Panel Install Help Needed

  • Thread starter Thread starter ck
  • Start date Start date
C

ck

Hello,

I am trying to get a NEC NL10276BC28-05D panel working with a Commell
LV-677 motherboard. I know the panel is in working condition, but no
matter what I try, I can seem to get any video to show up on the
panel. The system itself is running fine. I can see that on my
monitor running from the VGA port.

Any ideas? Ive been working on this for days and i'm at a loss.

Thank you in advance.

-Chris
 
ck said:
Hello,

I am trying to get a NEC NL10276BC28-05D panel working with a Commell
LV-677 motherboard. I know the panel is in working condition, but no
matter what I try, I can seem to get any video to show up on the
panel. The system itself is running fine. I can see that on my
monitor running from the VGA port.

Any ideas? Ive been working on this for days and i'm at a loss.

Thank you in advance.

-Chris
Click to expand...

Not enough information. The panel you are talking about doesn't appear to
have a standard data connector. And if you are using the VGA (DB15) port of
the mainboard to run a different monitor, then what we need to know is

How, EXACTLY, does the panel interface with the mainboard?

I'm guessing you are doing some kind of customized kludge to try to get it
to work. In that case, you need to find the exact signal specifications
(not just type, but signal level) on both ends and make sure that the input
requirement for EACH pin is exactly what the corresponding output is
sending.

It is likely you will find that you have made a wiring error somewhere, or
that the data the mainboard is sending is somehow incompatible with what the
panel expects to receive. -Dave
 
Not enough information. The panel you are talking about doesn't appear to
have a standard data connector. And if you are using the VGA (DB15) port of
the mainboard to run a different monitor, then what we need to know is

How, EXACTLY, does the panel interface with the mainboard?

I'm guessing you are doing some kind of customized kludge to try to get it
to work. In that case, you need to find the exact signal specifications
(not just type, but signal level) on both ends and make sure that the input
requirement for EACH pin is exactly what the corresponding output is
sending.

It is likely you will find that you have made a wiring error somewhere, or
that the data the mainboard is sending is somehow incompatible with what the
panel expects to receive. -Dave
Click to expand...

I am trying to use the onboard LVDS connector on the motherboard.
Assuming the cable is good (checking it now), and the connectors are
wired correctly on each end, the Windows side of the setup should be
simple, correct? Im using the IEGD drivers for the Intel 945GM
display chipset.

Thank you!


-Chris
 
ck said:
I am trying to use the onboard LVDS connector on the motherboard.
Assuming the cable is good (checking it now), and the connectors are
wired correctly on each end, the Windows side of the setup should be
simple, correct? Im using the IEGD drivers for the Intel 945GM
display chipset.

Thank you!


-Chris
Click to expand...

There is a datasheet for the panel here. The 20 pin has LVDS with
three data pairs and a clock pair. And needs up to 5V @ 600mA on
the cable as well. Lamp powering is separate I guess. The three data
pairs carry 21 bits, of which 18 are color, leaving an Hsync, Vsync,
and a "DE".

http://www.ortodoxism.ro/datasheets/nec/NL10276BC28-05D.pdf

This is a datasheet for a 24 bit NEC panel. I wanted to see this, to verify
what the hardware interface looks like. This uses 4 data pairs and a
clock pair. 7 bits times 4 pairs, is 28 bits, of which 24 are the color
bits, leaving two static (undefined) signals, a reserved signal, and
"DE". I presume the 24 bit panel extracts Hsync and Vsync from the DE
signal somehow - not really sure.

http://www.esskabel.de/Datenblaetter/LCD-Specs/NEC/TFT/NL10276BC30-10.pdf

Looking at the Intel 945GM datasheet, the LVDS interface only shows
three data pairs, and as near as I can tell, is intended for 18 bit
panels. Which makes me wonder why the Commell manual for the LV-677
shows four data pairs on the connector pinout ?

ftp://download.intel.com/design/mobile/datashts/30921903.pdf (945GM)

I'm guessing, when you made your cable, you connected ATX0, ATX1,
and ATX2 pairs to the panel ? JVLCD jumper set to 5 volts ? And been
extremely careful with the definition of pin one ? (Gotta watch
which view of the connector they are showing, and where the pin 1
actually is.)

When I look in the Commell downloadable BIOS, version 1.4, I see with
my hex editor, a module inside the BIOS called "cal_1284dvo.dat", and
my guess would be that it is the video BIOS included for the integrated
graphics. I notice the 945GM downloadable driver package, also includes
video BIOS files as separate files, and they have names with "cal_1413"
in them. The video BIOS is probably only important, if you wanted the
BIOS setup screen to be viewable on the flat panel. Otherwise, the
Windows driver would eventually display a picture of the Windows desktop,
if the video BIOS wasn't the right one.

