June 2007
A WORK IN PROGRESS
I am trying to interface an LCD panel from a vintage 486 Compaq Contura 420cx to the VGA port at 640x480 resolution with 3 bits of color (one bit per color -- RGB) for the purpose of displaying data from my upcoming ENSURE-2 near-space project. I couldn't find a datasheet or pinout for the Hitachi TX26D61VC1CAA LCD panel (and inverter) so I had to figure it out. Here is the pinout for both the panel and inverter.
I started with the VESA Feature Connector approach (English translation) with no success. The pixels never lined up. Perhaps it was just the graphics card that I was using, but I'm not sure. Here is the best image that I could produce from the feature connector:
I want a more compatible approach anyway, so I tried this circuit to get the video signal from any old analog VGA port. I was able to get a centered 3-bit color image this way, but the horizontal pixels would smear together, indicating that something was creating a delay in the color channels. I found that the clock signal from the graphics card was being loaded too much by the LCD panel, so I put the clock signal through another pair of NOT gates first. That fixed inconsistencies in the image, but did not fix the horizontal smearing problem. Photo:
I decided that the double NOT gate approach was not able to handle the high pixel frequency (~25MHz, see the circuit description). In then end, using a single OR gate for each channel and tying the B inputs to ground did the trick. Here is the relevant truth table...
| A | B | Out |
| 0 | 0 | 0 |
| 1 | 0 | 1 |
A single OR gate will perform the same logical function as two NOT gates in series, but faster, since the signal only passes through one gate. I am still getting the pixel clock signal from the graphics card, since I can't find a 25.175MHz crystal anywhere. To get the clock signal to the circuit without a separate cable I routed pin 18 (the PIXEL CLOCK pin) of the VESA Feature Connector on the graphics card to the unused pin 9 of the VGA connector as so:
Of course, I wired it up accordingly at the other end of the VGA cable as well. Success! Below are some test shots (BTW it can display all three colors, not just green, but I first tested each color channel individually):
YES! The pixels are perfect! except for that R there in 'R4'. I noticed that it would become distorted every time I moved my hand close to the panel or circuit. Some sort of RF interference? Also the image is not quite centered, and part of the image is cut off. This could be due to the fact that the resolution is not 640x480 -- I will have to look into this.
Unfortunately these images are the only ones I have before the PSU screw-up. Yes, I am a brilliant child, I am. I accidentally switched the +12V and +5V lines and fried the works. The panel no longer keeps a steady image. It appears that the part of the LCD that handles the VSYNC signal is no longer working, and the image just rolls continuously. Great. Now I've got to find another panel... Anyway, below is a link to the new schematic diagram:
New schematic (still being tested)
I'll finish this writeup as soon as I get another panel...
Questions? Comments? Email me: kc0wys AT gmail DOT com.