??? 02/13/06 19:33 Read: times |
#109880 - No illnes seen until today Responding to: ???'s previous message |
Erik Malund said:
bit-bang which is 1% of processing total is still 100% of processing when it takes place. If you can accept 100% at such times, who cares, if you can't then you need HW IIC.
A single byte on SW-I2C cost only 100% for 100µs of the main loop time. Since typically a main loop must be fast from the humans view, no human was able to detect a 100µs delay. You can also often see a HD44780 LCD connected over PCF8574 and SW-I2C without noticeable slowing down the whole program from the humans view, e.g. pressing a key was displayed instantly. Erik Malund said:
So, Peter, in many cases you can use bit-bang with no ill effect, but, as I post, not in all. Should that be a reason to shun HW IIC? that is where we disagree. Please can you explain what do you mean with "ill effects" ? SW-I2C need no interrupts and need also no interrupts to be disabled. So all fast things are fully unaffected by SW-I2C. And again, I know no main loop which can not tolerate several 100µs delay. On my experience a HW-I2C was only really needed in very rarely cases or to support I2C slave or multimaster. Peter |