??? 03/07/09 22:39 Read: times |
#163222 - But you don't do 80:1 Responding to: ???'s previous message |
80:1 multiplexing demands a lot from both the circuitry and the display. 10:1 demands less. Now you have reversed the multiplexing logic again :) A high multiplexing factor is cheaper than a low multiplexing factor. But the quota between peak current and average current means that you can't make the multiplexing factor too high without reduced light intensity. In this case, we have to assume that the OP can't afford to run a fully DC-driven display. In some cases, it may not even be possible with a fully DC-driven display because the LEDs are too close together so you can't manage drive electronics that fits on the available PCB estate - especially since the really large shift registers can't handle as much drive current because the power dissipation gets too high. You have another issue - if you scroll text horisontally in a multiplexed display, then the time delays between the rows will make the text look slanted. But if there isn't a constant time delay from row to row, then the text will not be slanted anymore but will instead be jagged. So that affects your ability to use multiple shift registers for a single text line. This makes it a bit problematic to use 2:1 or 4:1 multiplexing of a 8-pixel high text row. Looking at column multiplexing then. Lighting only one column at a time would represent a 80:1 multiplex factor - way outside the capabilities of a LED. It would either burn to cinders, or the display would be very, very, very dark. Lighting groups of columns would result in a number of visual effects that you wouldn't like, in case the information is scrolled. And in some ways it would also complicate the distribution of the power and the implementation of the code. After all, a 80-bit long shift register is very optimal for the processor. You want a minimum of "if" statements in the code to make best use of the processor. For a 12-clocker, the "if" statement would just be a wasted instruction where no data is emitted. For a faster processor, every jump represents a significant number of clock cycles where the processor have to first decide if it should jump, and then start fetching instructions from a new address. But anyway, in this case, the best multiplexing factor would be 4:1 or 8:1. Good light intensity but with limited number of resistors + shift registers. The most natural choice would then be 8:1. In the end, you will find that the world is full of multiplexed displays (and quite a number of fully DC-driven ones too) but a number of practical and visual limitations will interfere in the designs and you will normally end up with the traditional method of driving one scan line at a time while driving all columns at the same time. You may go for other designs if you never scroll information, or if the display isn't used for text but for example to show arrows scrolling up or down or flashing messages, ... |