I solved the problem on the slow response (for anyone who has the same problem)
The initial problem was the 350 ms delay caused by the MEAS:VOLT? and MEAS:CURR? commands.
According the documentation you can also use this commands without the question mark, so no answer expected from the PSU.
In the mean while question and set the other channels. and when you are back at this channel you can query the MEASure result with the FETC? and FETC:CUR? commands.
With this trick I get an update rate of about 0.5 - 0.7 seconds.
If I have time will release the C# code on GitHub (after I cleaned the code from all the prototyping)
Benno