Your Waifu is Still Trash: The Buildening

This will be a relatively short post because it turns out there wasn’t that much on this cursed postmodern robot to fabricate. To be fair, that was the intention. I didn’t want to spend a bunch of time manually drilling and tapping and fabricating, so this bot really was a paragon of my latest preference for unrefined laziness approaches to have the machinery do the work for you, keeping fabricated parts to a minimum.

We start with some 3mm HTD pulleys, namely the motor pulley and the main drive pod input pulley. These two together determine the motor reduction ratio. The motor pulley has a D-flat section designed into the center such that it can be directly pressed onto the motor shaft with a matching flat. I’ll make a little aluminum donut hub that presses into the big pulley.

For making the intermediate shaft gear and pulley assemblies, I designed up small drilling jigs that allowed me to just bump the parts against the jig stops and pull the lever. These are all basically fancy V-blocks. It doesn’t take much pressure at all to pilot and then drill the hole for the 2mm cross-pin.

The shaft material I used is just precision-ground mild steel, so no fancy carbide drilling through case-hardened shafts here. It’s not nearly a hardcore enough application!

The story was the same for the output shafts. These shafts were turned from 12L14 hex bar stock.

12L14, for when nothing at all matters specification-wise! I don’t know how much I trust steel that machines like brass in regular applications, but for this thing? Anything will do!

As a result of the little jigs, all of the machining for this bot was actually finished in one day and one session. I just cut the hex bar stocks, pounded out the driveshafts, then made a few one-operation cut shafts.

The bearings are simply pressed into the Mystery Micro-Center Something-filled Nylon drive pod prints, and little shims will keep the gears centered in the bearings. Something, because I’m not sure I believe it’s carbon fiber.

Here’s an assembled gear and belt drive ready to be stuffed into the casing!

The aluminum big pulley hubs were made by shaving a few thousands of an inch off a 0.75″ round bar, drilling and sectioning them, and then reaming the bore out to an 8mm sliding fit with a reamer I bought for such things a long time ago.

The printed-in hole in the pulley was actually enough to guide-start the drill bit, so I didn’t even use the printed jig. The hole is tapped for #1/4-20 and I’ll be using a flat-bottom set screw in conjunction with a small hand-ground dimple in the input shaft to hold it.

As I’ve said for a very long time: Set screws don’t suck, nobody uses them right. They have to be sized to hold the amount of torque you think you’ll be shoving through them, and they must sit on a dimple or a flat. The usual small power transmission part for an 8mm shaft will probably come with a teeny M3 or #4-40 set screw which is wholly inadequate for anything I’d want to do with one.

The next day, I fully assembled the drive pods. The motor mount idea was spot on and easy to install and adjust! These are some very smooth and quiet drives, even before being run in for 10 minutes apiece.

There’s only one “new design” motor clamp and one of the older design. I decided to not reprint the other side, since the old design Didn’t Not Work.

While all this was going on, the Mark 1.75 was making short work of the questionable spool of nylon I still had left over, turning it into the electrical system mounting brackets. Truly a testament to Markforged never leaving a stone unturned – their engineers have probably spent thousands of hours tuning and calibrating every aspect of the print behavior because they know irreverent sadists like me will avoid the company store and somehow find the cheapest, sketchiest nylon available to feed the machine with and complain when it doesn’t work.

And it will perform just fine.

An evening of populating the electrical components later and I’m left at the longest, most tedious process: wiring. A lot of the wiring from the last build was straight up unusable, whether from corrosion (remember, vape juice absorbs water from the air and vape clouds coat everything) or mismatched connectors. I basically started from scratch here. No reason to keep anything from the last build except the capital equipment.

Topologically, though, the power distribution design remained unchanged. The main battery power coming out of the vintage very-hard-to-toggle switch is split three ways. One for the RageBridge2 drive ESC, two towards the Pololu R/C relay which lets the power through to the vape and air pump power supply, and three a standalone BEC/power supply for the R/C receiver.

I made one change to the downstream power handling. Because of how much power I’m planning on putting into the vape cannons, I decided to move that power supply to after the R/C relay. The vape coils and the air pump will be permanently connected to the output, and I’ll be switching the power input on and off. This contrasts with the first build where the power supply was on at all times, and the relay handled current directly to the vape cannons.

It’s better for a power supply to “wake up with its load”, so to speak, than to have it sitting around twirling thumbs and then suddenly have the load applied and then removed. Each edge presents a transient voltage and current that could have other ripple effects on the system.

At the end of the night, the thing could actually power up and spin. Now I just need a dumpster to put it all in!

And to actually build the vape cannons. This is a funny little story that will get its own episode!

Bots away!