Intel describes in the Bible that the HCMOS outputs on ports 2, 3, and 4 actually have a kind of latch mechanism that is designed to keep the port pins at a high level when they have been commanded to go high. On a pin that then reverts to an input driven from the outside this latch actually goes off as the pin is driven low. This is pretty cool in that the off board driver does not need to sink the current of the weak pullup and will tend to save battery life in systems with the microcontroller V_cc being derived from a battery. After all, the pull-up current comes from the V_cc pin.

The really curious thing to me is to wonder whether all the derivative 80C51/80C52 type parts implement this same circuit configuration. Maybe not all parts work the same way. I do know that Cygnal parts for instance have an additional mode where each pin on the ports can be programmed as a push-pull output. For pins that operate in full time output mode this can offer quite an improvement in the rise times sent to peripheral chips. Do note that since Cygnal parts operate at a V_cc of 2.7 ^_> 3.3V. The push-pull output mode works great when you need swings up to 3.3V on the outputs. If you need to interface the Cygnal part to logic that needs 5 volt output swings (and yes the Cygnal part is 5V compatible) it is necessary to use the normal quasi-bidirectional port pin modes and then use external pullups to the +5V supply).

Michael Karas