You'll need to calculate the overturning moment as each major piece of the weldment is added to the fixture for any position it might be turned to to select a gearbox with at least that capacity X some factor of safety to account for errors by the users. You should also include a separate brake if there is a chance of injuring the user if the gearbox or connection to the gearbox fails. In the past I used a bevel set to position the crank out of the way of the item and the user outside of harm's way - I added a heavy handle that provided greater leverage to resist back-driving in case the brake was not used. This would not be required with a worm drive, but worm drives have enough reduction that users aren't happy turning the crank.
Also, if there is any chance the fixture will be loaded with material 'backwards' in a way that increases the expected moment, even if you think the user will have to be stupid to do it, try to make the fixture tolerant of that excess or work harder to make that an unlikely event. In the aforementioned fixture the original designer offset the mount from the axis so there was zero moment and the unit could be turned by hand and remain in any position; this led to one day an assembler put the unit on backwards so instead of 0 in-lbf of moment it was, I think 1000 in-lbf of moment. As the unit was in the CG-high configuration, the slightest tilt caused it to suddenly invert, nearly injuring the assembler. That's when the gear box was added - to make sure even if the assembler was an idiot again that he could not be damaged. I don't recall if the unit, about $80k, was damaged or not, but I suspect it was. You would think that an item that cost as much as an upper middle class house cost at the time would earn some extra care.
In another turning fixture the designer was more clever and made incorrect installation more difficult. They did, however, get stupid and use sliding, self-retaining clamps that were then slid into position and their bolts tightened to hold the unit. Eventually a distracted assembler forgot to tighten the clamps and, as the unit was rotated, one set of clamps followed gravity and slid out of position. This also nearly injured the assembler and certainly did about $6-10k in damage to the unit.
Even more expensive errors have been made by careless thought about what might get missed and how to ensure that the obvious measures have been taken. Check
and click on View Report.
