The most important piece in the puzzle when it comes to this terminal build has to be the choice of display, I already knew that I wanted a legitimate green phosphor cathode ray tube but I was also going to need something that was an appropriate shape and size and was also something I could interface with to actually produce images.
I wasn’t really sure where to start but I figured I needed some idea of the size I was looking for, I don’t think a huge variety of sizes of green CRTs exist anyway but to get some idea I started out by loading the game model and trying to establish a scale. As I’ve discovered from building various other props, objects from the game are very rarely scaled relative to the real world but usually you can take something of a known size and use that to base things off. For this model I picked the keyboard as my gauge as keys tend to have a fairly standard size of 19mm, scaling the model to fit that I can now take a measurement of the screen space to give me a rough idea what to look for, which is 304mm or 12" for the metrically impared.
The Amstrad GT-65
Now that I knew I was looking for a 12" green CRT I took to eBay to see what was on the market that wasn’t ultra rare unobtainium and quickly came across quite a few Amstrad GT-65 displays from the Amstrad CPC home computers from the mid 1980’s for actually very reasonable prices, so reasonable I actually accidentally ended up with two.
I suspect that the amount of these I found was primarily due to being located in the UK, these displays I learned were also sold under the Schneider name to the German market. The tube itself is an Orion 310GNB31 which I suspect may well have been used in other devices of the same time, as I understand it the now defunct Orion Electric also produced CRT’s under the JVC name and may well have sold to others.
The screen space on the Orion tube is pretty much 12" and the aspect ratio is also perfect for this use case, seen below the highlighted shape is the front of the tube and the outline is the bezel from the front of the RobCo terminal.
Interfacing
So I’ve found the perfect tube, but that’s only half the battle, can it be used with a modern computer? Well as it turns out this display does not have composite, nor does it use an RF modulated antenna type connection, instead Amstrad use a somewhat proprietary connection based around a 6 pin DIN connection. The connection was originally intended for use with the CPC computer series which was typically a colour display experience, the monochrome was being sold as the cheaper option at the time, hence why the pinout also includes signals for red, green and blue.
The fantastic thing about electronics of this era is how often service manuals were available that contain full schematics of the device and part numbers for sourcing replacement parts, this schematic allows us to see that on this monochrome display, the RGB signals are all pulled to ground and the only connections actually used are sync and luma on pins 4 and 6 respectively.
Signal Processing
Now that I know the sync and luma pins are what are used to interface to the display, I needed to understand what signals these expected. Fortunately I am far from the first person to hook more modern devices up to these displays and I found quite a few hints online that set me on the right track.
The luma signal here is more or less what you’d expect from a composite video signal, a voltage that signals the brightness or luminance of what is being drawn and then the sync signal is a CSYNC, a composite of both HSYNC and VSYNC. In a typical composite video connection the sync signals are part of the composite of red green and blue. I had read some cases where simply inputting a regular composite video signal into both luma and sync should be enough to get something to display, but I wanted a cleaner signal so I decided to take another approach.
Queue the LM1881. Texas Instruments create an IC (the LM1881) which is designed to strip the timing signals out of a composite video signal, the perfect tool for the job. I got my hands on a couple of these IC’s and put together a basic test circuit on a breadboard.
Prototyping an interface
Initially I built the prototype based off the example in the LM1881 datasheet, but reading some more of the datasheet I realised I didn’t actually need some of the elements so I stripped back my breadboard to just enough components to do the bare minimum functionality I required, splitting out the CSYNC signal.
Below shows an example of the processed signal, the yellow trace is the input composite video signal, you can see the sync followed by a series of levels. Then below that the blue trace is the output CSYNC signal, a cleaned up and simplified signal just containing the sync components.
Once I was happy with basic circuit I moved my prototype onto a more permanent proto board and also added a basic PAM8302 mono channel audio amplifier. The composite output from the Raspberry Pi cable now goes straight to this board and breaks breaks out both audio and video to produce our luma and sync signals for the CRT and an amplified mono audio signal ready to go straight into a speaker.
Configuring the OS
Raspberry Pi OS can output composite video by default but I took a few extra steps to ensure a correct resolution was being passed and some margins to fit the output to the tube.
To get the Pi to output composite video at a standard PAL resolution I first needed to update the kernel cmdline params in /boot/firmware/cmdline.txt to add video=Composite-1:720x576@50e. To remove any unnecessary signal I also added tv_mode=Mono to send a purely black and white signal.
Now it was just a case of some trial and error to set margins to fit content into the screen space I wanted to use, I ended up adding margin_left=80,margin_right=30,margin_top=20 which as shown in the resulting picture below does have some overscan, but for the application this screen will be running this will be perfect and allow content to reach right into the edges at the center, even if the corners are cut off here.
