I’m sorry I can’t really make this any simpler. There is no bug or other heat source causing it to have a negative slope. That’s how your furnace functions and if you looked at the picture from the manual you’d see that when controlled by a single stage stat it tries to use some of the modulating features like ramping up output the longer a heat call is made.
Another fun way to think of this is when it’s a nice 0C outside and your stat makes a heat call the furnace will click on and starts running. It is warm up so it clicks off after two minutes. In that two minutes it ran at 80% for a few seconds then went to 37% output as per the pretty picture. So for two mins it was mostly producing 20K btu.
Now when you have a really cold day and the furnace clicks on and runs for two hours you get very different behavior. First it clicks on goes to 80%, then 37%, then 58%, then 78% and finally 100% after twenty minutes. It then runs for the remaining hour and fourty minutes at a whopping 100% so let’s say it averaged 50K BTU over the whole cycle.
Now if you look closely you’ll notice 50K is a lot bigger than 20K. This means that when it is having short cycles and satisfying a heat call in less time it is actually heating the house at a slower rate because it never reaches the twenty minute mark and really opens up.
That’s why you get a negative slope. You are heating faster when it is cold outside like graph is showing. You’re free to disagree with me but that’s what is happening
Long cycles = Reaching 100% output and staying there
Short cycles = Heat call is satisfied before reaching 100% output
Cold days = longer cycles
Warmer days = shorter cycles