I think we're on the same page. To me, the larger scale items seem interesting from a perceived frequency balance standpoint. I.e., the wide depression from roughly 200 to 1500 Hz, followed by the peak from 1500 to 3500 Hz (frequencies open to interpretation depending on how you look at them). Their magnitude isn't huge, but combined with their width they're likely making the speaker sound detailed, but lacking warmth/guts in the lower vocal range.
Do you have a means of measuring electrical impedance and phase? If you do, that data can be used in typical crossover simulators if you want to model the low voltage side of the transformer interface. Combined with the acoustic measurements, that should get you to a usable simulation without being able to fully model the transformer and high voltage side. At least I think it will. I've done some basic sims like that, but haven't done enough real work with them to guarantee it.
Sorry it took so long to reply, I completely forgot I was going to look into this. It appears that the damping screen is 200 mesh (200 line/inch). Here's a picture, it's a little hard to see because the millimeter rule is out of focus and I'm taking the picture through one eyepiece of my stereo-microscope. 200 mesh is nearly 8 lines per mm.
Sheldon
Sheldon