Post 1: A Truly Pointless Mechanism

This is a writeup about a fun and truly pointless machine I've been working on. Inspired by the tool changers of CNC machines and (more recently 3D printers like the bondtech IDEX) this is a machine whose sole purpose is to automatically swap one screwdriver bit on my driver for another.

To be clear, if the title didn't give it away, this isn't a useful innovation. It's pointless, but that doesn't mean it can't be fun and maybe be a tool for myself and others to learn some cool things.

So with that said, at the outset of this project I'm setting some rules / goals for this design:

The design must swap one bit for another in under a second. This is quicker than I could do it when trying to change bits as quickly as possible with my hands.
The design must house all 64 bits that came in the iFixit driver kit.
The design can't make use of vacuum / compressed air. This is a practical concern as I don't own an air compressor.

So with these goals in mind, I set out to manifest this machine into existence. A first step for me was trying to break down a list of actions I needed the machine to make so that I could start building out the actual parts and assemblies. The actions to change a bit as done by me, a human, are as follows:

Grab the old bit
Remove the old bit from the driver
Place the old bit back into the driver case
Release the old bit
Select a new bit
Grab the new bit
Remove the bit from the driver case
Place the bit into the driver case
Release the new bit

Written out like that it does sound like a lot, at least to me, the person currently working on cramming all these actions into a span of time shorter than a second. But I'm going to do some optimization now to compactify this set of actions. That way we can get all these things done in approximately the same amount of time that it takes to sneeze or takes Jeff Bezos to earn about $3,750 as of the time of writing this post (he must really be working hard.)

Instead of doing these actions in a nice neat ordered list, we're going to squash pairs of them together to single moments in time.

T0: Detect presence of driver combined with user selection of a new driver bit (1 second timer starts now)
T1: Grab new and old bit
T2: Remove old bit from driver and new bit from storage
T3: Place new bit in driver and old bit in storage
T4: Release both bits

By letting some actions happen at the same time I can cut down on how long the bit change takes. And with the individual actions separated into steps in a process, I can start designing the mechanism that will be responsible for making this all happen.

Mechanism 1: Dual bit grabber — a pair of grabbing mechanisms to exchange both the new and old bit.
Mechanism 2: Take / place the bits into the driver and storage.
Mechanism 3: Storage for all 64 bits.

A little animation to better illustrate the idea:

I didn't come up with the concept for the swapping mechanism on my own. It's largely modeled on the mechanism from a DMG Mori CNC mill, just scaled wayyyy down.

DMG Mori Tool Swap Reference

Now that all the mechanisms are defined I started putting them together into a version 1 design. And wow did it have a lot of problems. Here's a look at the version 1 design. I was calling it version 1 which currently feels much to large a number for the level of refinement and much to small a number for the amount of concepts and failed attempts that preceeded this stage. So instead I'm going to call it version bad.0. Let me briefly talk about some of the things that make this design generally crappy. The bit grabber is huge, heavy and bad at it's one job of grabbing the bits. Two non-parallel jaws coming together is a bad way to precisely and forcefully grab a tiny object like a screw driver bit. For smaller bits no amount of tweaking the jaw design itself could grab the smaller bits used for M3 and below screws. Which is a problem because I want this to work with all 64 bits including the SIM bit and the 0.6mm hex and torx drives. I also realized that the linear axis that was responsible for moving and then replacing bits into the screwdriver / carosal had some additional design constraints I wasn't aware of. My first attempt using a pre-built leadscrew driven stage I had on hand moved at a sluggish pace but also lacked the thrust needed to pull bits out of the screw driver. Once I switched to a belt driven axis with a larger stepper the output speed and thrust could at least pull larger bits out of the driver. So with some initial testing done some constraints set I set about designing a version 2 that could fix some of these issues and as these things tend to go introduce even more problems. I started with the linear axis having burned out a stepper motor during my initial testing due to not configuring a drivers current output. I had harvested the motors from my old ender 3 but with one already acting strange and another completely dead I had a good excuse to upgrade. The ender 3 and most filament printers use a stand NEMA 28 stepper motor but for the low cost of a taller body a NEMA 28-34 motor offers more torque and speed in the same XY footprint. With this swapped in, I turned to the bit grabber. Not fully understanding what a saga this would be. My first goal was to try to shrink the mechanism down and upgrade it so that a single servo motor could drive each gripper bringing the total number of servos on the bit changing mechanism from 4 to 2. The first mechanism I tried for this was actuated by a scroll gear and functioned as a much smaller version of a lathe chuck. This did work but brought with it some new problems. The mechanism ending up being quiet fiddly and required the gears to be printed at orientations that resulted in the fear teeth constantly shearing off.