ThinkWiki - User contributions [en]

Problem with Pixel Error

2006-08-02T10:07:12Z

Atesti:

This page provides information about pixel and subpixel errors.

==What are dead pixels or dead subpixels?==
Dead Pixels are a phenomenon caused by the technology of LCD displays. TFT displays consist of a huge number of individual display cells, each controlled by its own transistor. I.E. a display with a resolution of 1024x768 pixels consists of 1024x768x3 = 2,36 million cells. A such enormous number makes the production of a display without any faulty cell difficult. The more strict the quality restrictions are, the more expensive the panels get, because this simply means raising the number of garbage panels.

==Types of pixel errors==
Here is a listing of possible types of dead pixels / subpixels:
{| cellspacing=0 cellpadding=2 border=1 style="text-align:center"
! pixel type<hr />state type !! pixel !! red subpixel !! green subpixel !! blue subpixel
|-
| always lit ||   ·   ||   ·   ||   ·   ||   ·  
|-
| always unlit ||   ·   ||   ·     ·   ||   ·     ·   ||   ·     ·  
|}

You can notice here, that always lit complete pixels, red subpixels and green subpixels are the most annoying ones. Also, dead pixels located closer to the center of the screen are more annoying than such located closer to the edge of the screen.

Dead pixels do not always have to be a constant phenomena. You might in fact experience your pixel in question be faulty at times and working perfectly at other times. Or you might experience it to be a little flickering. This is only logical since a faulty transistor might sometimes work and sometimes not, depending on environmental conditions or even just randomly.

{{HINT|The table above also shows you a way of detecting dead pixels and subpixels. Just use plain colored screens.}}

==The ISO 13406-2 norm==
However, to give the concerned customer more transparency when buying LCD displays, the ISO norm 13406-2 was introduced, providing a basis for warranty claims. By assuring one of the classes of this specification, the producer is bound to keep that quality standard throughout the whole production series. However, Manufacturers are not necessarily going with this standard. For values for IBM displays look at [http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4U9P53 this page].

The values in the following table state the acceptable errors per 1 million pixels. Dead pixels form a cluster when they are within a square of 5x5 pixels.

{| cellspacing=0 cellpadding=2 border=1 style="text-align:center"
! Error class !! Type 1 (always lit pixel) !! Type 2 (always unlit pixel) !! Type 3 (always lit or always unlit subpixel) !! Type 4 (clusters with more than one Type 1 or Type 3 faults) !! Type 4 (clusters of Type 3 faults)
|-
| I || 0 || 0 || 0 || 0 || 0
|-
| II || 2 || 2 || 5 || 0 || 2
|-
| III || 5 || 15 || 50 || 0 || 5
|-
| IV || 50 || 150 || 500 || 5 || 50
|}

==Causes for dead pixels/subpixels==
* As stated above, dead pixels are an inherent feature of the technology/production process of TFT displays. How many of them are in the final displays sold is just a matter of quality control standards.
* Furthermore, they can also accur at random time after production. Transistors can have faults that are not recognizable at production time. These faults may make the transistor stop working after it has been in use for a certain amount of time.
* Mechanical or environmental stress (i.e. heat) can possibly damage a transistor, leading to pixel errors.

==Don't panic!==
Dead pixels are a really common phenomena on TFT displays. In fact there are other things about TFT displays that matter a lot more, like brightness and contrast, picture stability, color reproduction and response time. People just tend to focus on those things that are the most obvious, that's why everyone cries for dead pixels and response time when it comes to TFT displays.

However, as stated above there can be truely annoying types of dead pixels. If their count is higher than the displays quality class permits you might get the device replaced under warranty.
The best thing, however, is to look for dead pixels right after purchase (you might have a chance to return it without special reason then):
* Do it three times over two days - different moods and different environemental light have an influence on your perception.
* [http://freshmeat.net/projects/lcdtest lcdtest] utility can generate predefined patterns, and may help you to detect damaged pixels.
* Be concious about how hard it was to find the dead pixel - if it was really hard it can't be too bothering.
* Also make sure that what you consider a dead pixel isn't just a piece of dirt or dust.
* Try pressing the (blue) scroll button on the UltraNav. There is a bug in the UltraNav driver which causes fake Bright pixels to appear on the screen, especially if FireFox is used as browser. See [[UltraNav]] for more information

==Solutions==
Besides some april jokes there isn't really any solution except looking for a replacement.

However, for {{Windows}} an interesting program exists that can help you making annoying subpixels and pixels more bearable. It works by trying to compensate a dead subpixel by adjusting the values for the remaining working subpixels. This way you can i.e. get rid of a most annoying always lit red subpixel. The program is german only and a download link can be found in its german [http://www.prad.de/board/thread.php?threadid=11797 PixChanger discussion].

