After the first test, I changed the airflow direction. The air is now pushed into the outside holes and pulled out by the second blower. Working now to cool down the grid tube and the air stream with a Peltier element. Another option is to cool the grid tube outside with water cooling and also inside water cooling on the cathode must be possible. There is a lot off heat to exchange, due the low efficiencies on 23 cM. I continue to work to reduce thermal drift problem.