Category: Software

The angle block joins the aluminum grip with the plug sticking into HQ Sixteen’s handlebar control base:

Because I don’t know the exact angle until Mary puts more hours on the machine, the OpenSCAD code can tilt the plug from 10° to 30° with respect to the original grip. The bent part of the model consists of a succession of hulls around adjacent slices:

An overall hull() then gloms everything into one solid lump, with all the negative features removed from it:

After a brief flirtation with heat-staked brass inserts, four setscrews threaded into steel square nuts secure the original grip in the bottom:

A screw behind that big washer pulled the nuts firmly into their sockets, where they stay without any adhesive. The square recesses include a little adder based on the curvature of the hole to sink the nuts deep enough:

I made the block’s OD large enough to accommodate the brass inserts and hope it’s chunky enough to withstand the force from the setscrews. The inserts tended to creep outward after being snugged down, but the square nuts seem stable against the recesses.

The block prints with the top surface against the platform to produce a clean recess for the plug, which requires support material for the ring around the bore. Because the ring sags slightly against the support, the model makes the recess 0.4 mm deeper, but the next iteration gets a little more:

Not that it makes much difference.

The bore from the grip meets the bore from the plug in a sphere centered at the bottom of the plug recess:

A ball joint seems the best way to join a pair of intersecting cylinders, if you have room for the sphere, and eliminates a whole bunch of computations figuring the cylinder lengths; they just meet at about the center of the sphere and you’re done without anything sticking out. I’d like to pretend that was the first idea I had, but …

The OpenSCAD code can add more length to the bottom of the block, in the event Mary wants the grips lower:

That obviously increases the lever arm applied to the plug, but we’ll burn that bridge when we come to it.

This lineup shows the progression from the first pass to something that might actually work:

Rapid prototyping FTW!

The OpenSCAD source code as a GitHub Gist:

The grip plug is the upper part of the assembly changing the angle of the HQ Sixteen’s front grips:

It sticks into the machine’s handlebar control base in place of the original aluminum tube grips:

Only the two lower setscrews held the original grip in place, but I made the plug tall enough to engage all four, which meant it needed a side port to ease the ribbon cable through on its way to the PCB inside:

The plastic tube is obviously thicker than the aluminum tube, with four dents capturing the 10-32 setscrews (using a 3/32 inch wrench) to align it within the hole; the bore juuust passes the original connector.

The plug glues into the angle block, which means all the stress from the grips passes through a thin ring of plastic just above the joint. So I added five 1.2 mm OD hard steel wires about 20 mm long:

Five wires, because four didn’t seem like quite enough and six seemed like crowding too much steel into too little plastic. The holes are offset to avoid the setscrew dents, with one lined up directly under the cable port.

A pair of alignment marks help get the orientation right while gluing:

The control base angles away from the grip, leaving a little more than half unsupported:

Pondering that picture suggested adding those steel wires.

The angle block prints with its upper surface against the MK4’s platform to get good dimensions inside the recess for the plug, so I can’t add a wedge to that surface, nor can it go on the plug. Maybe a separate wedge glued around the plug?

However, the pin header for that cable sits directly inside the base and a transparent cover (not shown here) extends outward over the casting against the grip:

So maybe it’s like that and that’s the way it is.

The OpenSCAD source code as a GitHub Gist:

With the handlebar assembly angled to let Mary see the LCD panel, the grips no longer meet her hands at the proper angle:

Each grip has two buttons intended for thumb operation, but at that angle her thumbs lack oomph.

So I added a compensating angle just under the handlebar control assembly:

Restoring the front part to vertical means she can walk up to the machine, grab the grips at a neutral wrist angle, and start sewing.

Which required several iterations:

The pictures show various setups as we installed, tried, tweaked, and replaced nearly everything along that progression from left to right. They’re similar, but the details made all the difference.

This is an overview of the adapter, with details to follow over the next few days.

The solid model shows how the pieces go together:

The white chip on top fills the space between the surface of the base and the top of the plug, with some lettering just for pretty.

The greenish plug (not its real color!) sticks into the handlebar control base, where its dimples capture four setscrews. The original grips extended only halfway into the base, leaving the top pair of tapped setscrew holes empty:

The ribbon cable carries signals from the pushbuttons into the base assembly:

Looks like a guillotine to me, too, but the foam rubber cover prevents the grips from sliding any further into the base:

Despite the metric socket head cap screws used elsewhere on the machine, those are 10-32 setscrews. Took me a while to figure that out, as 10-32 setscrews are visually indistinguishable from M5 setscrews, but neither screw will thread into the other’s nuts even though their wrenches are equally sloppy fits in the other screw.

The angle adapter block has an intricate geometry:

Because I don’t know the proper angle, the OpenSCAD model includes enough trig to adjust from 10° to 30°, with the default at 20° to set the front of the grips vertical. The lower part of the block can extend to lower the grips if that turns out to be necessary, but we’ll start with zero millimeters.

The grips slide into the bottom of the angle block where they’re captured by four M4 setscrews threaded through square nuts:

The big washer over on the right sits under the screw I used to pull the nuts into their recesses, where they sit firmly without adhesive.

The first iterations used heat-staked brass inserts that didn’t provide enough griptivity against the torque generated by shoving the grips sideways. I probably applied more force than they’ll ever see in real life, but I’m no Hulk and I didn’t like the feel.

The upper plug gets glued into the lower angle with JB PlasticBonder urethane adhesive. Had they been finished, the first two iterations would have had screws through the angle block into brass inserts in the plug, but I realized adhesives would work much better. A pair of index marks aligns the two pieces:

It’s early days, but the machine fits her much better than it did before.

The OpenSCAD source code as a GitHub Gist: