Why was it possible to cause an Apple //e to shut down with SHIFT and paddle button 2?

Question

Looking at the MAME driver for the Apple //e, I noticed an interesting comment:

NOTE: On real IIe and IIe enhanced h/w, pressing SHIFT and paddle button 2 will short out the power supply and cause a safety shutdown. (We don't emulate this "feature", and it was relatively rare in real life as Apple joysticks only had buttons 0 and 1 normally).

I've never heard of this before, though I don't recall ever utilizing paddle button 2 either.

Is this an actual problem that always existed? Would reproducing it by hooking up the right pins on the Game I/O port be unsafe or potentially damaging to the //e or is it completely harmless because the circuitry is designed to handle such an event? Why does this "feature" exist?

Answer 1

The keyboard on the Apple II+ was designed to type uppercase letters only, which rather limited the machine's usefulness for word processing. Because almost nothing used paddle button 2 and very few controllers even had a third button, a common user-installed modification was to run a wire from the shift-key signal to the paddle-button-2 input. Word processing programs could then be set to use paddle button 2 to check whether the shift key was down and use that to decide whether to type uppercase or lowercase. This was far from perfect, since if e.g. a user typed shift+A and then released both keys while the system was doing something, the system would then poll the keyboard, find out an "A" had been pushed, observe that the shift key was not (still) pressed, and type a lowercase "a". It was nonetheless superior to being unable to type lowercase letters at all, or requiring that users press ESC before any uppercase letter (the other approach word processors used).

Although the Apple //e allows lowercase letters to be typed directly, the designers wanted to be compatible with code that expected the shift-key mod to be installed, so they simply included the connection on the motherboard directly.

The problem is that paddle buttons were intended to be used by wiring a resistor that would weakly pull the input to +5V when the button isn't pushed, and then having the switch pull the input to ground when pressed. The shift key was wired the opposite way--passively pulled down, with a switch to +5V. If Apple had foreseen the potential issue this could cause, it could have been averted by simply adding a resistor in series with the switch. Alternatively, Apple could have included a jumper that would connect that input to either the appropriate pin on the internal controller socket or to the shift key. My guess is they simply failed to recognize the possibility of a problem since paddle 2 inputs would be used so rarely. It was certainly never intended as a "feature".

As for whether this would cause damage, it would certainly cause excessive stress on the switch contacts used for the shift key and/or paddle 2 input, and would likely stress the power supply components as well. Doing this once or twice might not cause the components to die instantly, but one probably wouldn't have to do it too many times to wear out the components.