Originally Posted by
iMalc
I agree. Knowing there are a moderate number of errors occurring at all in some cases made me nervous. Our software gathers statistics on packets with wrong CRCs (amoung other things), and some sites have reported error rates of at least 1 in 50,000. Even a 1 int 256 chance of not detecting the error is awful when you're receiving on the order of 500,000 packets per day.
Unfortunately I am working with a third-party product here, so whilst I don't strictly have the ability to change the protocol between devices, I am hoping to gather enough information to persuade the third party that a protocol change is necessary.