The similarity of the temperatures of the two panel headers is good since the panels should be behaving similarly and contributing an equal amount of heat to the overhead storage mass. Again, I am not sure why the lower headers suddenly increase in temperature during the middle of the night. A similar, though less extreme, drop in the upper headers occurs at around the same time, so perhaps the heat collected during the previous day doesn’t finish conducting through the storage tanks until the late night. However, this did not happen on the night of the 2^nd, so something else could be going on.