Data about HP Agilent Keysight 3458A
These meters are specified for operation in up to 55°C ambient temperatures(!) So in June 23 I disabled the cooling fan in the Keysight meter and watched it like a hawk. Normally a high internal temperature (CAL? 175) is said to be the main cause of drift and therefore bad.
Not for my Keysight box, as you can see its green trace is a bit more bumpy, showing a larger dependence on day-to-day ambient temperature, but there is no long-term tendency. My old HP box did not start out as solid as this.
Surprisingly the magic "U180" steering hybrid at the core of all 3458s is not hermetically sealed, it only has a lid epoxied on. So I thought that humidity might be a cause of drift and so a high internal temperature could be a cure for my slowly drifting HP meter.
On Nov 23 I disabled HPs cooling fan too and was delighted to see an immediate downwards jump, reversing months of long-term-drift. I was also delighted by perfect silence and no dust inhalation for a while. Unfortunately it started climbing again soon after.
So I had to let it cool itself again and what do you know? It has been perfectly steady ever since. Maybe the few months of heat did something good after all?