Building the Gear Unit: DIY Kayak Motorization Project

English

Started detailed design using a friction wheel and pulley system.

Detailed Design

Realizing I needed precise measurements for clearances and exact positioning, I decided to draw up full-scale blueprints using Excel. Who knew you could actually do 2D CAD in Excel? There were online 3DCAD tools available, but they felt like overkill for this.

  • The light yellow base is an aluminum plate, 5mm thick and 40mm wide.
  • The blue parts are Ø34mm pulleys (which are actually blue in real life too).
  • The orange parts are Ø24mm pulleys (also orange in real life).
  • Note that circles marked with a yellow '5' are Ø5mm holes, 'M5T' means an M5 tapped hole, and drill holes are 4mm.
  • Green circles represent aluminum pipes in two different sizes (large and small).
  • For some reason, the blue pipe guide on the right is inverted. This error happened while going back and forth between Linux and Windows while trying to get things set up.
  • The 3mm-thick L-shaped angle bracket at the bottom was some scrap material lying around the house, meant for securing the pipe.
  • The gray areas (6mm inner diameter, 8mm outer diameter) are holes to seat 850 bearings. The machining method here is key. It's 6mm to provide clearance for the 5mm bearing inner diameter.

I proceeded with fabrication according to this plan, but ran into one snag. I forgot to draw in the pulley belt (O-ring) connecting the upper blue pulley and the orange pulley. As a result, the M5 screw pillar used to connect both sides of the middle base interferes with the pulley belt. Oh well, I'm sticking with two pillars for now.

This is a side view showing just the shafts and pulleys; the pillars aren't drawn.

Fabrication

Now for the fabrication phase.

Here is the concept for the gear unit. I probably drew accurate blueprints because I had a sudden flash of inspiration: 'Hey, wouldn't this work great?' Now, I'll start machining the base that sandwiches all this together.

I mark out the design on the material using a Permanent Marker and a compass needle before machining. This hole is a 10mm pilot hole for the 11mm hole used to fit the pipe between the two base plates. I didn't own a 10mm metal drill bit, and they're crazy expensive, so I used a cheap woodworking bit from a 100-yen shop. Since it's aluminum, it cuts through well enough. However, quality seems hit-or-miss; the 8mm bit—the one I needed most for the bearing seats—had poor cutting edges, which caused me trouble later. Among my various drill bits, my metal-cutting bits only go up to 6mm.

The deeper side of the back base and the small center hole are for the 850 bearing: Ø8mm with a depth of 2.5mm, plus a Ø8mm through-hole... wait, actually Ø6mm through-hole. You need a drill bit with a flat tip for the Ø8mm part, but those are ridiculously expensive.

Woodworking drill bits have pointed tips that make a small pilot hole first, guiding the rest of the bit in, and their cutting edge is flat. I took advantage of this: I drilled an accurate 2-3mm pilot hole, used that as a guide, and then used an 8mm woodworking drill bit to create the flat space for the bearing.

Here is what it looks like after drilling the pilot holes for the bearings and the 10mm hole for the 11mm section. I actually tested the method on a scrap 10mm aluminum plate lying around first. Confident it would work, I started on the real piece, but messed up the first one. The 8mm drill bit wandered, ruining it, so I had to start over.

For some reason, only the 8mm woodworking drill bit had a dull edge, and even filing it didn't help. After agonizing over whether to buy another set or just a single woodworking bit...

I found a two-step metal drill bit for 8mm holes that cuts in two stages, has a fairly flat outer profile, and comes in an extra-short type which is quite affordable. Since it's a high-grade tool, it cuts cleanly. Plus, its two-step design is practically made for bearings! Ultimately, since the hole needs to go all the way through, I still needed a longer drill process anyway. For the bearing holes (8mm + 6mm), since the thinner part of the drill is 4.4mm, I drilled a 4mm pilot hole, then used this bit to cut the 8mm outer circle, and finally drilled the 6mm through-hole.

The precision of these drilled holes determines how well-crafted the gear will be, so I worked with extreme nervous tension. Afterwards, I cut it with a hacksaw. Maybe my guard dropped, because the cut went in completely crooked... Well, precision isn't really critical right there, but still, it's pretty diagonal.

Machining the small reduction gear pulley. I drilled a set-screw hole sideways, but remember that aluminum is softer than steel—as soon as the drill bit bites, it tends to pull itself in, so be careful. It's not a set screw, just a regular 3mm screw.

The photo is a bit out of focus, but this shows the friction wheel section for the first-stage reduction: two thick G5 O-rings (O-ring names consist of a letter plus inner diameter in mm), a P45 O-ring, a sliding door roller, and a small sliding door roller. The pillars joining the base plates together use 5mm bolts and spacers.

Anyway, I reassembled the remake of the failed piece. It actually looks surprisingly good. Wait—the belt is interfering with a bolt! That was my reaction in this photo. I removed the bolt on the interfering side to keep going. There simply aren't any better positions. It looks rigid enough as it is, so I might just leave it like that. Or maybe keep that position and use it as a tensioner pulley? The paper underneath is a printout of the blueprint where I'm scribbling down various assembly notes and feedback.

Here I'm previewing the layout using another mix of failed and remade parts, connecting the input-side shaft and pulley. Hmm, regarding the interference with the pulley belt—if I tighten it loosely, the spacer might act as a tensioner on its own, but resistance looks like it would be too high, so I'll remove it. It would be nice if I could place a bearing in this spot, though it's starting to look pretty complicated.

Working on the second base plate for real. I machined the bolt holes, bearing holes, and pipe holes the same way. The bolt holes on this side are drilled to Ø4 and then tapped to Ø5.

For the large circle in the pipe section, I drilled multiple holes in a circle and then finished it off with a file. It's a technique I used to use a lot with cars.

Most of the parts machining is done. All that's left is putting on the pulley belt and screwing everything together.

Machining the Input-Side Pipe

The input side depends on the dimensions of the brushcutter: a 5mm shaft goes inside a 22mm outer diameter pipe (which has a 20mm inner diameter). This particular bearing is slightly oversized, but being a standard size, they regularly stock it at Handsman. It's a 635zz, priced at 340 yen each. The outer diameter is 19mm, but 20mm doesn't exist. Coincidentally, I found a cheap, usable alternative of the same size in the sliding door roller aisle... at just 260 yen for a pair, so I decided to give them a try. Turns out the smoothness is within an acceptable range. I wonder about durability, though? Since cheap parts are great for prototyping anyway, I might upgrade to actual bearings in the final version.

I bought a 1mm rubber sheet with double-sided tape to make up for the 1mm shortfall, but... it wouldn't fit. Even after stretching it, no go. Ah!! Since it's a 1mm difference in *diameter*, wrapping a 1mm layer adds 2mm in total! That won't work! They also sold a 0.5mm version without double-sided tape, but instead, I used some spare material I had lying around the house—what do you call it? I adjusted the size using some automotive gap-sealing tape.

This is it.

I figured out the pipe length by actually test-fitting it inside the brushcutter. Next up is making the mounting bracket to secure it.

This post is getting a bit long, so I'll save the rest for a separate article. Writing this stuff is exhausting and hard to progress with... man, it's tough. Doing actual workshop crafting is way more fun.
Next up is connecting the input section to the brushcutter, complete with video footage, so look forward to it!

Leave a Reply