Here's the final CAD! Here are the details of how I got to there!
I decided to use MDF because it's very quick to lasercut, inexpensive, light to be able to lug around to various shops on campus, and readily available in my lab. I enjoy and have a lot of experience making 3D structures out of 2D components and machining processes.
Here, I design the main clamping structural area to handle the clamping forces. I cut it up into 90-degree chunks and modeled each beam as a rectangular channel.
And this is the final side plate. There are four of these in total.
Here I begin dimensioning the linear slide axes for the components I can find readily online based on the calculations and decision I made in the last PUPS.
Here I consider the moment about the tool axis from tool crash forces. Because angular compliance isn't an issue here, (it is in fact desired to handle the possible +-5deg misalignment that may occur from the fastening operation), I designed these axes to handle the stress, but not be very stiff.
And here are the twin axes, shown with the front plate removed.
Showing off some linear motion in CAD.
It's important to note that both clamps and both axes all have the same linear guide rail and leadscrew architecture, for ease and homogeneity of design and components.
A cutaway of the vertical tool axis and of the leader clamp axis (The follower only has a pulley, no shaft coupling or servo). Note the twin bushings on the motor side, which are preloaded by a shaft collar and a conical spring. The other end of the shaft has a single bushing to handle radial loads, but allows for expansion and other linear motion of the leadscrew to prevent overconstraint. The distances between the components is tight, and the stacking of the 6 parts of the carriage presents a risk of overconstraint. To mitigate this, carriages were all assembled while on the guide rails, and the bolts that hold them together were the last to be tightened, ensuring a fit that is smooth and not interfering.
Showing off clamp range of motion from a side cutaway view...
And a top cutaway view.
The specific dimensions of the clamps were designed here to ensure stiffness under gravitational load.
These blue balls (heh) are used to couple 5 DOFs of the FASBot to the airframe. The 6th DOF is constrained by the pair of clamps.
Here, the FASBot is shown in its natural environment... Note the three balls on the top of the FASBot. These will couple kinematically to the Parent Robot (not shown) using a magnet as preload. A hunk of steel will be bolted onto the hole that's in the center of stiffness of the three balls. The clamping force will be stronger than the magnet prleoad force, letting the Parent Robot to simply pull away from the FASBot once it's clamped to the wall. To pick it back up, it couples to the FASBot and the magnet engages with the steel to provide coupling preload. The FASBot then unclamps itself and is then attached to the Parent Robot, which can than remove it and place it at a new location inside the aircraft for further fastener installation.
Side view of the FASBot to Airframe coupling.
A closer look.
The FASBot-Airframe coupling procedure ideally places the nut driver tool right above the first fastener hole of the Shear Tie. The composite Shear Tie will have temporary tack fasteners at all sets of fasteners, and the adjacent set of fasteners will be installed, ensuring that the Shear Tie is held well against the Fuselage Skin despite the loading of the ~3-5lb FASBot.
As the tool goes down, it finds its target everytime. No need for heavy and expensive visual sensorheads!