Due to increasing the fan voltage, I calibrated the fan delta T contribution again. Now it is 0.45 C instead of 0.25 C. So the power into the fan seems be heating the air ultimately. Of course this extra bit of delta T is included in all output measurements, calibration and any excess events. It should probably be tested at several different, higher ambient temperatures, to be certain that the fan adds heat to hot air as much as to cold air.
Then I started to mess with heating up the fan... and after a few short experiments I worked out that any heating up of the fan requires ultimately heating the air going through it, and the same input power, minimum, as the delta T would normally require is required to raise the delta T at the downstream TC (if the air is mixed). So short of directly heating the outlet thermocouples there is no free lunch of delta T increasing due to heating the fan by some mysterious way.
I did shine a very bright incandescent flashlight at the outlet thermocouples with a possible 0.1 C increase.
And then I did a full 8-crossing (every 45 degrees) hot wire anemometer traverse, and also the vane anemometer capturing 100% of the airflow. More on that later.
.