I've run into a very strange issue with the AtoD inputs, or more correctly, one of the AtoD inputs. I have a project where I am using both AN0 and AN1. On each pin I have an external 10.0K Pull up to 3.3V and then a series filter of two 1.0K resistors with 1.0uF ceramic to ground in the middle between the pull up and the CPU.
I am using the standard reference Coldfire lite project which does continuous conversions on all 8 inputs even though I only use AN0 and AN1. AN0 works perfectly correctly, but I have an input impedance issue on AN1. AN1 is behaving as though it has a pull up to VCC enabled internal to the part. If I ground the input, the AN1 pin sources 24.9uA through my T network, introducing error in my measurements. I plotted the input current on AN1 and the pin will source current into any input voltage below 3.3V.
Repeating the same test on AN1 I see very little current, only a 2 or 3 uA and as predicted the part sinks current if the input voltage is above mid scale and sources current if you are below the mid range.
Everything seems to point to AN1 having an internal pull up even in AtoD mode which according to the data sheet it doesn't have at all, regardless of mode. I've tested a few boards and they all behave the same way. AN0 is fine and AN1 has this odd pull up behavior.