bitwhacker
New Member
- Joined
- Apr 11, 2024
- Messages
- 6
I'm in the process of fixing a Compaq Portable Plus. One of the many, many, many problems it has is that the VDU board has a fault which results only in some characters being corrupted. Like this:
The first corrupted character has pixels changing randomly while the screen is static, the second character is a static zero with a light background. If actual characters a displayed, they are corrupted. The pattern repeats every 8 columns, most of the time, but sometimes every 24 columns.
A little tangent, since I don't have a convenient way of writing 5.25 inch disks, I hooked up some random 3.5" 1.44MB floppy drive I had laying around to the Compaq, and it works fine as a 720kB drive! I connected an old USB floppy drive to my 12-core Windows 10 PC with 32GB of RAM, and was able to format a DD disk in CMD with the command format a: /f:720! I created an MS-DOS 3.3 boot disk using WinImage, mounted it in DOSBox-X and typed in the BASIC the video RAM test program from the SAM'S manual in BASICA, saved it, created an autoexec.bat to run it on boot ... and it's completely crazy that this all worked!
I changed the LPRINT to PRINT, since I don't have a printer hooked up. Anyway, the test program tells me that the problem is U26. This is the output (it runs in CGA mode):
Now, the schematic in the SAM'S manual is for a different revision of the board, but U26 is hooked up to data bit 4. I confirmed that on my older revision board (ASSY 000007-001 REV P), that's U66. I then desoldered U66 and booted up the machine without it. This is the result:
Unfortunately, I'm not at all familiar with how the video memoy on MDA/CGA works. Is it plausible that one of the RAM chips is defective, or is the culprit more likely to be one of the buffer or latch chips or even the MC6845? Where do I go from here?
The first corrupted character has pixels changing randomly while the screen is static, the second character is a static zero with a light background. If actual characters a displayed, they are corrupted. The pattern repeats every 8 columns, most of the time, but sometimes every 24 columns.
A little tangent, since I don't have a convenient way of writing 5.25 inch disks, I hooked up some random 3.5" 1.44MB floppy drive I had laying around to the Compaq, and it works fine as a 720kB drive! I connected an old USB floppy drive to my 12-core Windows 10 PC with 32GB of RAM, and was able to format a DD disk in CMD with the command format a: /f:720! I created an MS-DOS 3.3 boot disk using WinImage, mounted it in DOSBox-X and typed in the BASIC the video RAM test program from the SAM'S manual in BASICA, saved it, created an autoexec.bat to run it on boot ... and it's completely crazy that this all worked!
I changed the LPRINT to PRINT, since I don't have a printer hooked up. Anyway, the test program tells me that the problem is U26. This is the output (it runs in CGA mode):
Now, the schematic in the SAM'S manual is for a different revision of the board, but U26 is hooked up to data bit 4. I confirmed that on my older revision board (ASSY 000007-001 REV P), that's U66. I then desoldered U66 and booted up the machine without it. This is the result:
Unfortunately, I'm not at all familiar with how the video memoy on MDA/CGA works. Is it plausible that one of the RAM chips is defective, or is the culprit more likely to be one of the buffer or latch chips or even the MC6845? Where do I go from here?
