This display problem is stumping all the experts...

Okay, something weird is going on right now. I don't know what triggered it
but I've been doing a lot of troubleshooting lately, before giving up and
reformatting my HD (because of another unrelated issue).

The monitor (a "Dell 2407WFP-HC" LCD screen) gets no signal on boot-up until
Windows is loaded. Only then does it turn on. The screen is connected via
DVI (looks noticeably better than VGA).

MB is Asus A8N-SLI Deluxe, running latest Asus BIOS for this MB.

Video card is Asus EN6600GT.

---

Troubleshooting :

When I plug a second simultaneous monitor in the video card's VGA slot (a
BenQ CRT), BOTH the CRT and the LCD display the full boot-up routine (DOS
visuals). And of course, the Windows portion.

Leave the VGA slot unplugged (with only the DVI plugged-in), and the whole
DOS boot-up routine is bypassed by the monitor again. It gets no signal
until Windows is loaded.

Loading BIOS defaults solves nothing.

What gives? Why can't the DVI slot emit anything until Windows is loaded,
unless something is plugged into the VGA slot? I need to plug in a 2nd
monitor each time I want to boot in Safe Mode, or else I can't see the
menus.

Mike J.S. wrote:
> Okay, something weird is going on right now. I don't know what triggered it
> but I've been doing a lot of troubleshooting lately, before giving up and
> reformatting my HD (because of another unrelated issue).
>
> The monitor (a "Dell 2407WFP-HC" LCD screen) gets no signal on boot-up until
> Windows is loaded. Only then does it turn on. The screen is connected via
> DVI (looks noticeably better than VGA).
>
> MB is Asus A8N-SLI Deluxe, running latest Asus BIOS for this MB.
>
> Video card is Asus EN6600GT.
>
> ---
>
> Troubleshooting :
>
> When I plug a second simultaneous monitor in the video card's VGA slot (a
> BenQ CRT), BOTH the CRT and the LCD display the full boot-up routine (DOS
> visuals). And of course, the Windows portion.
>
> Leave the VGA slot unplugged (with only the DVI plugged-in), and the whole
> DOS boot-up routine is bypassed by the monitor again. It gets no signal
> until Windows is loaded.
>
> Loading BIOS defaults solves nothing.
>
> What gives? Why can't the DVI slot emit anything until Windows is loaded,
> unless something is plugged into the VGA slot? I need to plug in a 2nd
> monitor each time I want to boot in Safe Mode, or else I can't see the
> menus.
>

The video card has a BIOS add-in module. When the computer POSTs, all of
the add-in modules are loaded. In the motherboard BIOS flash chip, there
might be three add-in modules, for things like PXE boot (network booting),
RAID add-in ROM, and so on. The chip on the video card, also contains
code, and that code is also an add-in ROM. 64KB of memory is allocated for
its use, in low memory.

One of the things the video card BIOS code should do, is initialize the video
card and configure it to support a certain small set of VESA modes. The
BIOS wants to drive the screen at 640x480 or whatever, and there needs to be
a way to load the registers in the GPU, to support that mode.

Now, part of that sequence *could* include the video BIOS code, reading the
EDID information from the monitor. There is a serial EEPROM (electrically
eraseable programmable read only memory) inside the monitor. It is connected
to a serial clock and data interface on the cable. The VGA cable has this
serial DDC interface and so should DVI. As part of the BIOS code loading, the
BIOS code may expect to see sane EDID info, before proceeding. Part of the
purpose of EDID, is so that the video card won't use illegal resolutions or
refresh rates.

Now, there have been cases, where certain monitors and certain drivers, have
resulted in the EDID chip getting corrupted. On the one hand, it would make
a lot of sense, if the EDID chip was read only. The difficulty with that, would
be if the manufacturing people wanted to put a custom serial number, electronically,
for each monitor made. To support that, the manufacturer may leave the chip
read/write. There have been tales of the EDID being corrupted. Once the EDID
is corrupted, connecting your monitor is a lot like connecting your projection
system to the computer - the computer hasn't a clue what is safe. The video
BIOS policy might vary from design to design, as to what to do.

http://www.rage3d.com/board/showthread.php?t=33779480

(More fun with monitors here. Save this one for later)
http://www.geocities.com/jgeneedid/

That is about the only explanation that makes any sense to me, because the
BIOS sequence is not a function of the OS used or any OS settings. Changing
something in the Display control panel, installing a Microsoft Service Pack
or the like, is not going to affect the Video BIOS, motherboard BIOS, or
hopefully, the contents of the EDID chip. They are part of a more primitive
subsystem, that makes decisions before the first sector of the boot is read
off the hard drive.

