TimeSquare Multiplexing

[adammhaile]

I'm trying to wrap my head around how the TimeSquare watch handles the multiplexing of the LEDs on the display...
A lot of matrix tutorials I've seen have all said to drive it using some transistors, but this is obviously not happening in the TimeSquare.
How is it driving each column without putting too much current stress on each I/O pin? Whether done as a current sink or current source setup, those LEDs are usually around 10mA each (maybe more) and 10mA x 8 per column is definitely greater than the 40mA max for an ATMega I/O pin.

I'm trying to do something similar and get an 8x8 matrix working without one of the driver backpacks, I just don't fully understand how this is supposed to be done (which is not helped much by the fact that the adafruit matrix datasheets aren't really clear on which pin is which).

I generally understand how the code is working, I just am unsure of the current limiting.

[adammhaile]

Upon looking at the build guide, I found another oddity. The limiting resistors are all 47 ohm. And since most LEDs have a forward voltage of 1.7V - 2.2V, that would make it seem that they are probably shooting for running the LED at 20mA, which would drop ~1V from the 3V battery and bring the LED to 2V. So that makes the current calculations even worse with 20mA x 8 = 160mA per source or sink.

[pburgess]

There's a lot of shenanigans going on...part of it is that the LED matrices were selected for their high brightness and low current...they switch on at just a couple mA. Then there's the software driving the display, which seldom actually does 1:8 multiplexing (it's usually 1:16, or even 1:64 in certain cases). The current also is inherently limited by the battery.