This week another of the familiar yellow bubblewrap envelopes from China dropped on my doormat, delivering some ds89c21 rs-422 line driver chips.
The ds89c21 chips fit the board designed in the previous post and led to some first tests with a rs-422 differential link. First of all, I found that I had omitted the bias resistors necessary for the rs-422 link on the link input in my design. These were easily bodged to the board for a quick test, the ugly results of impatient soldering shown to the right. But no success, the link input remained pinned to a logic high signal.
A few days of debugging later, it suddenly dawned to me to cross the wires of the link and the link immediately sprang to life. Doh. For now the link seems to be reliable even at 20 Mbit/s, great.
I rechecked the schematic and the board silkscreen, seems that even though the ds89c21 seems to match the board, these plus/minus lines on either the receive or transmit lines are swapped in respect to the board layout.
The oscilloscope picture to the right shows that the negative signal has some nasty degradation, while the positive is pretty decent, strange. It may just as well be the terrible bodged-on bias resistors, but I can’t really explain why this would affect only one of the signal lines.
Oh well, time for more interesting tests than a simple loopback.