Thanks MacieKN and Steve for chiming in.
The reasoning behind K. Gilmore point of view is that the lack of mechanical load in the orthodymanic drivers, contrary to dynamic drivers, would lead to a potential lack of control. The analogy that's often used is that dynamic drivers have their own "brake" built-in, as the magnet makes resistance. The orthodynamic drivers, being built differently and without a driving magnet attached to the ribbon, would then need to be properly damped to avoid getting out of their mecanichal operation range. This is what is referred to as "flapping" on magnepans' boards.
Steve Eddy's experience is interesting because he seems to hear no difference, even though the electrical damping is close to a ratio of 0.15. I wish I had the equipment to measure this too, I'm really wondering if distortion is going up, or if that's still within the mechanical range of the drivers (for the Audeze at least, but I see no reason to have a difference with HifiMan).
MAcieKN, HifiMan models have close to zero mechanical damping, and I think it's the same for Audeze's models. Yamaha and Fostex, however, do use that kind of damping. Hence my question on the applicability of the numbers from the Yamaha brochure to the Audeze and HifiMan models (open-back orthos).
Dynamic drivers are only indirectly controled my magnet (if amp's Z out is low enough), the resistance you speak of comes from driver suspension, and that is why some drivers are less sensitive to high impedance outputs- they have stiffer suspension that does most of the job of controlling the membrane. I'm not an expert but it's logical that ortho drivers do not simply sit loosely between magnets but are suspended just as regular dynamics, something holds them in place.
Btw there is no magnet attached to the membrane (or ribbon, whatever you call it) in dynamics, just the voice coil.
Even if we speak of an open- back ortho like Audeze there still is mechanical damping from the felt or similar stuff that sits behind (and, I guess, in front) of the driver. For example there is a mod for LCD- 3 which uses toilet paper (purrin's idea if I'm right): it simply restricts the airflow allowing less driver movement after it was excited by the signal.
As far as my basic knowledge of accoustics goes I can say that orthos are indifferent to output Z variations (and therefore high Z out) at least in FR terms. If they are a purely resistive load you could aim for current bridging. If Z out is equal to the load, then current transfer is most efficient.
Low Z out allows for more efficient voltage transfer, meaning more dB at your ears at a given gain, so if you used like 10kOhm Z out to drive 50 Ohm ortho headphone you would need a lot of a gain for satisfactory loudness, and most of the voltage would be wasted into heat.