|
4.865 bytes |
Service Hints & Tips |
Document ID: COBN-385CF6 |
|
This document is provided to IBM and our Business Partners to help sell and/or service IBM products. It is not to be distributed beyond that audience or used for any other purpose. |
G40 Monitors - Wobble problem
Applicable to: World-Wide
Service Information:
This is a summary of the G40 'wobble' problem and actions taken to date for your information. Please distribute this information carefully.
Monitors Affected:
G40 (6542-1xx)
Systems Affected:
Attached to PC330/350 Pentium Units (Reading Planar) with S3Trio764 video chipsets. Potentially any system unit out putting VGA at 31.3 KHz is affected.
Modes Affected:
BIOS setup mode/Boot Manager/Diagnostics DOS text mode/ Windows or OS/2 full screen DOS sessions etc (720x400 at 70Hz VGA modes)
Symptoms:
Screen appears to wobble or ripple in height, width or position. Can sometimes be described as "jitter", but it is very regular appearance.
Cause:
The Motorola microprocessor in the G40 is very sensitive to HSync frequencies that are very close to harmonic of the micro clock. When this happens, the DAC control outputs ripple at the beat frequency between the two signals.
The G40 uses a 0.5% ceramic resonator to run the microprocessor. A small part (believed to be 10%) of this distibution come close enough to the abnormally low frequency of the S3Trio on the PC330/350 to show the problem.
Fixes:
1. Increased filtering on the DAC outputs reduces the ripple except in a few cases (believed 2%) where the frequencies lie within 10Hz of each other.
2. A 0.3% tolerance ceramic resonator can be used to reduce the number of failing units to about 5%.
3. A crystal oscillator (+/- 80 ppm) can be used in place of the ceramic resonator. It is believed that this will completley fix the problem for the PC330/350 planars.
4. Change oscillator frequency to avoid clash with existing modes. This requires a new mask in the micro to change the counter values.
5. Change BIOS in system unit to correct VGA frequency.
Implementation:
Fix 1: Broken into card production Dec. 1995 at Liteon. Box build from Feb. 1996 onwards.
Fix 2: As above.
Fix 3:
a. Rework required for 30k cards WIP in Greenock.
b. Request Liteon to implement in new cards.
c. Qualify source of crystal/ELP test
Fix 4: Not pursued for G40 - it will be well implemented on follow-on.
Fix 5: RTP system PE working with Intel (planar design)
Field Fix:
Fix on fail critical customers with crystal oscillator (3) above.
Service:
Issue crystal fix as mandatory EC to repair to cenres.
|
Search Keywords |
| |
|
Hint Category |
Display/Monitor | |
|
Date Created |
11-03-96 | |
|
Last Updated |
20-01-99 | |
|
Revision Date |
20-01-2000 | |
|
Brand |
IBM Monitors | |
|
Product Family |
G Series (G40/G50/etc.) | |
|
Machine Type |
6542 | |
|
Model |
V05 | |
|
TypeModel |
| |
|
Retain Tip (if applicable) |
| |
|
Reverse Doclinks |