At the OS level, the OS has a couple choices of info to consult. The OS
can look at the EDID. If the EDID is missing, a second opportunity (for
max res and refresh) can come from the "monitor driver". The "monitor driver"
file is not available from all manufacturers, and an example would be people
who buy 42" LCD TV sets and expect their VGA option to work. TV makers aren't
likely to have a download page with the tiny monitor driver needed.

This utility, will attempt to read the EDID. The above rege3d thread also has some
suggestions.

http://www.entechtaiwan.com/files/moninfo.exe

This is what moninfo reports for my 17" LCD. I presume my EDID is working.

Standard timings supported
640 x 480 at 60Hz - IBM VGA
640 x 480 at 67Hz - Mac II
640 x 480 at 72Hz - VESA
640 x 480 at 75Hz - VESA
720 x 400 at 70Hz - IBM VGA
800 x 600 at 56Hz - VESA
800 x 600 at 60Hz - VESA
800 x 600 at 72Hz - VESA
800 x 600 at 75Hz - VESA
832 x 624 at 75Hz - Mac II
1024 x 768 at 60Hz - VESA
1024 x 768 at 70Hz - VESA
1024 x 768 at 75Hz - VESA
1152 x 864 at 75Hz - VESA
1152 x 870 at 75Hz - Mac II
1280 x 1024 at 60Hz - NEC
1280 x 1024 at 60Hz - VESA
1280 x 1024 at 75Hz - VESA

So, that is my theory, that something has affected the EDID. You'll be able to
verify that pretty quickly, using the above utility.

Finally, if you don't want to screw with the EDID, or contemplate this
problem any longer than necessary, there is another alternative. You mentioned
that connecting both the VGA and DVI, allowed the DVI to work. You can make
a "dummy" VGA connector, to convince the monitor detection feature of the
video card, that a monitor is present. I've done this on another computer,
where the computer would freak if you wake it from standby, and the monitor
happens to be disconnected. I built one of the following, without consulting
this web page, and it worked for me. I didn't need to do more, than terminate
the three color signals, with 75 ohm resistors to ground. The video card
thinks a monitor is connected, although it won't get any EDID from this
"dummy" :-)

http://www.soerennielsen.dk/mod/VGAdummy/index_en.php

HTH,
Paul

"Paul" <> wrote in message news:fc7if5$v31$...
>
> This utility, will attempt to read the EDID. The above rege3d thread also
> has some
> suggestions.
>
> http://www.entechtaiwan.com/files/moninfo.exe

Here's what it reports with BOTH the CRT (vga) and LCD (dvi) monitors
plugged-in :

Monitor
Windows description......... Plug and Play Monitor
Manufacturer description.... BENQ P992
Manufacturer................ BNQ
----------------------------
Plug and Play ID............ BNQ0106
Serial number............... 00137
EDID data source............ Registry (stored)
----------------------------
Manufacture date............ 2003, ISO week 50
EDID revision............... 1.3
Display type and signal..... Analog 0.700,0.000 (0.7V p-p)
Sync input support.......... Separate, Composite
Screen size................. 350 x 260 mm (~18")
Power management............ Standby, Suspend, Active off/sleep

Color characteristics
Display gamma............... 2.90
Red chromaticity............ Rx 0.639 - Ry 0.322
Green chromaticity.......... Gx 0.274 - Gy 0.597
Blue chromaticity........... Bx 0.143 - By 0.062
White point (default)....... Wx 0.282 - Wy 0.297

Timing characteristics
VESA GTF support............ Not supported
Horizontal scan range....... 30-98kHz
Vertical scan range......... 50-160Hz
Video bandwidth............. 170MHz
Extension blocks............ n/a
Timing recommendation #1.... 1280x1024 at 85Hz
Modeline................ "1280x1024" 157.500 1280 1344 1504 1728 1024
1025 1028 1072 +hsync +vsync

