Eddie
this reply has nothing to do with gearboxes, but all to do with flow potential through a port!
Jim's flow tests indicate that there is promise of a performance improvement for your engine with the valving and porting modifications he has carried out.
Another view of this is to examine the situation when the valve is lifted by 25% of the inner seat dianeter, which by and large can be considered to give a curtain area equivalent to that of the inner seat diameter, (referred hereafter as LD).
For the standard valve in conjunction with the PW3 cam, the duration through which the valve is lifted to the LD value is equivalent to the flow aperture being at optimum area for 31.25% of total duration.
Changing to a 3 mm larger valve, flow at the LD value is significantly reduced to only 15.45% of total duration.
Losses on swings outweighs gains on the roundabout (or vice-versa!).
As a matter of interest to some, of all the Commando cams I have data for, most have a LD flow potential with the 38 mm intake valve, of 30% or better, except one - the Norris RX which lags slightly at 29%.
To really maximise on the larger valve, a different cam would be required with increased duration above the LD value, but as Jim has stated, with the pushrod engine this becomes increasingly difficult to manage as a number of things begin to stack up against it. As it is, the increase in curtain area is sufficient to provide an increase in the flow potential, so the overall result is still coming out on the right side.
All will inevitably depend on how well you configure the rest of your engine to get what you want. If you are at Donnington for the Easter meeting, I will look forward to seeing you there, meantime let us all know via this forum the results of your testing.