| ??? 10/16/06 22:06 Read: times |
#126542 - Erik, go back and read the whole thing Responding to: ???'s previous message |
look at the illustration, too.
Your're right, it uses, actually, 13 pins, right? but you get 24 for them, and a lot of flexibility that you didn't have before. Further, for ONE additional bit, you can have another entire 8255, which gives you yet another 24 bits. Under some circumstances, that's not too high a price to pay. Erik Malund said:
1) when using external memory the 8255 can be attached at a low "pin cost" but will require most modern uCs to run at less than maximum speed. 2) when not using external memory, attaching a 8255 can be done at a very high "pin cost" but will not require the uC to slow down (except for 8255 access stretch which typically will have only a nominal effect). In this configuration the gain is miniminal since it eat about as many pins as you gain I've never denied either of these assertions. I don't think sacrificing 13 pins to get 24 is a particularly good deal, though it is a substantial benefit. Sacrificing 14 to get 48 is a bit better, and 16 to get 96 is quite a bit better. That won't come up often, thank goodnes, but it may coe up some day. The only novice that needs more than 32 I/O pins is one that needs more than 32 I/O pins. RE |