Standard timings supported
640 x 480 at 60Hz - IBM VGA
640 x 480 at 72Hz - VESA
640 x 480 at 75Hz - VESA
720 x 400 at 70Hz - IBM VGA
800 x 600 at 56Hz - VESA
800 x 600 at 60Hz - VESA
800 x 600 at 72Hz - VESA
800 x 600 at 75Hz - VESA
800 x 600 at 85Hz - VESA
1024 x 768 at 60Hz - VESA
1024 x 768 at 70Hz - VESA
1024 x 768 at 75Hz - VESA
1024 x 768 at 85Hz - VESA
1280 x 1024 at 75Hz - VESA
1280 x 1024 at 85Hz - BNQ
1600 x 1200 at 75Hz - VESA

Raw EDID base
00: 00 FF FF FF FF FF FF 00 09 D1 06 01 89 00 00 00
10: 32 0D 01 03 6C 23 1A BE EA A7 B4 A3 52 46 98 24
20: 0F 48 4C AF CF 00 A9 4F 45 59 61 59 01 01 01 01
30: 01 01 01 01 01 01 86 3D 00 C0 51 00 30 40 40 A0
40: 13 00 68 0E 11 00 00 1E 00 00 00 FC 00 42 45 4E
50: 51 20 50 39 39 32 0A 20 20 20 00 00 00 FF 00 30
60: 30 31 33 37 0A 20 20 20 20 20 20 20 00 00 00 FD
70: 00 32 A0 1E 62 11 00 0A 20 20 20 20 20 20 00 B1

Display adapter
Adapter description......... NVIDIA GeForce 6600 GT
Adapter device ID........... 0x014010DE
Display settings............ 1920x1200, 32bpp

----------

Oddly enough, the drop-down on top does not let me choose the 2nd monitor!
It's as if this thing refuses to acknowledge the Dell LCD even exists.

Cold-unplugging the BenQ and running the program again, I get... the BenQ
info again. And still no sign that the Dell LCD exists.

Note that I booted with both monitors plugged-in before running these tests.

