Ok,

Here's some more.

I have another hunch that the harshness may have to do with the 12k peak. For this purpose I made seperate FFTs of a 12.5kHz sine and ones with both a 12.5kHz and a 2kHz tone.

Unit1 @12.5kHz:
8

The above gives rise to some elbrows. There is actually a new peak at 1/2 the fundamental, that is also echoed at the other side of the 12.5kHz. This one is not there in unit2, so is not related to the sound of -both- units. This must be some side effect of my imperfected centering capabilities.

Unit2:
9

In the dual tone test this doesn't show up, I think because both tones are much lower in amplitude individually.

Unit1:
10

Unit2:
11

No real surprises here I'd say. With a lot of midrange in the music I can get a little crowded around the 12.5Khz though...

Well, I'm still waiting for the new GPA diaphragms. In the meanwhile I will do an impedance measurement. I might be that the highish output impedance of my favourite tube amps doesn't match well with the impedance peak likely to be associated with the 12.5KHz peak (amps increase output with rising impedance). Favourable for LF response, yet with an allready firm 12kHz bump, it can be too much on the HF.

Thanks all for the support. The story continues...

Coen