Thanks for your inputs regarding the choice of transistors. I decided to go for the BDY20 2U50 and I dont regret - it sounds very nice!! Highly recommendable for as an amp for daily indoor use. Depending on the input device (in my case a Focusrite Scarlett 2i2 (max. output voltage: 1.2V pk-pk @ 1kHz)) it's easily loud enough to disturb neighbours

The housing just contains the amplifier and a small RC input-filter with 1 Ohm and 10mF on each side. Very handy! A mid-size fanless CPU-chiller sinks the heat of both 15Ohm resistors:
Attachment:
MCA-JL.jpg
For the moment, I use my switching mode lab power supply and there is no hum at all. A CLC-supply (20mF-50mH-20mF) is close to be ready to use, too.
By driving the input at 28V, I burn almost 50W.
Playing with the input voltage, I only observe a clip in the positive half wave. Is this normal?
Because the MCA shows an amplification which is over a wide range independent on the input DC-driving voltage, I think I reduce it to say 15VDC, reset the bias and burn only a bit more than 10W all the time. Does this change the overall performance drastically (sure, I will shift the Q-point by some volts, but is this a 'serious' thing?)? My ears and my oscilloscope say it's fine as long as the input voltage is restricted by the preamp such as in my case to a relatively small value of 1.2V pk-pk. Unfortunately, I don't really have a detailled idea about the distortions and how to measure them..
In general, I'm curious how to calculate/estimate the damping factor, that will be provided by this circuit. Anyone an idea?
Thank you!
Do you have an image of the amp all closed up from the outside. Still using it?