Vmca as published can be very misleading
Vmca is a static speed. It is established in testing with asymmetric power/thrust set by incremental decreases in speed until one of the limiting parameters is met.
Because it is static and determined from a higher initial speed, it has some direct relevance to the Queenair video of slowly bleeding off the speed until LOC occurs. Well and good.
However, there is no element that accounts for sudden engine failure, slow rudder reaction or existing bank angle. The speed at which one of the limiting parameters will be met in the dynamic case will be much higher than the published static figure. There may be little margin, if any, between the target V2 and the dynamic Vmin if you have a sudden failure just after liftoff or at the start of a go-around. And in those cases, the priority is on control before climb.
The propeller related issues of loss of blown lift and discing exacerbate the dynamic issues substantially - I nearly lost a Conquest some 12-14 knots above the published Vmca during a practice engine failure in the GA, despite quite minimal reaction time as I saw the initiation begin. It seemed to take forever for the full rudder and full aileron to stop the roll and yaw, let alone bring it back to a sensible place.
JT: I seem to remember that, while certification standards limit the bank angle to "no greater than 5 degrees", the optimum is around 3 degrees.
And as for the good old V2 issue: it is a target speed from below (ie when accelerating), it is not the best anything (just the lowest speed at which the OEM can meet the minimum certification gradient) and, if you have already got it, V2+10 to V2+15 gives a much better gradient.
Stay Alive,