> I just had that happen to me with the Philips I2C PCF8563 RTC chip
> where my first read/write scheme worked great until the test people
> had at it and we discovered that the chip has some strange
> "undocumented" behaviours when its I2C bus is used in a certain
> way with bit banging and then a long interrupt comes along and makes
> a 270 uSec hole in the I2C sequence to the chip!!. It was re-design
> time last weekend....ARGH).

Can you provide more details about that problem? I have been (and will
be) using the 8593 which is probably basically the same chip without
the RAM (and so with much less current consumption), and might
have the same "undocumented" behaviours...

Thanks, Tilmann