> It seem to me that each doubling of the charge voltage would equal a 6dB increase in signal to noise ratio.
In acoustic-electric conversion sensitivity, yes.
If noise is limited by your tube, then higher bias overwhelms noise.
All condenser mikes require damping, resistance, usually acoustic resistance. This makes noise too. Changing the bias will increase signal but also damping-resistance noise.
> Why "only" 60V?
200V is common in some measurement mikes. Lower voltages than 60V have been used. I have seen a 5V-bias condenser mike for digital telephone use (not a hi-fi job, I grant).
> What is the maximum charge voltage of a Schoeps MK4 capsule?
Almost certainly the voltage it has, plus a margin for production variation. It may well be a 50% margin: changing to 90V would only increase electric output 3dB, and it is better to work every time all the time than to try for "3dB better" and risk total failure.
When you apply bias to a condenser mike, the diaphragm is sucked into the backplate. It resists because the diaphragm is under tension (or is stiff). But as it gets closer, the electric field strength increases, which sucks it harder, which increases the field strength, it sucks more..... There is a voltage at which the diaphragm will suddenly fall all the way onto the backplate. It is a basic and well-understood limit on condenser mike design.
You say, then, that you could increase the bias if you also increased the stiffness. But over much or most of its range, a condenser mike's sensitivity is proportional to inverse stiffness. Make it twice as stiff, get half as much mechanical motion and electrical output. You also change the frequency response upper limit. For maximum midrange output you want a less-stiff diaphragm, for maximum -3dB frequency you want a more-stiff diaphragm. Considering the fairly sharp ~20KHZ cut-off of the ear, the "best" stiffness is the value that puts the resonance between 6KHz and 30KHz depending what kind of mike you want.
I believe that for small acoustic levels, you can design to any voltage you want, if you change all other parameters to suit. But the optimum spacing for low voltage is too small to manage in production; anyway for many uses the spacing is limited by maximum acoustic levels. You could use a very large spacing with a very large voltage; that's silly, if you can meet your goals with a lesser voltage. I gather that the 30V-200V range we see happens to work well for normal levels and reasonable production tolerances.
You will find that Neumann et al were not lazy: the bias is high enough to make tube noise "small", but not excessively high. Higher bias would relatively decrease tube noise but leave damping noise unchanged, for little net improvement. Trading-off spacing and voltage will give a family of different mikes, each better at one thing at the expense of another.