What lens design limits derive from the width of the lens mount?

I'm familiar with the lens design limitations imposed by the flange-to-focal-plane distance. But what constraints are caused by the width of the lens mount (that is, the size of the hole in the camera body)? Clearly there's the fact that if your lens mount needs electrical contacts, they have to fit in somehow and you need to make arrangements for a diaphragm if the lens needs one. But what other limitations are imposed? For example, does it limit the maximum aperture of a lens?

-
Do you mean the width of the ring itself, or the total width of the mount? – Håkon K. Olafsen Apr 21 '12 at 6:54
Clarified (I think). – James Youngman Apr 21 '12 at 20:05
Regarding aperture, the only thing that really drives the size of the maximum aperture is the front lens element, since relative aperture is based on entrance pupil, which is observed through the front of the lens. With the proper post-diaphragm element groups, you can control the size of the light cone pretty much as you please, so long as it maintains the proper magnification...so the size of the lens mount shouldn't really affect lens design limits. It might affect sensor design limits because of the image circle size. – jrista Apr 22 '12 at 13:45
@jrista: I think you've almost answered the question as I understand it: Suppose we have a sensor of diameter x and flange distance y, and we want to project onto the full sensor. Then what is the lower bound on the diameter of the light cone at its narrowest point, given that the last lens element has to be on the opposite side of the flange from the sensor? And of what is that Minimum Diameter a function? – feetwet Sep 4 '14 at 1:35
I don't know that there is necessarily a limit, not within reason. So long as the light "cone" is in focus by the time it reaches the sensor, that is all that matters. You could narrow the cone at the last element...or you could enlarge it. Whatever is necessary to support the entrance pupil of the mount. Problems ultimately arise when the flange distance is very small. In mirrorless cameras, it actually becomes a challenge to bend light the right way to reach the periphery of the sensor, and at an angle that is conducive to the photodiodes actually receiving light. A larger exit pupil... – jrista Sep 4 '14 at 18:12

A larger lens mount makes it easier to design compatible tilt/shift lenses.

-

The mount's throat diameter limits the exit pupil diameter. It also has strong control over vignetting, so it restricts the pursuit of ultra-large aperture lenses with acceptable light falloff.

-
To clarify: You mean that as the "mount width" (i.e., throat diameter) decreases then vignetting increases, right? Is there a formula that describes the point at which vignetting begins due to the occlusion of the exit pupil by the throat? – feetwet Jul 8 '15 at 18:11
There are formulas, but they include the distance from image plane to exit pupil and from exit pupil to mount as well as the "proper" size of the exit pupil in the design, which will itself be restricted by element diameters (an undersized element induces vignetting). "Hard cutoffs" such as baffles and hoods that are distant from the aperture stop lead to an abrupt vignetting, while closer cutoffs such as an undersized element near the stop cause more gradual vignetting. – Brandon Dube Jul 8 '15 at 23:52
Assuming a lens isn't designed to vignette, then every lens has a "proper" exit pupil given distance to sensor and size of sensor, right? Furthermore, in practice we can assume that consumers prefer their lenses a short as possible, so the last lens element will be placed approximately at the throat. Given these practical conditions: What happens if we begin to occlude that exit pupil at the throat? Gradual vignetting, or total occlusion of the image? (At this point it's only a function of (1) throat-to-image-plane distance and (2) ratio of throat area to sensor area, right?) – feetwet Jul 11 '15 at 19:31
@feetwet all lenses vignette. It's a valuable correction technique and is required for most even semi-large aperture lenses to work on bigger sensors. The exit pupil distance varies considerably model-to-model; some retrofocus wide-angles place it within 25mm or so of the sensor, other lenses are nearly telecentric and the exit pupil is hundreds of mm away. If you obstruct the aperture you will see hard clipping in the spot and it will be obvious in anything out of focus. It will increase vignetting at first and eventually show as a loss of contrast and obstruction. – Brandon Dube Jul 12 '15 at 0:25
Vignetting is not necessary, but it is a valuable correction technique when a lens has significant coma, petzval, and/lateral color. It clips the most poorly behaved rays that would contribute the most to the edge blur, providing a good boost to resolution. – Brandon Dube Jul 13 '15 at 12:36

With larger mount you can have larger exit pupil diameters. Ratio of entrance pupil and exit pupil is a measure for a lens symmetry. So with larger mount I suppose you can achieve more symmetric designs. The level of asymmetry affects among else depth of field (at given focal length, distance and aperture) and depth of focus. The look of out of focus highlights at very wide apertures is also impacted by size of the mount (and the chamber).

-