This week, I spent way too much time on an all-paper linear stage; (hopefully) picked a project and started preliminary analysis; and looked at a neat self-helping mechanism.
For PUPS 3, we were assigned to pick a machine to build as our final project, and perform preliminary analyses to ballpark cutting force, required stiffness, and how error would be apportioned out over our machine.
I had a fair amount of trouble with this assignment – starting with the fact that I still hadn’t settled on which project I wanted to do! I started off my work by listing the project ideas I’d come up with and quickly triaging them in terms of suitability for the class; feasibility in light of the amount of time I have to devote to class this semester; likely cost; and (most importantly to me!) long-term usefulness. For each of the projects that made it through triage, I then completed Part 1 of the assignment (Problem Statement) for each. I then presented these ideas to my peer reviewers.
Pre-Review Design Notes: Summary of ideas, triage, and Problem Statement for each viable idea.
Out of the five ideas that made it through triage (measurement probe, CNC mill conversion, single-point threading stage, full lathe stage and Michelson interferometer), both my reviewers and I liked the CNC mill conversion idea the best. I took this idea, and began completing the rest of the problem set around it.
Post-Review Design Notes (all parts except FRDPARRC table)
- Clarification (post-class 2016-02-22): Because my machine’s structural loop length and section dimensions are already fixed, I approached the problem from the opposite direction – I made different assumptions about my loop length and beam section, and tested them to see what ballpark range of stiffnesses I can expect –> whether my performance specs are reasonable. I’ll be refining my understanding of machine stiffness as the project goes on.
Design Workbook (FRDPARRC Table, Preliminary Stiffness Analysis, Preliminary Error Analysis Spreadsheet. I’ve punted on the preliminary error analysis spreadsheet, but I do know how to do these – I’ll refine it later.)
Error Apportionment (spreadsheet by Alex Slocum)
I really like this project. Because the frame and bearings of the machine already exist, I won’t be able to change these variables as I iterate on the complete system’s design, and won’t have the opportunity to do a number of design tasks (selection of structural loop, design of frame dynamic performance, balancing cost between bearings/frame and other components when error budgeting). However, having the physical machine already exist brings up a new set of challenges. Because I can’t modify the mechanical performance of the frame or bearings, I need to have a much more detailed understanding of those components’ performance before I begin designing the motion systems – I’ll have to actually measure (and simulate) the stiffness of the machine in multiple axes, measure the static and dynamic friction in the bearings, and do at least a first-order characterization of geometric errors in the machine. It will also allow me to really focus on the design of the components I do have control over: the linear actuators, the sensors, the mounting components for the hardware, and the electronics to tie the system together. I’ll have a slightly different list of tasks than some projects will, and (likely) less machining work. However, I’ll be able to focus my energy on the design tasks I do have – and I’ll need to, as I’ll have a lot fewer handles to pull while trying to meet my functional requirements.
Going forward, I’m planning to complete the following steps:
- Perform basic geometric error characterization and stiffness measurement for machine to enable refinement of error budget; identify potential drive systems (Overlaps PUPS 4 and 5).
- Two ideas for stiffness measurement. 1) C-clamp with high-K spring and force cell in series to push vise relative to tool (or just use vise jaws!) 2) Cable and pulley system to pull table relative to spindle. In both cases, indicate from spindle to reference on table.
- Source potential drive technologies (ballscrew, leadscrew, cable drive) and think about layout of components (PUPS 5).
- Characterize bearing friction and stiffness (deflection of one bearing under a given load – if possible) to support individual axis detailed design (PUPS 6).
- Complete detailed design of actuator & support structure for MCM, plus electronics (PUPS 6 and 7)
- Test MCM (PUPS 8)
- Complete detailed design of rest of machine axes (PUPS 9)
- Assemble machine, including electronics (PUPS 10)
- Finalize control system & document (PUPS 11)
This week in lab, we were assigned to create a linear stage with scrounged materials. I set myself the challenge of making a stage entirely out of scrounged paper products (including paper tape) and glue.
The final product came out pretty well! I had planned to make the drivetrain out of paper as well, using a reverse-wrapped cordage paper rope, wrapped a few times around a drive drum (thanks, capstan equation!). However, I was running low on time, and when I found out we didn’t need to automate the stage, I put this part aside. I have the parts I need to do it, though (as well as reasonable confidence I can transmit enough force to move the stage – see design notes) – I’ll definitely try to do this later.
This week is a bit of a punt, but I still think it’s worth showing. As I was casting around today for a cool mechanical device to showcase, I opened my mailbox and found the following junk mail flyer, courtesy of MIT:
While it is unfortunately a piece of junk mail, it’s a cool pop-up piece of junk mail. Plus, it also has a self-helping bistable mechanism that helps it stay open. In its normal configuration, the paper on the bottom edge of the card keeps it folded up. However, as you fold it down and the two bottom-most “wings” pass the horizontal and start to point downward, the card seems to find a second lowest-energy state, and tries to close (weakly) in the other direction. It’s just enough force to keep the card stuck in the open state – all made for probably a fraction of a penny per unit.