What now? :-(

Mike J.S. wrote:
> "Paul" <> wrote in message news:fc7if5$v31$...
>> This utility, will attempt to read the EDID. The above rege3d thread also
>> has some
>> suggestions.
>>
>> http://www.entechtaiwan.com/files/moninfo.exe
>
> Here's what it reports with BOTH the CRT (vga) and LCD (dvi) monitors
> plugged-in :
>
> Monitor
> Windows description......... Plug and Play Monitor
> Manufacturer description.... BENQ P992
> Manufacturer................ BNQ
> ----------------------------
> Plug and Play ID............ BNQ0106
> Serial number............... 00137
> EDID data source............ Registry (stored)
> ----------------------------
> Manufacture date............ 2003, ISO week 50
> EDID revision............... 1.3
> Display type and signal..... Analog 0.700,0.000 (0.7V p-p)
> Sync input support.......... Separate, Composite
> Screen size................. 350 x 260 mm (~18")
> Power management............ Standby, Suspend, Active off/sleep
>
> Color characteristics
> Display gamma............... 2.90
> Red chromaticity............ Rx 0.639 - Ry 0.322
> Green chromaticity.......... Gx 0.274 - Gy 0.597
> Blue chromaticity........... Bx 0.143 - By 0.062
> White point (default)....... Wx 0.282 - Wy 0.297
>
> Timing characteristics
> VESA GTF support............ Not supported
> Horizontal scan range....... 30-98kHz
> Vertical scan range......... 50-160Hz
> Video bandwidth............. 170MHz
> Extension blocks............ n/a
> Timing recommendation #1.... 1280x1024 at 85Hz
> Modeline................ "1280x1024" 157.500 1280 1344 1504 1728 1024
> 1025 1028 1072 +hsync +vsync
>
> Standard timings supported
> 640 x 480 at 60Hz - IBM VGA
> 640 x 480 at 72Hz - VESA
> 640 x 480 at 75Hz - VESA
> 720 x 400 at 70Hz - IBM VGA
> 800 x 600 at 56Hz - VESA
> 800 x 600 at 60Hz - VESA
> 800 x 600 at 72Hz - VESA
> 800 x 600 at 75Hz - VESA
> 800 x 600 at 85Hz - VESA
> 1024 x 768 at 60Hz - VESA
> 1024 x 768 at 70Hz - VESA
> 1024 x 768 at 75Hz - VESA
> 1024 x 768 at 85Hz - VESA
> 1280 x 1024 at 75Hz - VESA
> 1280 x 1024 at 85Hz - BNQ
> 1600 x 1200 at 75Hz - VESA
>
> Raw EDID base
> 00: 00 FF FF FF FF FF FF 00 09 D1 06 01 89 00 00 00
> 10: 32 0D 01 03 6C 23 1A BE EA A7 B4 A3 52 46 98 24
> 20: 0F 48 4C AF CF 00 A9 4F 45 59 61 59 01 01 01 01
> 30: 01 01 01 01 01 01 86 3D 00 C0 51 00 30 40 40 A0
> 40: 13 00 68 0E 11 00 00 1E 00 00 00 FC 00 42 45 4E
> 50: 51 20 50 39 39 32 0A 20 20 20 00 00 00 FF 00 30
> 60: 30 31 33 37 0A 20 20 20 20 20 20 20 00 00 00 FD
> 70: 00 32 A0 1E 62 11 00 0A 20 20 20 20 20 20 00 B1
>
> Display adapter
> Adapter description......... NVIDIA GeForce 6600 GT
> Adapter device ID........... 0x014010DE
> Display settings............ 1920x1200, 32bpp
>
> ----------
>
> Oddly enough, the drop-down on top does not let me choose the 2nd monitor!
> It's as if this thing refuses to acknowledge the Dell LCD even exists.
>
> Cold-unplugging the BenQ and running the program again, I get... the BenQ
> info again. And still no sign that the Dell LCD exists.
>
> Note that I booted with both monitors plugged-in before running these tests.
>
> What now? :-(
>
>

The two URLs I provided, are as good a place to start as any.

I don't have two monitors to test here, on my video card, but I
expect if Entechtaiwan put a drop-down menu there, it was to support
the existence of more than one monitor at the same time.

It suggests to me, that something is up with DDC/CI and the EDID EEPROM.
It could be that the Entechtaiwan utility, won't deal with the EEPROM if it
has a bad checksum. Or, depending on what API the Entechtaiwan utility is
using, it could be that what the OS uses to "sniff" the data, has rejected
the second monitor, causing moninfo to be "blind" to the second monitor.

I tried the Viewsonic utility, but it is not obvious how you select between
monitors.

http://www.viewsonic.com/drivers/edid.exe

The program here, runs under Linux. You could use Knoppix or Ubuntu (700MB
ISO9660 download, burn to CD). They allow booting a computer from the CD,
and running Linux, without installing any software. Once booted into Linux,
you could visit this site and follow the instructions. There is also a DOS
version here, which I tested. It went into an infinite loop, saying it had
more blocks of data to read. Which means for the time being, the Viewsonic
utility is the next best alternative.

http://john.fremlin.de/programs/linux/read-edid/

What I was hoping to find, was a utility that could read the two
EDID interfaces, put the results in two files, and exit. Then it would
be possible to use other tools to deal with the small file that results.

Since the affected monitor is a Dell, you could try phoning their tech
support, and see if they know anything about an EDID problem. Or fire
up Google, and see if that model of monitor has a known problem like that.

I'm going offline now, boot into Linux, and try the Linux version of utility...

Paul

Paul wrote:

>
> I'm going offline now, boot into Linux, and try the Linux version of
> utility...
>
> Paul

I tried the Linux version of read-edid, and got different results than
the DOS version.

The Linux version does correctly return the EDID binary data block. Unlike
the DOS, it doesn't go into an infinite loop. My suspicion is, the DOS version
needs to be run in a real DOS environment, and not a command window. A
real DOS environment in my case, would be to boot the computer with my
Win98 prepared DOS floppy diskette, with a copy of the utility copied onto
the floppy. Bootdisk.com has various boot images, which could be used to do
something similar.