Time to fire up that oscilloscope :-) You'll need 100 ohm diff termination
on the end of the cable for the clock pair and the data pair (i.e. if trying
to scope the end of the cable, or at the connector). LVDS uses
current source/sink drive, and the resistor termination across the
diff data pair, is necessary to convert the current flow, into a voltage.
The voltage is quite low amplitude. Then probe single ended with your
high bandwidth, light loading, FET probe. Probing the clock signal first,
would be the most satisfying choice. The driver would only enable
the LVDS, if the display was supposed to be active. The driver should
be "power optimized" and won't drive the LVDS, unless commanded to.

Isn't debugging fun :-) I've even had to do it, with a couple
managers looking over my shoulder :-)

Found another doc. Might want to have a browse thru this - won't fix
your problem, but might provide a bit more background and references.
At least this explains what Intel was thinking, when they put a
3 pair data interface on the LVDS.

"Using 24-bpp LVDS Panels with Intel Mobile Chipsets for Embedded Applications"
http://www.intel.com/design/intarch/papers/315975.pdf

Paul
 
There is a datasheet for the panel here. The 20 pin has LVDS with
three data pairs and a clock pair. And needs up to 5V @ 600mA on
the cable as well. Lamp powering is separate I guess. The three data
pairs carry 21 bits, of which 18 are color, leaving an Hsync, Vsync,
and a "DE".

http://www.ortodoxism.ro/datasheets/nec/NL10276BC28-05D.pdf

This is a datasheet for a 24 bit NEC panel. I wanted to see this, to verify
what the hardware interface looks like. This uses 4 data pairs and a
clock pair. 7 bits times 4 pairs, is 28 bits, of which 24 are the color
bits, leaving two static (undefined) signals, a reserved signal, and
"DE". I presume the 24 bit panel extracts Hsync and Vsync from the DE
signal somehow - not really sure.

http://www.esskabel.de/Datenblaetter/LCD-Specs/NEC/TFT/NL10276BC30-10...

Looking at the Intel 945GM datasheet, the LVDS interface only shows
three data pairs, and as near as I can tell, is intended for 18 bit
panels. Which makes me wonder why the Commell manual for the LV-677
shows four data pairs on the connector pinout ?

ftp://download.intel.com/design/mobile/datashts/30921903.pdf(945GM)

I'm guessing, when you made your cable, you connected ATX0, ATX1,
and ATX2 pairs to the panel ? JVLCD jumper set to 5 volts ? And been
extremely careful with the definition of pin one ? (Gotta watch
which view of the connector they are showing, and where the pin 1
actually is.)

When I look in the Commell downloadable BIOS, version 1.4, I see with
my hex editor, a module inside the BIOS called "cal_1284dvo.dat", and
my guess would be that it is the video BIOS included for the integrated
graphics. I notice the 945GM downloadable driver package, also includes
video BIOS files as separate files, and they have names with "cal_1413"
in them. The video BIOS is probably only important, if you wanted the
BIOS setup screen to be viewable on the flat panel. Otherwise, the
Windows driver would eventually display a picture of the Windows desktop,
if the video BIOS wasn't the right one.

Time to fire up that oscilloscope :-) You'll need 100 ohm diff termination
on the end of the cable for the clock pair and the data pair (i.e. if trying
to scope the end of the cable, or at the connector). LVDS uses
current source/sink drive, and the resistor termination across the
diff data pair, is necessary to convert the current flow, into a voltage.
The voltage is quite low amplitude. Then probe single ended with your
high bandwidth, light loading, FET probe. Probing the clock signal first,
would be the most satisfying choice. The driver would only enable
the LVDS, if the display was supposed to be active. The driver should
be "power optimized" and won't drive the LVDS, unless commanded to.

Isn't debugging fun :-) I've even had to do it, with a couple
managers looking over my shoulder :-)

Found another doc. Might want to have a browse thru this - won't fix
your problem, but might provide a bit more background and references.
At least this explains what Intel was thinking, when they put a
3 pair data interface on the LVDS.

"Using 24-bpp LVDS Panels with Intel Mobile Chipsets for Embedded Applications"http://www.intel.com/design/intarch/papers/315975.pdf

Paul
Click to expand...

Paul,

Thank you SO much for the help. My engineer is going to start
debugging first thing tomorrow morning.

He did however just inform me that the actual panel is an NEC
nl10276bc28-11a. I can't seem to find a datasheet for it anywhere.
Happen to know where it can be located.

I'll definitely update you with my progress as I make some.

Thanks again!

-Chris
 
ck said:
Paul,

Thank you SO much for the help. My engineer is going to start
debugging first thing tomorrow morning.

