OK ... you've got to build a system to recover data from a serial channel at nominally 250 kbps (+/- 1.5%) in unknown format and modulation. You need to look at each "bit-window" to see where, with respect to that window, there's a change in the signal level from 1 to 0 or vice versa. Based on when that happens, you decide whether the bit reperesents a one or a zero, but, aside from that, you've got no information about where the bit-window lies. Occasionally, there are specific artifacts that appear in the asignal sequence, of which you have to be aware. Further, there are external signals that occur at a rate not specifically related to the bit-rate.

Several different modulation schemes are possible, and you probably have to deal with that but not necessarily during the sample interval.

There are also other signals that occur from time to time that are related in one way or another to the operation of the equipment from which this data comes, so you have to pay attention to those, at least one of which tells you when to start and when to stop sampling.

You store the accumulated samples in buffer memory, and, when the sample is complete, you transfer the data over an async comm link to a remote PC or whatever, using a standard communication rate.

The value of the data is high, time is short, as the need is urgent, and the volume set for production is exactly ONE.

Now, where does that $3 UART fit into the greater scheme of things?

RE