MSP430F5438A rev G problem

We are observing some unusual behavior in the MSP430F5438A.

 Please refer to the attached scope screen shots.

Our firmware developer has limited the firmware to simply set P8 BIT 4 (RED TRACE) as an output and drive it with a continuous square wave out.  P3 BIT 6 (Yellow) is set to default floating input.

 In the execution of the program, the output we expect to be driven is behaving like a floating or high impedance pin.   The first screen shot (no load) shows the output as an attenuated square wave.  When we load the pin (second screen shot) with 100k to ground it has a loading effect we would not expect to see on a driven pin.  Also, the pin that is supposed to be floating is driven with the square wave and loading has no effect.

 These pins are not connected to anything else since they leave the board to connect up to a sensor which is disconnected during these observations.  They are designed to be SPI but our developer discovered the problem when he tried to implement SPI in the firmware.  They are also not even next to each other.

 We traced out the board and the pins are indeed what they are supposed to be.

 All of the 8 prototype boards do this.

 The TI Demo board does not do this.

A second port does the same thing.

 The Rev on the Demo is not the same as the prototype boards.  The boards are rev G.

Could this be observed if the layout had one or more GND pins to MSP open? VCC pins?

 Any thoughts?

Thank you.

NO LOAD

100k LOAD to ground