Atari CX40 joystick

An exploded view of the Atari CX40 joystick shows its basic construction. The top lid is not shown in this diagram.

When the stick is moved, one side of the hemisphere (right) moves down to press on the buttons while the opposite side is forced upward (left) to provide a recentering force.

The Atari joystick works by connecting the ground pin to one of several pins in the Atari joystick port, thereby dropping the voltage on that pin and creating a signal that can be noticed by a controller in the computer.[8] For this reason, Atari-style joysticks are sometimes referred to as "digital joysticks", largely to differentiate them from the analog joysticks found on systems like the Apple II and IBM PC.[9]

The main structure of the CX40 is formed from a concave moulded thermoplastic base with a separate flat lid that covers the opening on the top of the base. Four cylindrical protrusions on the inside of the base hold a printed circuit board (PCB) above the bottom, roughly centered vertically. A conical post on the base passes vertically through a hole in the middle of the PCB.[10]

The PCB has five membrane switches mounted on top. Four of the switches are arranged in a cross pattern around the hole in the middle of the PCB; the fifth is offset near one of the corners. The PCB connects the switches to pins in the 9-pin D-connector that leads to the console via a cutout in the upper right corner of the base. The cutout is shaped to grip a moulded strain relief on the cable.[10]

The stick itself is a moulded polypropylene form consisting of a hollow vertical cylinder with a hemispherical dome at one end. The stick is placed on top of the conical post in the base, and the lid is placed over it. This clamps the stick between the post and the circular cutout in the lid. When pressure is applied to move the stick, it can rotate on the post by sliding the hemisphere within the circular cutout.[10]

The hemisphere has four small "fingers" at the bottom, which are positioned to lie over the switches on the PCB. Cutouts in the hemisphere make the fingers mechanically separate from the main section of the stick, allowing them to flex so they do not press too hard on the switches. When the stick is moved, the hemisphere rotates on its support post, bringing the fingers down to press on the appropriate buttons. At the same time, the opposite side of the hemisphere rises, where it comes into contact with short protrusions on the lid. These push down on the flexible section of the hemisphere, providing a centering force that returns the stick to the upright position when pressure is released.[10]

The stick is assembled by placing the PCB within the lower case and routing the cable out of the box. The stick is then placed on top of the post, and a flexible rubbery cover is placed over the stick. A red plastic button is placed over the fifth switch, providing the fire button; and the top lid is then placed on top of the entire assembly. Four machine screws inserted from the bottom of the case through the PCB support protrusions hold the lid on, holding everything in place through compression.[10]

Although the stick can be pressed in any direction, the four fingers on the hemisphere allow only two switches to be pressed at a time. This allows a total of eight directions: four for up, down, left. and right, and four more for combinations of neighboring switches: up-left, up-right, down-left, and down-right. The physical movement on the post prevents non-neighbouring switches from being pressed; one can not press up and down at the same time, for instance.[10]

The base and its cover are moulded with some roughness in the plastic. The portion directly around the stick and the fire button are smooth. Just around the opening for the stick are a series of rectangular embossings with the four cardinal directions indicated by small wedge shapes. The 12 o'clock position replaces the wedges with the word "TOP". These markings are all painted orange, but this tends to wear off over time.[10]

Right-handed users would normally hold the stick in their left hand by gripping the left side of the case with their thumb over the fire button, and fingers wrapped around the left side and bottom, while their right hand operated the stick. Some games allowed the stick to be rotated for left-handed users, or this could be done with some minor re-wiring.[11]

The ergonomics of the base and the need to hold it while pressing the fire button led to fatigue of the left thumb that became known as "joystick thumb".[12] This was especially common in games that required the user to fire repeatedly or hold down the button for long periods, like Eastern Front (1941). This led to various "autofire" attachments, available as do-it-yourself plans,[13] pre-assembled, or as parts of 3rd party sticks. These simulated rapid presses of the fire button as long as the button on the stick was held down.[12]

Small rubbery pads on the bottom allowed for stick placement on a desktop. However, the torque needed to operate the internal switches was more than enough to flip over the lightweight stick. Some 3rd party replacements used a much larger base and/or a shorter vertical stick so the torque was reduced. Generally those with bases large enough for one-handed use tended to be too large to use two-handed. Modifications using suction cups, also found on some 3rd party sticks, were generally regarded as more trouble than they were worth.[12]

The CX40 was infamous for its eventual breakdown from one of two problems. One was that the control ring on the bottom of the stick moulding was not very robust, and either the ring itself or the four small tabs that connected it to the stick could be broken by applying too much force. The ring transmitted the force from the stick to the switches, so if it broke, the stick would no longer operate properly in that direction. This happened with "distressing regularity".[6] The other problem was that the switches themselves would eventually wear out due to repeated use, normally with the top membrane cracking or the tape holding it in place wearing off.[14][15]

A common modification was to replace the switches with miniature snap-action switches (microswitches). Some of these, however, have the problem that their actuation is non-linear; they require considerable force to start moving compared to the force needed to complete the motion. When used in an otherwise unmodified CX40, this caused it to be more difficult to move along the diagonals,[12] as one of the two switches being pushed would normally reach the threshold first, causing motion in that direction while the other was not yet pressed. In games that required fine control, like Jumpman, these solutions were generally unsuitable. Newer switches improve this action.[16]

The Commodore VC1311 was introduced after their clone of the CX40 was ordered off the market.

