3 ms·

For anyone interested in how the "motion circuit" works: The basic building block is a counter chain, some simple counter chips in series. The horizontal count

For anyone interested in how the "motion circuit" works: The basic building block is a counter chain, some simple counter chips in series. The horizontal counter chain counts clock pulses for a single horizontal video scan line, which is also used to generate the horizontal blank interval. Reset it each time, it reaches the end of a scan line. Now we have a digital counter in sync with the video signal. (Actually, we'll generate the video signal from this.) Another, vertical counter chain is used to count the rows and is pulsed each time, the horizontal counter chain is reset. Put a simple AND gate on the output of individual pins of the horizontal counters to put the output signal HI and you get vertical stripes. If we combine some output signals of the counters, we may narrow this down to single bar of arbitrary width. Do the same with the vertical counters and AND it with horizontal signal and you get what is known as a window, a basic rectangle. So, how can we move this on the screen? Enter another set of counter chains, but this time with a bit of head room, which may be populated by a preset. These are known as slipping counter chains. E.g., have a 3-bit overhead and a preset of up to 4 bits, now the slipping counter chain will take head of the static one or fall behind it by up to 7 pixels in either direction, depending on the preset. (The extent of the "head room" is also known as the stop code, as setting the preset to that same value will result in a stationary signal. If the preset is zero, the counter chain will take 7 pulses longer to reach its end than the static one. If we fully populate the preset, it will finish early by the same amount.) If we start the slipping chain at a specific position of the static one, we may now move relative to it. Voila, we have a moving "window", which may be controlled by these preset values. Basically, you need a set of slipping counter chains for each moving object and a set of static ones to generate the basic video signal and its sync pulses, as well as any static on-screen objects. For hit detection, just compare the signals emitted by those counter chains. For some more detail (on the example of Computer Space) see: https://www.masswerk.at/rc2017/04/02.html https://www.masswerk.at/rc2017/04/02.html An amazing technical source on the subject is "The Textbook of Video Game Logic; Volume 1" by Arkush, William (Kush N' Stuff Amusement Electronics, Inc., Campbell, CA, 1976) [1] [1] (PDF) http://www.tokensonly.com/images/resources/manuals/video-game-logic-textbook-vol-one.pdf http://www.tokensonly.com/images/resources/manuals/video-gam...

weinzierl 5y ago
"Now we have a digital counter in sync with the video signal." This principle was maintained for many years until the homecomputer era came to an end. The CPU (and all the auxilliary chips) used to run in lock-step with the video signal. The whole architecture of many games from that time was built around the position of the electron beam on the screen. We exploited that to switch video modes in the middle of the screen (text/graphics split screen), draw in the overscan area (raster bars) or multiply the number of hardware sprites by quickly alternating different sets at certain scanlines. The V-sync pause was a welcome rest and spent to calculate the game logic. Another interesting consequence of this was that US homecomputer models ran at faster clock speeds than the Euro models because of the higher refresh rate of NTSC. With clock speeds in the single digit megahertz range the 60/50 ratio was not negligible. In Europe we did hw mods to add a switch that made the computer run at NTSC speed even if that meant no video. You had the choice: a fast computer with no output or a slow one with video.
- sokoloff 5y ago
  We’d turn off the video system to speed up calculations (by removing its periodic accesses to memory, so not the same as NTSC/PAL above). I haven’t touched an Atari in over 30 years and I still remember the memory location is 559 (decimal).
JKCalhoun 5y ago
With all the Apple I clones and the like, I am wondering why I have not come across a hardware Pong clone — using original TTL logic (not a modern microcontroller or FPGA).
- opencl 5y ago
  It's been done, the schematics are readily available and the chips are still easy to find (technically not the original 74xx models, but the pin-compatible 74LS and 74HC series). https://forums.arcade-museum.com/threads/reproduction-pong-pcb.422533/ https://forums.arcade-museum.com/threads/reproduction-pong-p...
- 5y ago

weinzierl 5y ago

"Now we have a digital counter in sync with the video signal." This principle was maintained for many years until the homecomputer era came to an end. The CPU (and all the auxilliary chips) used to run in lock-step with the video signal. The whole architecture of many games from that time was built around the position of the electron beam on the screen. We exploited that to switch video modes in the middle of the screen (text/graphics split screen), draw in the overscan area (raster bars) or multiply the number of hardware sprites by quickly alternating different sets at certain scanlines. The V-sync pause was a welcome rest and spent to calculate the game logic. Another interesting consequence of this was that US homecomputer models ran at faster clock speeds than the Euro models because of the higher refresh rate of NTSC. With clock speeds in the single digit megahertz range the 60/50 ratio was not negligible. In Europe we did hw mods to add a switch that made the computer run at NTSC speed even if that meant no video. You had the choice: a fast computer with no output or a slow one with video.

sokoloff 5y ago
We’d turn off the video system to speed up calculations (by removing its periodic accesses to memory, so not the same as NTSC/PAL above). I haven’t touched an Atari in over 30 years and I still remember the memory location is 559 (decimal).

JKCalhoun 5y ago

With all the Apple I clones and the like, I am wondering why I have not come across a hardware Pong clone — using original TTL logic (not a modern microcontroller or FPGA).

opencl 5y ago
It's been done, the schematics are readily available and the chips are still easy to find (technically not the original 74xx models, but the pin-compatible 74LS and 74HC series). https://forums.arcade-museum.com/threads/reproduction-pong-pcb.422533/ https://forums.arcade-museum.com/threads/reproduction-pong-p...
5y ago