The Linux read-edid package is a bit peculiar. The manual page for the package,
claims the command takes no parameters. If you read the source, the program
does have options.

In get-edid.c , the options are listed as:

if ( argc > 3 )
{
fprintf( stderr, "%s: syntax %s [output-filename] [controller]\n",
argv[0],argv[0] );
return 3;
}

Thus, a Linux command line prompt would look like:

./get-edid output.bin 0

And if you are running under Knoppix, you would use the sudo command
to give yourself administrator privilege to run the command, as in:

sudo ./get-edid output.bin 0

That would get the EDID block and store it in "output.bin" in binary
form, from "controller 0". It is not at all clear to me, if you had
two video cards, with four controllers total, how the indexing of the
controllers would work. Values of "0" and "1", should select the two
main connectors on a single video card. Yet, when I tried it on my
ATI 9800Pro video card, with one monitor connected, the command
returned the same results for both controller values. In the text
messages on the screen, the CX register value printed on the screen,
does seem to be using the "controller number" passed on the command
line, so at least to that level, the right things are happening. The
program does seem to be passing the parameter to the video bios call.

I do own a second monitor (stored in my basement), but it is so old it
doesn't even have DDC on it. Which is quite frustrating. It is an old
RGBHV monitor, with five BNC coaxial connectors. A bargain at only
$3000 CDN when it was new :-(

So at the moment, that leaves "moninfo" and the Viewsonic program. On the
Viewsonic program, again with my one monitor connected, there is no indication
of any selector, to select which monitor.

Another thing that came to mind, while doing these experiments, is that
making the dummy VGA connector I suggested to you, won't work. If the
video card is relying on the EDID it is getting from the working monitor,
to make it happy, then making a dummy VGA connector that delivers no
EDID, isn't going to help at all. Sure, it would fool the video card
into thinking it has *some* kind of monitor connected, but if the problem
is somehow related to finding valid, non-corrupted EDID, then the dummy
connector likely won't change the symptoms at all.

Talk to Dell Tech Support. Maybe they know something about this problem.

In terms of me experimenting here, the "Gefen Box" is a more complete
emulation of a monitor being present, and would be the cheapest way I could
find of fooling my video card. For a component cost of about $10,
you could probably fashion something similar, using a serial EEPROM, two
pullup resistors for the serial bus, a VGA connector, three 75 ohm resistors
for coax termination, and build your own. Just a bit cheaper than the $70
they want for this.

http://www.gefen.com/kvm/product.jsp?prod_id=1378

Paul

"Mike J.S." <> wrote in message
news:46e73552$0$47102$ reenews.net...
> Okay, something weird is going on right now. I don't know what triggered
> it but I've been doing a lot of troubleshooting lately, before giving up
> and reformatting my HD (because of another unrelated issue).
>
> The monitor (a "Dell 2407WFP-HC" LCD screen) gets no signal on boot-up
> until Windows is loaded. Only then does it turn on. The screen is
> connected via DVI (looks noticeably better than VGA).
>
> MB is Asus A8N-SLI Deluxe, running latest Asus BIOS for this MB.
>
> Video card is Asus EN6600GT.
>
> ---
>
> Troubleshooting :
>
> When I plug a second simultaneous monitor in the video card's VGA slot (a
> BenQ CRT), BOTH the CRT and the LCD display the full boot-up routine (DOS
> visuals). And of course, the Windows portion.
>
> Leave the VGA slot unplugged (with only the DVI plugged-in), and the whole
> DOS boot-up routine is bypassed by the monitor again. It gets no signal
> until Windows is loaded.
>
> Loading BIOS defaults solves nothing.
>
> What gives? Why can't the DVI slot emit anything until Windows is loaded,
> unless something is plugged into the VGA slot? I need to plug in a 2nd
> monitor each time I want to boot in Safe Mode, or else I can't see the
> menus.
>

Is your CRT plugged into DVI Port #1 using the DVI-to-VGA converter and your
LCD panel into DVI Port #2? If it is then swap it around.

