After working on mounting both the encoder and stepper to the existing banjo I realized it was too fragile. It has already undergone a repair via brazing in several locations.
I managed to get so far as to attempt running the encoder on the lathe but I was to impatient and broke the belt I joined. Using this tool and tire patching cement. The cement needs 1 to 3 days to cure and I gave it an hour.
At this point I realized I would be making a harp designed to handle the motor and encoder. But I had to decided how to make it. I ruled out fabricating from steel because I don't have a mill. My next choice was to cast it using cores to form the slots. Then I started thinking about 3D printing patterns and realized I could give a 3D printed banjo a try. The limitation here is the size of the printer bed, 8 inches or 200 mm square.
It took a while for me to figure out how to cut rounded corners in openSCAD. Maybe someday this will morph into a function but for now I only need to use it a few times so I will alter parameters as needed.
// Code fragment ot create a curved section to cut corners round.
// The object ccreated here must be subtracted from the oject needing rounding
// after positioning it properly with the first translate.
hubRadi = 56 / 2;
s = 100;
difference()
{
#rotate_extrude($fn=200)
translate([hubRadi,0,0])
square(20,true);
difference()
{
cube(s,true);
translate([s/4,s/4,0])
cube([s/2+1,s/2+1,s+1],true);
}
}
The 'cutter' generated by the code fragment.
It was applied to the boss in the lower center of the banjo template. The idea was to test the basic shape on the lathe. It was printed with the body .5 mm thick and the boss 5mm tall to save filament.
The next video was taken while testing
The next task is to correctly locate and construct the curved slot for the lower, and only, mounting bolt. I printed a button with a center that fit the banjo. This allowed the measurement of the distance if at a bit of an angle.
From the boss center to the inside edge measured 112mm, the outside 126nn.
(112+126)/2 = 119
The slot takes a 1/2" bolt. For the first test 12mm will be used.
I am going to start by generating a 12mm thick ring centered at 119mm from the hub.
Created a ring and used two objects like sector arms from the hub to trim the ends. Added a slot for the encoder. I will round the other corners and time excess past the curved slot later
// curved slot for mounting bolt
difference()
{
cylinder(100, 126, 126, true);
union ()
{
cylinder(110, 112, 112, true); // hole in cylinder to make ring
// two sector arms to close end of slot
rotate(a = 46, v =[0, 0, 1]) // set angle
translate([0,-10,-5])
cube([130,30,10], false);
rotate(a = 110, v =[0, 0, 1]) // set angle
translate([0,-10,-5])
cube([130,70,10], false);
}
}
// encoder slot
rotate(a = 20, v =[0, 0, 1]) // set angle
translate([-15,40,-10])
cube([11,110,20], false); // 11 mm slot for encoder
The printer was not advancing the Z axis enough it at all. Stopped the print after a few layers. It was enough to get the comparisons I wanted.
Used the button printed for measuring the hub to mounting slot distance to register the pattern with the existing banjo. This image is viewed from headstock looking back.
The curved mounting slot is where it needs to be or very close to it. Looks like it could be shorter on left providing more strength. Right is about right. The slot for the encoder need not have to be in a specific place. It will tighten with rotation about the hub or movement in the slot.
With luck the area between the hub and curved slot is enough to mount the stepper motor.
The encoder in its slot. The slot need to be a touch wider. Here we can see that the curved slot maybe a tish to close to the encoder.
As of now the current plans are to replace the controller on the printer. The new controller has stronger drivers. This may help even if the problem is mechanical. Not sure. But I have the stuff may as well give it a try. Next step would be new bearings on the Z axis but it seems to be moving easy enough. Thinking if the controller swap dowes not help I may temporialy counter weight the Z axis.
