There's something not right here.
I haven't understood how the second (front) belt is tensioned – there seems to be some sort of screw under the edge of the intermediate pulley.
On my RF25 (Alpine Engineering badged) the intermediate pulley stack is an idler mounted on an arm swinging around the centre column. For all combinations of belt and pulley positions the movement of the motor is sufficient to tension the pulleys. It gets a bit close the the limits of the guard in some combinations, but it does work provided the belts are the intended length.
Swinging the motor tensions both belts. Is that the case with this machine?
I wonder if the second belt (on the spindle) is a bit too long – using up more than its fair share of free travel?
Ostensibly it looks as though the machine has been fitted with the wrong pulleys from birth – but that statement has other ramifications.
Would you like to measure the OD of each sheave on the three cones, and also if possible identify the length of the belts as fitted. Oh, and the nominal speed of the motor.
Then a photo of the label showing the nominal spindle speed for each belt position please.
I think I've got a programmable calculator program to calculate the PCD length of the belt path for a given shaft spacing and sheave size somewhere, if not there's sure to be one on the 'net. We can then model the dimensions of the drive and see what's going on.
An intriguing riddle, but it's only geometry.
Best rgds Simon