Turns out that shorting the unused inputs made the 1.25pins stable again and I was not able to get the noise to come back.

Seems like a rookie mistake. Always connect unused inputs per manufacturers datasheet recommendation - don't just let them float or you'll see oscillations.

 

Seems like a rookie mistake. Always connect unused inputs per manufacturers datasheet recommendation - don't just let them float or you'll see oscillations.

Haven't read the datasheet myself. But I've seen cases aplenty in which the manufacturer's don't clarify what to do with unused inputs. Sometimes they recommend grounding them, other times pulling them high, and other times leaving them floating. But many times they fail to mention what to do with them.

 

Seems like a rookie mistake. Always connect unused inputs per manufacturers datasheet recommendation - don't just let them float or you'll see oscillations.

Yes, this is defiantly a lesson I will remember. I just didn't expect such weird and wonderful things to happen though.

I am guessing that when I try to read an unused input is when it messes up, and the datasheet probably doesn't think the user would need to read in an input not connected to anything, so they don't mention it.

The design was so that the user can use as many or as few inputs as they need. But the micro tries to read all 8 (x2 ADC) inputs anyway, Discarding unused inputs in software.

Anyway, thank you everyone for taking their time to even read my problem and hopefully I can give back to this community in the future!

 

Yes, this is defiantly a lesson I will remember. I just didn't expect such weird and wonderful things to happen though.

I am guessing that when I try to read an unused input is when it messes up, and the datasheet probably doesn't think the user would need to read in an input not connected to anything, so they don't mention it.

The design was so that the user can use as many or as few inputs as they need. But the micro tries to read all 8 (x2 ADC) inputs anyway, Discarding unused inputs in software.

Anyway, thank you everyone for taking their time to even read my problem and hopefully I can give back to this community in the future!

Just to double check - at this point have you done the long, multi-hour runs like before and confirmed that not only is the 1.25V reference clean, but you've also eliminated your drift problem? It certainly sounds like you've found the key issue, but I was just wondering if you've got the good data now to confirm that there wasn't more than one underlying problem.

By the way, good detective work on the reference noise!