I have found with some graphics cards (usually the older ones) that Port #2
only becomes active once the graphics card driver gets loaded in Windows.
This seems to describe the symptoms you have posted.

DVI Port #1 is the primary port and always generates a video signal.

"Mike J.S." <> wrote in message
news:46e73552$0$47102$ reenews.net...
> Okay, something weird is going on right now. I don't know what
> triggered it but I've been doing a lot of troubleshooting
> lately, before giving up and reformatting my HD (because of
> another unrelated issue).
>
> The monitor (a "Dell 2407WFP-HC" LCD screen) gets no signal on
> boot-up until Windows is loaded. Only then does it turn on. The
> screen is connected via DVI (looks noticeably better than VGA).
>
> MB is Asus A8N-SLI Deluxe, running latest Asus BIOS for this
> MB.
>
> Video card is Asus EN6600GT.
>
> ---
>
> Troubleshooting :
>
> When I plug a second simultaneous monitor in the video card's
> VGA slot (a BenQ CRT), BOTH the CRT and the LCD display the
> full boot-up routine (DOS visuals). And of course, the Windows
> portion.
>
> Leave the VGA slot unplugged (with only the DVI plugged-in),
> and the whole DOS boot-up routine is bypassed by the monitor
> again. It gets no signal until Windows is loaded.
>
> Loading BIOS defaults solves nothing.
>
> What gives? Why can't the DVI slot emit anything until Windows
> is loaded, unless something is plugged into the VGA slot? I
> need to plug in a 2nd monitor each time I want to boot in Safe
> Mode, or else I can't see the menus.
>

Mike,

I had the same problem using an ATI X1600 board on an Asus P5W-DH
Deluxe motherboard with my Dell 2407 and it was driving me nuts.
I plugged both the VGA and DVI connectors into the monitor (I
know that someone on the Dell forum says not to do this) and
reinstalled using the inf file from Dell. I did the install
routine twice, once for the VGA and once for the DVI. Then I
unplugged the VGA cable from the monitor and I now see the full
boot sequence. The only drawback is that the monitor shows as
monitor #2 when you go into properties. I haven't gone back in
and tried to reinstall with just the DIV connection...if it works
don't fool with it.

I am using Vista Ultimate 64 bit so I am not sure the same will
work for you.

Regards, Dan

Thanks for your aid, Paul. I will call Dell tomorrow.

"Homer Jay Simpson" <> wrote in message
news:wLYFi.264197$...
>
> Is your CRT plugged into DVI Port #1 using the DVI-to-VGA converter and
> your LCD panel into DVI Port #2? If it is then swap it around.

There is only 1 DVI port, and 1 VGA port.

The LCD is using the DVI port (naturally).

> I have found with some graphics cards (usually the older ones) that Port
> #2 only becomes active once the graphics card driver gets loaded in
> Windows. This seems to describe the symptoms you have posted.
>
> DVI Port #1 is the primary port and always generates a video signal.

Problem with that theory is that it would make the VGA port the #1 port. But
I had this working just fine with DVI only a couple of weeks ago. Something,
somewhere, got corrupted... or a switch was, er, switched.

The DVI port now relies on something being plugged into the VGA port to be
able to display either a non-native resolution, OR to be able to display
*anything* without Windows drivers (not sure which of the two it is, and
there's no way to find out - not that it would bring us closer to a
solution).

"Dan Schumacher" <> wrote in message
news:kAZFi.533925$...
>
> I had the same problem using an ATI X1600 board on an Asus P5W-DH Deluxe
> motherboard with my Dell 2407 and it was driving me nuts. I plugged both
> the VGA and DVI connectors into the monitor (I know that someone on the
> Dell forum says not to do this) and reinstalled using the inf file from
> Dell. I did the install routine twice, once for the VGA and once for the
> DVI. Then I unplugged the VGA cable from the monitor and I now see the
> full boot sequence. The only drawback is that the monitor shows as
> monitor #2 when you go into properties. I haven't gone back in and tried
> to reinstall with just the DIV connection...if it works don't fool with
> it.
>
> I am using Vista Ultimate 64 bit so I am not sure the same will work for
> you.

I am using WinXP 64, so maybe this is 64-related?

I will give this a shot. Thanks, Dan.