The CX40 was the only model in widespread use until about 1981, when the introduction of Space Invaders on the 2600 quadrupled that platform's sales.[24] This led to a thriving 3rd party market for better sticks for use on the 2600.

Many were almost identical copies of the CX40. Other examples of barely-modified sticks include the Gemini GemStick, which was essentially a CX40 with a somewhat larger base and the fire button replaced by a larger yellow version. The Gemstick 2 added a second button in the upper right. The Suncom Slik-Stik was slightly more modified with the case rounded off and a large red plastic ball at the tip of a shortened stick. The shorter stick made it more suitable for single-handed use. Suncom later introduced the TAC-2, essentially a larger Slik-Stik with buttons on both sides, for left-handed users.[25]

For the VIC-20, Commodore International introduced a joystick that was essentially identical to the Atari model with the exception that the top plate on the base was white instead of black. This prompted Atari to sue for patent violations.[26][27] Commodore responded by introducing the 1311, a modified version that made the case more rectangular, changed the stick from a cylinder to a triangle, and replaced the single round fire button with a larger one running across the top of the stick, making it suitable for use in either hand. These changes also rendered it painful to use and it was widely panned.[28][29]

Happ's Competition Pro used microswitches in place of Atari's membrane switches. The stick (red) presses on the switch actuators (white part on blue components).

The Wico Command Control was one of the earliest examples of a stick that differed radically from the Atari pattern in mechanical terms. Wico was a major supplier of arcade game joysticks and adapted these mechanisms for home use.[12] These replaced the membrane switches with leaf switches that were much more robust and provided the recentering force internally. A large rounded-corner square block of plastic provided the moving surface similar to the hemisphere in the CX40, but was shaped to provide more obvious directions when pressed, sliding along the edges of the square. The stick, a steel tube covered in a bat-like red plastic moulding, featured a button on top that pressed a shaft running through the center of the stick to another switch below the main assembly.[30]

The base was greatly enlarged with the specific intent of making it useful in one-handed play, which operated in concert with the stick-top button. A switch on the base selected whether the button on the handle or the base was active. The case also featured moulded indents in the bottom that made it easier to hold in two-hand use. The result was a system that was significantly more robust than the Atari sticks, but also much more complex and expensive. Later versions added a switch for "autofire" that caused the fire button to be repeatedly pressed while held down, and even later the bat-like handle was replaced by a ball like those found in the arcade. It was consistently rated very highly.[6][25]

The Epyx 500XJ was a radical restyling of the basic CX40 concept. This version, for the NES, has additional buttons.

The Epyx 500XJ[lower-alpha 3] was among a very few designs that broke from the Atari mould completely. Like the Competition Pro, the 500XJ's mechanism was based around a steel shaft pressing on microswitches and offered a similar feel. Unlike the Command Control, the 500XJ's shaft pressed directly on the switches, making it harder to press into the diagonals. More importantly, the base of the unit was completely changed, consisting of a moulded form that was designed to be easily gripped by the left hand with the index finger naturally positioned over a fire button located on the bottom right side of the case. Both robust and unique, the 500XJ garnered widespread praise and was the only other joystick of the era to achieve widespread popularity to the point of becoming a standard of sorts.[31]

Early gamepads were essentially CX40s with the base exposed to the user and the joystick on top removed. Sega's Master System and Genesis/Mega Drive controllers are compatible with the 2600, although the console can only read one fire button on the Sega controllers – the 2 button on Master System controllers, and the B button on Genesis controllers. The Sega Control Stick[32] was the only Sega-made controller actually marketed for its compatibility with Atari and Commodore systems. Due to the unique setup of the 7800's fire buttons, Sega controllers are not compatible with 7800 games that require two fire buttons, although they can be used with 7800 games that only use one fire button.

The Atari joystick was so popular that adaptors for its standard of connector were available for platforms like the Apple II and IBM PC.[33][19] There were even factory modified CX40s, notably the Trisstick from Big Five Software, that placed the converter electronics inside the CX40 case and replaced the cable with one suitable for the TRS-80.[34]

While its single-button configuration makes it unsuitable for general use with Sega's consoles, an Atari joystick can be used in the player two port for certain two-player games with minimal controls, so long as a standard Sega controller is plugged into the player one port. For example, a CX-40 can be used to control Tails in two-player mode in Sonic the Hedgehog 2, as all three Genesis fire buttons are mapped to the same jump/spin-dash function (and with the Master System for the 1 and 2 buttons), although player two cannot pause the game due to the absence of a Start button.

• ATARI VCS/2600 CONTROLLERS, lists dozens of Atari-compatible joysticks