Finally, I installed the spool on the top of the Fisher with a very basic setup, no bearing, no complex part.
I have made a simple spigot support on a wood shelf support bolted on the top.
To have the filament path as straight as possible, I moved the extruder and inclined it by 30°. I have drilled the top plate, somewhat in the PLA support. I had to cut the extruder panel to make room for the retaining tongue.
With that new path, it is quite impossible to guide manually the filament to the inlet, so I have reused the part cut from the bowden and printed a termination to be locked in the extruder. It shall be more simple to use a brass termination as set on the bowden inlet, but I had not one at hand.
The reclined position of the extruder have the advantage that you shall reduce the length of the bowden tube (on the setup, it is reduced from 460mm to 410mm) and the shape of the bowden is fairly less tortured than on the original setup. Looks really better at all effector positions.
The support is installed with two butterfly nuts, which allow its dismantling for settting the printer upside down for maintenance, or simply for transport.
The rubber band you see on the spigot is cut from a bike tire tube. It is used as a 'brake' to limit the spool free running and help maintain proper winding of your filament. Unwinding the spool needs a tension of approximately half of the spool weight. It is so not quite constant but it seems to work. However, if as me you have an aggressive retract, the spool will tend to balance. In the supplied file, I have designed a groove to install an O-ring 18x1, which will be more elegant than my tire tube segment. The spigot is reclined by 2° to have the spool staying near the shelf support.
1 spigot (on photo, it is 100 mm long, but i will propose also a 80 mm one)
1 tube inlet part
One template to drill the extruder panel.
Drill the holes for the support, their axis is approximately aligned with the side of the motor breating holes. I have glued the head of the bolts to help them remains locked while you unscrew butterfly nuts.
Inlet filament access: you shall drill a hole 2/ 2.5mm diameter first with the drill a little behind the top screw (toward the rods), as near as possible from the screw (more than on my setup). You will drill through the PLA support but at this location, it will have no effect on the structural integrity.
Drill first vertically before inclining the drill toward the motor shaft end. After having drilled the pilot hole, drill with a 5mm bit. To avoid any risk of acrylic crack, your bits shall be well sharpened and your drill torque clutch be set at low value. Go slowly and beware of the endstop cable. Do not hesitate to file if needed. Look the result from inside (printer upside down).
Use the template you printed to drill the new holes for the stepper and for the tongue. You shall file sowewhat the panel to make room for the tongue.
use the motor existing holes to attach the template (the two holes on top)
The bigger holes at the bottom gives the position of the tongue. One may be partially positioned on the existing panel opening. You will have to file the space between the holes.
Cut the bowden tube by 50mm minimum to have a new length of 410mm (400mm may be better, but 410 is what you see on the photos) and redo the brass installation - no fun.
Use the small bowden part with the end formerly in the brass part screwed into the printed end. set this small end into the inlet of the extruder and lock it while rotating. There is a small locking extension.
Reinstall you extruder.
And last, reload modified load/unload macro with your new bowden length.
Unexpectedly, the most challenging part of this setup was to print the small part for the filament inlet. It is so small and tall that it tend to melt, to be pushed away, and so on. I melted a lot of parts, pushed them away, broke one while separating from the support column and finally I melted my bit in the PLA. But I learned a few things in the process :)
The final part is shorter than the photo, to limit the bent at the extruder inlet.
So to print this tiny part - in PLA, no hope with PETG :
- First, your extrusion shall be calibrated (I find on my machine that with the out of the box setup -144 step/mm, I was overextruding by 6%).
- Layer thickness 0.2 mm
- Speed set to 20~30mm/s for perimeters and infill - do not slow for small perimeters, it tend to overextrude on the hole, and below 20mm/s is too slow.
- external perimeters first.
- sufficient Z-lift (I am at 0.6 mm)
- the extrusion coefficient may be reduced, but I cannot advised any value.
It is more work, but what is possible is to :
- move the endstop to the other panel (the holes are already there) then  redrill a top plate fixation hole. (hole of the screw which prevent straigth path) at 12mm from the existing (nearest to column)  and reinstall the top panel /corner screw. Drill the inlet path in the top plate and PLA, as near to the side panel as possible. This will straighten better the inlet path.
- Test shorter Bowden length (390~400 mm).