While testing the KIM Clone I/O Board I was using a real ASCII terminal and everything was working exactly as it should. After the board was in product a dedicated tool was written to run complete tests of the board and the serial port was no longer receiving data.
As it turns out, it is possible to set the various jumpers on the board so that the DCD line on the 6850 is floating, and it tends to float to a state where it ignores incoming data. My test jig simply loops back pins 2 and 3, leaving DCD unconnected.
The easiest solution is to pull the jumper off JP8, then place it between pins 6 and 7 on JP10. This forces the DCD to ground, which allows the receiver to work as expected.
Depending on your application, whatever is connected to the DB-9 might pull DCD to the active state which also allows the receiver to work properly.
Future revisions of the board will have a pull-down resistor on both DCD and CTS so unconnected inputs will default to the proper levels for normal operation of the ACIA.