Baslinux mailing list

[sindi keesan <keesan AT sdf.lonestar.org>]

I just got hold of some 'new' video cards - a 1995 and a 1997 Matrox (VESA 
2.0, framebuffer), and an even newer (VESA 3.0) SiS (32MB).  The SiS won't 
work with zgv in BL2, complains about something not being defined, which 
is another reason to upgrade svgalib to SW91.  The other three probably 
need me to change to or from chipset VESA, and are having their own 
problems with svgalib (crash on exit, can't reuse any VGA terminal).  But 
maybe a later svgalib would fix these problems too.  They work with Xvesa.
The Mach64 crashed something.

The cards go to 1024 on my 15" monitor (31.5-54KHz horiz sync, found on 
the web) and to 1280 on the 17" monitor (goes to 66KHz) but not to 1600 on 
that monitor (they output 72KHz-76KHz at that resolution).  My 17" monitor 
tells me the frequency, and that it is out of range.

When I try to use them with David's framebuffer kernel, even at 640 
resolution or 800 resolution, on the 17" monitor, I get a black screen.
Not even an 'out of range' message from the monitor.

I had my older Matrox working fine with a few monitors, in framebuffer 
mode, same for two Tsengs and some AGP cards.  Are some monitors just 
incapable of doing framebuffer?

Is there some way, using a linux tool rather than a monitor with error 
messages, to determine which frequencies a video card puts out, so I can 
match up the cards with the monitors without having to drag around either 
a computer or a 17" monitor?  I have now determined that the SiS does 
89KHz at 1920 res, and the Matrox cards do 72 or 74KHz at 1600, and none 
of these work on the 15" monitor which goes to 54KHz but they do work on 
the 17" which does 66MHz so that 1280 is between 54.1 and 65.9, but is 
there a better way to do this?

I have at least figured out that 15" monitors generally go up to either 
49.9/50.5 KHz hsync (1993-1997) or up to 54-56KHz (1997-1999) and will do 
1280 with S3 or Tseng but not Matrox.  S3 seems to work with older 
monitors best, also with all of BL software.  Trident will go to 1024 on 
even older monitors where S3 only does 800.

Most 17" monitors (1996-2001) that we have do 1600 res up to 66, 69, or 
70MHz, still not enough for the Matrox, but okay for S3 including the AGP 
Trio3D.  ATI and Riva will work at 1400 res.  One 17" NEC says it only 
does up to 1280 but it goes to 70KHz so I will try it with S3.  It is 
unlikely to do 1600 with a new card.

A 15" HP '1024' from 1997 goes to 49.4 KHz and did do 1280 with S3, but 
not with newer cards.

This implies that newer monitors are labelled as to what resolution they 
can achieve with newer cards, which put out too high hsync for them at the 
higher resolutions, but they should be able to go higher than rated.

Is there some software method of determining the hsync of video cards?
I am getting tired of moving hardware around for testing.

Digiview made a monitor in 1996 that did 24-82KHz hsync.  Still looks 
great and I should try it with the newer cards at 1600.

CTX made something in 1993 that goes to 1600 resolution.  We had three but 
they are all rather fuzzy by now.

We have two 1997 HP monitors, one goes to 49.9 (nice a sharp but skewed) 
and the other to 50.4 (this one got very dim).  HP does not make monitors, 
they sell them.  Compaq sold a still-clear 50.4 in 1993 and a really nice 
54KHz in 1999.  A 1992 Viewsonic still looks great (up to 1024) as does a 
1996 Viewsonic (69KHz, same as some 1999-2001 monitors).  IBM monitors are 
clear but not cutting edge in frequency (1998 56.4, 1999 69).

So the manufacturer (or seller) is more important than the year, for both 
quality and frequencies.

Xvesa was able to handle all these cards, at least up to 1600 on the 17".

Some other newer cards also messed up svgalib.

I will report back about the svgalib upgrade and USB kernel, and also SiS 
card.  And the smartmedia USB card reader - we have made a backup copy of 
the 55MB of photos at 19200 download speed (overnight) so can now 
experiment with both of the new kernels (2.2.26 and 2.4.31) on two 
computers (and on several more once the video/monitor experiment is done).

keesan AT sdf.lonestar.org
SDF Public Access UNIX System - http://sdf.lonestar.org