He did however just inform me that the actual panel is an NEC
nl10276bc28-11a. I can't seem to find a datasheet for it anywhere.
Happen to know where it can be located.

I'll definitely update you with my progress as I make some.

Thanks again!

-Chris
Click to expand...

There is a ref here - 1998 makes the panel 9 years old.
http://www.necel.com/english/news/9802/2601-01.html

This is a case where, the datasheet should be in hand,
before you buy the panel. If you bought a pile of these
panels, maybe you'll end up tearing one apart, and try to
reconstruct the pinout. Either that, or convince someone
at NEC, to provide you with a nine year old datasheet.
The sites that archive datasheets, probably weren't
around nine years ago. And since NEC didn't put the
datasheet on their web site, using web.archive.org probably
isn't going to help. (I checked, and the archived page is
unusable.)

One unrelated web site, mentioned plugging in that unit,
and having it ground a power rail. Implying you should be
careful with the pinout.

Paul
 
In alt.comp.hardware.pc-homebuilt "ck said:
Paul,

Thank you SO much for the help. My engineer is going to start
debugging first thing tomorrow morning.

He did however just inform me that the actual panel is an NEC
nl10276bc28-11a. I can't seem to find a datasheet for it anywhere.
Happen to know where it can be located.

I'll definitely update you with my progress as I make some.

Thanks again!
Click to expand...
The best *I* can find, is this cable for the beast:
http://www.lcdfriends.com/picture/200541222305130.pdf
What's annoying is that I can find spec-sheets for just about every
*other* NEC LCD panel except that one!

Geesh.
Perhaps one of the other LCD panels that said cable also supposedly
works for, has similar specifications?
 
Frank said:
The best *I* can find, is this cable for the beast:
http://www.lcdfriends.com/picture/200541222305130.pdf
What's annoying is that I can find spec-sheets for just about every
*other* NEC LCD panel except that one!

Geesh.
Perhaps one of the other LCD panels that said cable also supposedly
works for, has similar specifications?
Click to expand...

Well, that helps. It shows that at least the pinout is the
same on the 11A and the 05D. And there is a datasheet
downloadable for the 05D.

Paul
 
Well,

I am able to get an image on the lcd panel. The problem is that the
panel is native 1024 x 768. When I use the Intel Embedded Graphics
Driver to "clone" the two monitors (lcd attached to VGA port & lcd
attached to LVDS port), I can get an image on the LVDS panel. When
the driver sets the resolution of the LVDS panel to 1024 x 768, all I
can see is some vertical lines jumping around the screen. When I set
the LVDS panel to 640 x 480, i can see the desktop, but it is too
large for the screen horizontally, and then the bottom is cut off and
the top repeats again on the panel. Basically I see a double image of
everything, shifted down the panel.

We checked the cable according to the lcdfriendly.com PDF (thank you)
and it seems correct.

Any other ideas?

-Chris
 
ck said:
Well,

I am able to get an image on the lcd panel. The problem is that the
panel is native 1024 x 768. When I use the Intel Embedded Graphics
Driver to "clone" the two monitors (lcd attached to VGA port & lcd
attached to LVDS port), I can get an image on the LVDS panel. When
the driver sets the resolution of the LVDS panel to 1024 x 768, all I
can see is some vertical lines jumping around the screen. When I set
the LVDS panel to 640 x 480, i can see the desktop, but it is too
large for the screen horizontally, and then the bottom is cut off and
the top repeats again on the panel. Basically I see a double image of
everything, shifted down the panel.

We checked the cable according to the lcdfriendly.com PDF (thank you)
and it seems correct.

Any other ideas?

-Chris
Click to expand...

To figure out what is going on, you need a well defined stimulus. Not a
full color 24 bit image of some sort, but a saturated simple geometric.
Like a green horizontal line at a defined location. Then a green
vertical line at a defined location. With a pure black background.
See how the mapping responds. Change to red and repeat. Change to
blue and repeat. Perhaps you can see a pattern to what is happening.

The fact that you got an image, is a small miracle. I congratulate you :-)
I figured getting the backlight enabled would have been a problem.

Paul
 
In alt.comp.hardware.pc-homebuilt Paul said:
To figure out what is going on, you need a well defined stimulus. Not a
full color 24 bit image of some sort, but a saturated simple geometric.
Like a green horizontal line at a defined location. Then a green
vertical line at a defined location. With a pure black background.
See how the mapping responds. Change to red and repeat. Change to
blue and repeat. Perhaps you can see a pattern to what is happening.
Click to expand...
Better yet would be a diagonal line.
It's hard to figure out overlaps and repeats with strict X and Y.
 
Back
Top