If you have a '''line''' of dead (bright or dark) pixels, this is usually caused by a failure of one of the address-lines where it contacts the edge of the panel. You can sometimes restore it by (''carefully''!!) flexing the display. Find the point on the bezel where the line ends, then push it gently with a thumb (or from behind) while supporting the outer edges of the display with 2 fingers. The LCD needs to be
"folded" by about 1 degree; the fold direction depends on the problem. I can report success with an X20. Note that this is a last resort; it '''might destroy your LCD''', and the LCD should thereafter be treated exceptionally gently!

==External sources==
* [http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4U9P53 IBMs page on defective pixel display replacement criteria]
* [http://graphics.tomshardware.com/display/20030319/index.html "An autopsy of Dead Pixels" at Tom's Hardware Guide]

Problem with Pixel Error

2006-08-02T10:06:39Z

Atesti:

This page provides information about pixel and subpixel errors.

==What are dead pixels or dead subpixels?==
Dead Pixels are a phenomenon caused by the technology of LCD displays. TFT displays consist of a huge number of individual display cells, each controlled by its own transistor. I.E. a display with a resolution of 1024x768 pixels consists of 1024x768x3 = 2,36 million cells. A such enormous number makes the production of a display without any faulty cell difficult. The more strict the quality restrictions are, the more expensive the panels get, because this simply means raising the number of garbage panels.

==Types of pixel errors==
Here is a listing of possible types of dead pixels / subpixels:
{| cellspacing=0 cellpadding=2 border=1 style="text-align:center"
! pixel type<hr />state type !! pixel !! red subpixel !! green subpixel !! blue subpixel
|-
| always lit ||   ·   ||   ·   ||   ·   ||   ·  
|-
| always unlit ||   ·   ||   ·     ·   ||   ·     ·   ||   ·     ·  
|}

You can notice here, that always lit complete pixels, red subpixels and green subpixels are the most annoying ones. Also, dead pixels located closer to the center of the screen are more annoying than such located closer to the edge of the screen.

Dead pixels do not always have to be a constant phenomena. You might in fact experience your pixel in question be faulty at times and working perfectly at other times. Or you might experience it to be a little flickering. This is only logical since a faulty transistor might sometimes work and sometimes not, depending on environmental conditions or even just randomly.

{{HINT|The table above also shows you a way of detecting dead pixels and subpixels. Just use plain colored screens.}}

==The ISO 13406-2 norm==
However, to give the concerned customer more transparency when buying LCD displays, the ISO norm 13406-2 was introduced, providing a basis for warranty claims. By assuring one of the classes of this specification, the producer is bound to keep that quality standard throughout the whole production series. However, Manufacturers are not necessarily going with this standard. For values for IBM displays look at [http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4U9P53 this page].

The values in the following table state the acceptable errors per 1 million pixels. Dead pixels form a cluster when they are within a square of 5x5 pixels.

{| cellspacing=0 cellpadding=2 border=1 style="text-align:center"
! Error class !! Type 1 (always lit pixel) !! Type 2 (always unlit pixel) !! Type 3 (always lit or always unlit subpixel) !! Type 4 (clusters with more than one Type 1 or Type 3 faults) !! Type 4 (clusters of Type 3 faults)
|-
| I || 0 || 0 || 0 || 0 || 0
|-
| II || 2 || 2 || 5 || 0 || 2
|-
| III || 5 || 15 || 50 || 0 || 5
|-
| IV || 50 || 150 || 500 || 5 || 50
|}

==Causes for dead pixels/subpixels==
* As stated above, dead pixels are an inherent feature of the technology/production process of TFT displays. How many of them are in the final displays sold is just a matter of quality control standards.
* Furthermore, they can also accur at random time after production. Transistors can have faults that are not recognizable at production time. These faults may make the transistor stop working after it has been in use for a certain amount of time.
* Mechanical or environmental stress (i.e. heat) can possibly damage a transistor, leading to pixel errors.

==Don't panic!==
Dead pixels are a really common phenomena on TFT displays. In fact there are other things about TFT displays that matter a lot more, like brightness and contrast, picture stability, color reproduction and response time. People just tend to focus on those things that are the most obvious, that's why everyone cries for dead pixels and response time when it comes to TFT displays.

