>The I2C spec is not explicit about when clock stretching can occur.

I thought it was pretty explicit, at the time I looked at it.

>It implies that it is done by a slave after a the slave sends
>an acknowledge, which is how Microchip implements it in PICs
>by using the CKP bit.

Umm, it really does it at (or during) the time of sending the ACK bit, not 
after.

And that is how the MChip devices do it (although some have found problems 
with 16F88 IIRC). The slave device holds the CLK low during the ACK bit time 
so that it doesn't go high when the master allows it to go high, resulting 
in the stretched clock bit.

Technically I think you can have a slave device do this on every clock bit 
if it cannot clock the data fast enough, but I personally don't know of any 
devices that do this. This may be the area where you feel the spec is not 
explicit. 

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist