the 1335 is what's
responsible for controlling the row scan rate.

>
> > You may want to look at the SED1335 data
> > sheet (or SED1330's). I did see a reference to the EG2801 in here:
> >
> > http://www.eio.com/cgi-bin/byteserver.pl/sed1330.pdf
> >
>
> Yes, I've found that doc too, but I'd like to avoid using an additional
> chip. I think that driving of LCD panel can be done by a 16F877 (PIC of
> course ;-) ) in the final device.

Agreed!


> YSCL: period=220usec => freq. = ca. 4.5kHz
> In each period of YSCL there should be 512/4=128 XSCL pulses, so it makes
> XSCL freq= 128/220usec. = ca. 582 kHz
> So when using the 18432MHz xtal for 16F877 (good frequency for RS232
> compatibility) I should get 8 instructions for each pixel.
> With a small hardware support (8bit port for output and 2x4->4 MUX between
> that port and D0-D3 lines of LCD module addressed with T flip-flop driven
> by XSCL) I can achieve 16 instructions for byte. It should be enough
> to access the external RAM and generate the necessary pulses.
> However the modification of RAM will disturb the LCD timings. I don't know
> if it is safe and acceptable...

Too bad the 877 does not have enough internal RAM storage...

I suppose you've done this already, but the way I'd approach this problem is by
trying to emulate the 1335. From what I've seen (but I haven't seen your
display) the LCD vendors use Row and Column drivers that can be driven by the
1335. So if you were to hook up your Epson display to a 1335, how would you
program the 1335 to drive it? From that you can determine the timing of the
signals. However, insufficient LCD data sheet info would preclude this approach
(and is probably why you're posting this question to the piclist!).


Scott

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.