However, as stated above there can be truely annoying types of dead pixels. If their count is higher than the displays quality class permits you might get the device replaced under warranty.
The best thing, however, is to look for dead pixels right after purchase (you might have a chance to return it without special reason then):
* Do it three times over two days - different moods and different environemental light have an influence on your perception.
* [http://freshmeat.net/projects/lcdtest lcdtest] utility can generate predefined patterns, and may help you to detect damaged pixels.
* Be concious about how hard it was to find the dead pixel - if it was really hard it can't be too bothering.
* Also make sure that what you consider a dead pixel isn't just a piece of dirt or dust.
* Try pressing the (blue) scroll button on the UltraNav. There is a bug in the UltraNav driver which causes fake Bright pixels to appear on the screen. See [[UltraNav]] for more information

==Solutions==
Besides some april jokes there isn't really any solution except looking for a replacement.

However, for {{Windows}} an interesting program exists that can help you making annoying subpixels and pixels more bearable. It works by trying to compensate a dead subpixel by adjusting the values for the remaining working subpixels. This way you can i.e. get rid of a most annoying always lit red subpixel. The program is german only and a download link can be found in its german [http://www.prad.de/board/thread.php?threadid=11797 PixChanger discussion].

If you have a '''line''' of dead (bright or dark) pixels, this is usually caused by a failure of one of the address-lines where it contacts the edge of the panel. You can sometimes restore it by (''carefully''!!) flexing the display. Find the point on the bezel where the line ends, then push it gently with a thumb (or from behind) while supporting the outer edges of the display with 2 fingers. The LCD needs to be
"folded" by about 1 degree; the fold direction depends on the problem. I can report success with an X20. Note that this is a last resort; it '''might destroy your LCD''', and the LCD should thereafter be treated exceptionally gently!

==External sources==
* [http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4U9P53 IBMs page on defective pixel display replacement criteria]
* [http://graphics.tomshardware.com/display/20030319/index.html "An autopsy of Dead Pixels" at Tom's Hardware Guide]

UltraNav

2006-08-02T10:05:24Z

Atesti: Ultranav bug documented. Please help!

{| width="100%"
|style="vertical-align:top;padding-right:20px;width:10px;" | [[Image:ultranav.jpg|IBM UltraNav]]
|style="vertical-align:top" |
<div style="margin: 0; margin-right:10px; border: 1px solid #dfdfdf; padding: 0em 1em 1em 1em; background-color:#F8F8FF; align:right;">
===IBM UltraNav===
The classic integrated pointing device in ThinkPads was always a trackpoint. With the T30 IBM introduced UltraNav, a combination of both the classical [[TrackPoint]] coupled with a programmable touchpad. The technology for this combined pointing device comes from Synaptics. The TouchPad features all kind of customizable ways of input, from standard pointing over scrolling by movement along the edges, tap zones, to ignoring accidential touches.
</div>
|}

==Linux Support==
Both the TrackPoint and the TouchPad work with the standard ps2/psaux driver of linux kernels.

To get advanced configurability for the touchpad working there are several drivers for Synaptics TouchPads available for X:
*[[Synaptics TouchPad driver for X]] is an X input driver
*[[tpconfig]] is a tool to configure the TouchPad device before the actual input driver takes over

To get advanced configurability for the TrackPoint, look at the [[Patch to enable advanced trackpoint configuration | TrackPoint kernel patch]].

==Windows Support==
The UltraNav driver from IBM is based on the Synaptics driver. It contains a bug leading to "defective pixels":
If the scroll-function of the touchpad is used (moving finger on right border of touchpad), often taskmanger shows
an application/window on its first tab named "Syn Visual Window". Sometimes this stays on the screen and in task
manager. It is a 1x1 pixel large window that is usually white. If you move the mouse directly over this pixel,
a little icon appears like if the middle trackpoint button is used to scroll.
This pixel can be removed by terminating SynTP* processes in taskmanager. Some claim that it can be removed by
using the middle trackpoint button, althought this is not always the case.
So far this problem has been reproduced on T60P and T41P, but only by scrolling large wegpages in Firefox.
After Firefox was closed, the pixel is still there.
IBM was able to reproduce this by installing Firefox on a fresh recovery image in their labs. But they
refuse to fix this or pass it to Synaptics or their driver developers, because Firefox "is not supported".
Guess IBM only wants you to use IE.
This bug is quite annoying because people think it's a wandering defective pixel

==Models featuring UltraNav==
* {{R40}}, {{R50}}, {{R50p}}, {{R51}}, {{R52}}
* {{T30}}, {{T40}}, {{T40p}}, {{T41}}, {{T41p}}, {{T42}}, {{T42p}}, {{T43}}, {{T43p}}, {{T60}}, {{T60p}}
* {{Z60m}}, {{Z60t}}

[[Category:Glossary]] [[Category:Components]]