Reference Voltage input on Nano54415

[dciliske]

You're likely running into an unpublished errata of the MCF5441X.... Here's the testing we did to uncover this last spring. We have repeatedly asked Freescale to update the errata regarding this, but (if had to guess) legal appears to not want to publish it.

In our testing, all
values are based around a simple resistor voltage divider as follows:

1.2k tied to 3.3V
3.3k tied to ground
Sample node between them connected to the ADC input

Now, the datasheet lists the ADC with a nominal 2k ohm input impedance.
However, when the DAC is powered down for the corresponding ADC
submodule, the ADC pins will read the sample node without obvious
loading. However, when the DAC is powered up, suddenly the sample node
behaves as if the ADC input is a 2k ohm resistor tied high, low, or
floating, depending on the input number.

The behaviour as I understand/observe it is as follows:
ADC0/ADC4: floating
ADC1/ADC5: tied to AVDD_ADC
ADC2/ADC6: tied to AVSS_ADC

My measurements are (approximately):
ADC0/ADC4: 2.40-2.42V
ADC1/ADC5: 2.68V
ADC2/ADC6: 1.80V

[dciliske]

Since you are now the fourth or fifth person I've seen get bit by this waiting on Freescale, we will be adding an addendum to the Freescale errata on our website in the product documentation for MCF5441X based devices.