Playing with the voltage divider chain (changing absolute value of divider, types of resistors, mounting of resistors) one can find quite optimal arrangements in terms of low drift and startup behaviour
Christoph is going on with thess tests. Probably he will find a suitable solution in terms of drift and power dissipation by trial and error (divider after divider...).
At least, it seems conclusive that parting (large) resistances along the divider into several smaller value resisitors lessens the susceptibility to drift (maybe from lower voltage developing across the resistors? Voltage rating, however, is/was not exceeded in any case)
As Malte is going to JLAB for a period of one year he hands over the SADC cooling test and construction work to Tobias H. and Tom.