[Raul]

The name of the second case that Joe J. mentioned is called a "Dutchman" pin. I talked about it last year extensively. I do not know the formula, nor am I inclined to figure it out right now - maybe some other time. I do know that it is a lot stronger than you might think by doing normal stress calculations or estimations. It has to do with the unique orientation of the pin in relation to the direction of applied forces (which is longitudinal to the pin).

If you ever try to abuse one of these joints and examine it, you will notice that it fails (but only after extraordinary forces) in a fashion that I would call "a wiping shear" or "skinning". It does not shear in the normal sense or break (unless you use a brittle material). It just deforms into an undescribeable shape. In fact, you can only get the pin to fail when the material surrounding the pin yields to allow some freeplay in the joint. And in some cases with free play, rather than breaking the pin, it gets pushed out of the joint. So you need a very tight fit for this type of pin.

The first case described by Joe would also work quite well. It would behave somewhat like a clutch bearing where the rollers in the bearing get wedged in one direction of rotation. But in this case it wedges in either direction.