Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 03 Jan 1995 23:11:25 -0700
From:      Steve Passe <fbsd@clem.systemsix.com>
To:        dusio <dusio@ws062.bytex.network.com>
Cc:        questions@freebsd.org
Subject:   Re: Can't get 1024x768 in Xfree 
Message-ID:  <199501040611.XAA08269@clem.systemsix.com>

next in thread | raw e-mail | index | archive | help
>> Every time you make a change, be sure that your new ModeLine
>> still fits within your monitor's horizontal and vertical sync
>> limits.  Significantly out-of-limits sync rates can damage
>> some monitors!

xf86-3.1 uses HorizSync and  VertRefresh to protect the monitor.
If the modeline math works out to be outside of the ranges declared
in these two fields the server will drop them from usage during
initialization.  Notice of this goes to stdout (stderr?) on the
tty that you invoke startx from.  It happens pretty early in the
sequence and usually will be scrolled off the screen by the time
you leave X.  To capture it do:

% startx 1>/tmp/Xlog 2>&1

now from within an xterm you can:

% more /tmp/Xlog

to see what happened on the way up.  Or try:

% tail -f /tmp/Xlog

to keep a running view of what happens

often times when in the fine-tune stage you use numbers that
push you just beyond the limit.  eg, I was tuning numbers
for a monitor with an Hsync range of 30-65 and came up with
numbers that calculated to an actual value of 65.4 by the X
server, which then promptly threw that line out.  It took
awhile to notice, once I did I bumped the high end of both
ranges up a little to prevent this while fiddling with the
numbers.  Once I arrived at something I liked I reset the ranges.
(use this method at your own risk!).


>> I didn't include the Bandwidth specification because it doesn't
>>  seem to be used by drivers.

as above, I believe the server would toss the line if the stated 
clock rate was above the Bandwidth spec, but I have never actually
tried this to prove it.  In anycase, the clocks listed in the 
sample XF86Config file are all below 100 and thus this is not an
issue here.


>> You may not even want to use 1280x1024.  Monitors are generally
>> built with a 75% aspect ratio -- the viewing area is 3/4 as high
>> as it is wide.

I agree, keep things 'square'.  1280x960 also gives room for a
slightly higher refresh rate.  unfortunately, none of the listed
clocks would give acceptable refresh rates at 1280. 1280x1024
@60Hz requires approx a 110Mhz clock.  the weak link here is the
ramdac speed, check it carefully when buying a new board.  4 MB
of memory might get you true color, but unless the ramdac is at
least in the 110-120 Mhz range you will NOT be able to use that
memory for the 1280 dot modes.  (unless you really like the 'strobe
effects')  to get above 70Hz refresh you need around 135Mhz ramdac
(and clock).  note again that the driver will toss a line if it
attempts use of a clock higher than the ramdac's rated speed,
even if the clock is truly available on the board.  I should state
here that all of the above claims I make about the driver doing
the 'good thing' with these numbers is based on my experience with
the XF86_S3 driver, I do not know for a fact that this applies to
all the XF86_xxx drivers,  have not looked at the source to see
at what level this is accomplished....


>>    Clocks 19.9 22.3 24.8 25.0 28.1 31.3
>>    Clocks 32.2 35.7 35.9 37.0 39.8
>>    Clocks 39.7 44.6 49.6 49.9 56.2 62.6
>?    Clocks 64.4 71.4 71.8 73.9 79.6

check out the file modeDB.txt, "MONITOR SECTION".  at the head
there is a list of VESA timings which will work as starting points
for most any newer monitor.  find the one with the highest clock
rate provided by your board @ the desired resolution and start there.
with the listed clocks I would pick "VESA 1024x768@70Hz"

the ModeLine is:

 "1024x768"  75  1024 1048 1184 1328    768 771 777 806

(but change clock from 75 to 73.9)
 btw, these clocks look very fishy to me also!

						Steve Passe
						smp@csn.org




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199501040611.XAA08269>