This page documents what I went through to adapt a Sony
GDM-20D11 monitor, in late May 2003 to my PC. Laurent
Granjon's
prior effort started my investigation, along with Adam
Kropelin's modification of the
same for Sun.
This page first created: 30 May 2003, last mod: 08 June 2003
There are many different video connector standards but for this document, only two are discussed: HD15, for the PC and 13W3, used by workstation makers.
HD15
|
13W3As used by workstation makers chiefly for fixed frequency monitors, 13W3 has no well-defined pin definitions. As show on gruftie.net, IBM, Sun, SGI and Intergraph placed different meaning on the pins. On the GDM-20D11, pins A1 to A3 are R, G, B, 'normal' enough, but pins #2 through #5 are NC while the rest (pin #1, #6 through #10) are ground. |
I am no authority on this, but as far as I can tell, there are at least 3 ways to transport video synchronization signals from your graphic card to the monitor.
The Sony GDM-20D11 is a 21" Trinitron™ monitor, made circa 1995. (Somewhere on the web mentioned 1994, but the two GDM-20D11s I came across are dated June and September 1995.)
Gathered from all over the web, including MonitorWorld.com, its specifications are:
| Dot pitch | 0.30 mm |
| Max Resolution | 1600 x 1200 @ 66 Hz |
| 1280 x 1024 @ 76 Hz | |
| Freq, Horizontal | 48-82 KHz |
| Freq, Vertical | 50-150 Hz |
 |
Step 1: Remove the back cover by taking out the 4 screws indicated |
 |
Inside, it is neat and tidy. Everything is compartmentalized and modular. You begin to feel the GDM-20D11 is a well-engineered piece of work. Next, we need to remove the aluminium cover furthest to the back. |
 |
Step 2: Remove this screw .. |
 |
.. and this screw |
 |
With the aluminium cover removed, our main target is exposed. Step 3: Remove the 4 screws indicated. |
 |
Our module of interest is loose. Note down the order and color of the 6 sets of wires plus the 3 R.G.B. so that you can reconnect them later. Strictly speaking, noting down the order and color is not necessary, as the connectors are color-coded. Together with the difference in number of pins, you cannot reconnect the wires wrongly. Evidence #2 of a well-engineered product where pains were taken for connectors to be color-coded. Step 4: Disconnect the wires. |
 |
Our 'Sync-Signal cum Service-port' PCB. Click the image for a larger view. |
|
Evidence #3 of good engineering: signal pins are generously labeled. Click image for a larger view. |
 |
As far as video synchronization signals are concerned, the input to this module/PCB is, naturally the 13W3 video connector, while the output (to the rest of the monitor) are 4 signals (in addition to the basic R.G.B. signals of course):
On the extreme top left corner, there is a test signal switch. With this, the GDM-20D11 by itself can test its picture tube by generating a whole screen of white.
There is a reset switch (below the test signal switch) which is used to reset the video settings you made. This reset switch is accessible without taking apart the monitor. I guess this reset switch functions like the reset button on the remote control.
The left one-third of the PCB is dedicated for the service port which is an 8-pin DIN-like connector. Signals are RS-232 standard since a Maxim MAX232 driver/receiver IC is used.
| Click image for a larger view. |  |
The CXA1616 integrated circuit is a sync discriminator for CRT displays that automatically selects one of three types of sync signals to output; separate sync, composite sync, or sync-on video. What this means in English is that the chief function of the CXA1616 is to act as a 3-to-1 selector; selecting from a choice of 3 inputs to 1 output. (Physically, separate sync is a pair (horizontal and vertical) signal, composite sync is one signal, sync-on is one signal, and output is a pair (horizontal and vertical) signal.)
How it decides is a matter of priority; from high to low: separate sync, composite sync, and sync-on. If separate sync signals are available, it will be used regardless of the other 2 inputs. If separate sync is not available, composite is up next, and so on. Lastly, sync-on will only be used if the other 2 input types are not present.
'Byproducts' of the CXA1616 are:
No longer made after 31 December 2000 (as noted here), its datasheet (.PDF), available here and here, is where the above is gleemed.
The Sony CXA1616 on our 'Sync-Signal cum Service-port' PCB is of a SDIP form factor.
Its circuit surrounding its 3 input pins (#1, #4 and #7) is as follows:
We need to modify it to be the following, where the changes are highlighted in red.
The component values of 75 ohms and 4.7 microfarad are as suggested by the CXA1616 datasheet and are not computed to be correct/optimized. They are used by me without adverse effects that I can see.
Modified as shown, I have the GDM-20D11 working in 1280 x 1024 @ 75 Hz on my nVidia GeForce 2 MX400 in Windows XP. When booting up, (ie, during POST, etc), the GDM-20D11 is not blank but shows a triple image of what is going on, typical of not able to sync to the 31.5kHz(?).
I plan to delve next into the GDM-20D11's synchronization circuit(s) see what can be done about this. If you've any ideas, please do leave a comment below.