I’ve watched a lot of videos on cam clamps, and I think I’ve only seen one where there was any analysis of the shape of the curve that’s needed for an effective clamp.  What I’ve been doing (and I don’t claim it’s the only shape that works) is basing the cam on an equiangular spiral.  With this kind of spiral, all radial lines intersect the curve at exactly the same angle.  If you choose an angle that’s close to 90 degrees, there’s enough friction between the cam and the clamped surface to hold it in place.  These cams are based on an angle of 85 degrees, which seems to work pretty well.  The force against the work isn’t perfectly perpendicular to the clamped edge, but it’s close enough that friction will keep the cam pretty secure.