It turns out that some of the SCW oscillators on the the model samples are inclined to cause resonance because they are rather pure or something like that.
I had this problem whenever I used a D#3 D#4 D#5 and possibly upwards, lower octaves don’t create the problem.
I turns out that the steel pipe of my speaker stand has a hole to hide the speaker-cable at exactly 55cm which corresponds with the wavelength of D#5.
Plugging the hole helps a bit, detuning the pitch helps a bit but I really have to find a way to non-destructively detune the pipe, would filling with a light material help ?