Charlieplexing, Gugaplexing & now Chipiplexing

[rWAVE]

The November 27, 2008 issue of EDN (page 59) or located here describes the latest of the "plexing" approaches to driving multiple LEDs with few uC ports.

Chipiplexing (derived from the author's nickname) has some advantages over Charlieplexing, including driving N-1 LEDs simultaneously, using only N additional NPN/PNP (very cheap) transistors.

Richard

 

[MikeGyver]

Can picaxe chips even do this 'high impedence' output state required for charlieplexing?

 

[Jeremy Leach]

Yes, on picaxe pins that can be either inputs or outputs, assigning them as inputs makes them high impedence.

 

[rmeldo]

Can the opposite be done and save input pins when reading from keypad?

I probably said an heresy,....... but there you go: novices.

 

[Jeremy Leach]

Not sure what you mean Rmeldo ??

 

[rmeldo]

What I was trying to ask is:

can the circuit above, or better, the concept above (use 3 pins to light > 3 LEDs) be applied in reverse: use 3 pins to read the inputs from > 3 keys from a keypad, including simultaneous key pressed?

 

[BeanieBots]

As virtually EVERY concept can be 'reversed' I'd stick my neck on the line and say YES without even looking at either the method described or what might be involved with reversing it.

For inputs, the analogue resistor ladder and ONE input works very well if you actually want a tried and tested method.

 

[Tom2000]

Well, it's interesting, but I question its practicality.

The author says that you can turn on N-1 LEDs simultaneously. Cool. But in your program, how do you decide which LEDs can be turned on at any given time. Or, more precisely, how can you keep track of which LEDs are constrained from activation at any given time?

For small panels, this might not be a factor. But for large LED arrays where a multiplexing scheme becomes attractive, tracking LED permissions might be a real trick.

Tom