I am trying to make a new Atari 5200 compatible controller based around an optical joystick wired to an up/down Digipot. I believe that I will need to place several digipots in series on each axis to make this work. Is that even possible? If interested, please read on for further details.
Because the "null" resistance value (ideally 250K) as seen by the game logic is not consistent among all 5200 consoles let alone all games in the library, the axes on my controller need to be trimmable to guarantee a joystick-centered neutral directional state in-game. I have determined that the range of resistance change in the original controllers from extreme to extreme in either orientation is around 280 kohms. For trimming, I can source inexpensive and reliable 15-turn trimpots with a 220K range. I chose 220K because when summed with 280K it gives a nice round number of 500K, and that total happens to match the resistance range of the original rotary potentiometers in the original controller. Thus if for a given axis I put that trim pot in series with an "effective digipot" of 280 kohms, I can shift the resistance value at the "stick-centered-position" back and forth by 110K, which should be ample to make the game logic see zero deflection on a wide variety of unit/game combinations. By "effective digipot" I mean that since a single 280K digipot from Analog Devices does not exist so far as I can tell, if the total required current is not excessive perhaps I can put several digipots in series, that have their up and their down inputs hooked to the same optical outputs from the optical stick? I can source cheap optical thumbsticks that give 80-90 pulses from center position. Assuming the lowest total pulse count of 160, if I go with 128 step digipots, I am certainly able to reach their full step count so long as there is neither mechanical impediment to limit the stick travel, nor excessive accumulation of missed pulses in any one direction to make drift an issue. Thus I would require a total of six digipots: qty2 of 100K and qty 1 of 80K on both axes.
Is this even realistic? If so, what type of digipot would be best?