Originally Posted by Billfred
May, but testing has shown a fan blowing on the motor after hard running makes a meaningful impact
. Maybe not enough to totally cool a CIM-based drivetrain between finals matches, but an 83% drop in cooling time for the air inside is significant in my book.
A 775-class drivetrain should be able to cool even faster, since it has less thermal mass, and I imagine big radiator fans cool a bit better than a 120mm KOP fan. (5402 used those in testing in 2017 mounted axially with respect to the CIMs, and they made a difference but didn't reach to the outside edges of the cans so there was definitely meat left on the bone.)
So I suppose the lesson is "let the fan run beyond the time the outer can is cooled to ambient", but how long is a matter for more instrumented testing (and then better data logging than Aren's post, though I appreciate him sharing what he has).
The testing described in the post linked does not show if the internal ambient air temperature rose again after the various forms of airflow were stopped. If the internal ambient temperature does not rise, the armature has truly been cooled. If the internal ambient does rise, the airflow has mainly flushed out the air that was heated.
I agree with Kevin's post, after the linked post, that the internal ambient temperature is not a good proxy for the armature temperature. From the many thermal tests I have had to do at several of my day jobs, I have found that a gap of 1-3 mm due to a thermocouple that detached from a heatsink will read between 5 and 20 degrees C lower than an adjacent one that is properly attached.