Re: [avr-gcc-list] ADC anomaly

[Mattias Svensson]

Douglas Dotson wrote:
> I have a strange anomaly with the ADC on a MEga32.
> The ref voltage is 2.5, input voltage to channel 2
> is 1.25. ADC reading settles at 511 or so. After
> sitting with no activity for several hours, the ADC
> reading drops to 508 or so but the input voltage
> and ref voltages are stable at their proper values.
> Any ideas?

A difference of 3 LSB sounds a little to small to be an anomaly...
I do not have the Mega32 datashet at hand, but for Mega16 under "ADC 
Characteristics" it says that the "Absolute Accuracy" (INL, DNL, QE, GE 
and OE included) is typically 1.5 LSB.

The 3 LSB difference in your case could be temperature drift due to the 
device heating up during your several hours test.

/Mattias S