Design
For my linear motion module, I decided to make a slider. Find links to my full write-up and results, part drawings, my solid model and the equations / code used for my predictions here:
Having considered various options for my linear motion module and having played around with a sketch model I had made to evaluate its successes and failures, I decided to go with the concept of mounting two widely spaced Delrin sliders (the wide spacing was selected to avoid jamming) on a C-Channel carriage and sliding it along an 80/20 Aluminium Extrusion. The solid model is shown in Figure 1.
Having considered various options for my linear motion module and having played around with a sketch model I had made to evaluate its successes and failures, I decided to go with the concept of mounting two widely spaced Delrin sliders (the wide spacing was selected to avoid jamming) on a C-Channel carriage and sliding it along an 80/20 Aluminium Extrusion. The solid model is shown in Figure 1.
Figure 1: CAD model of linear motion slide.
The specific details I paid attention to in my design were:
- Use a bar with cylindrical ends with the maximum cylinder diameter that will fit in the groove with two contact lines and maximize the width of the neck such that it won’t jam into the edge of the groove but also will not bend too much.
- Apply St. Venant’s jamming principal (the axial spacing between the sliders should be 3-5 times the lateral distance between contacts) to ensure the carriage is long enough and mount the sliders at the end of the carriage.
- Use a C-channel that is inwardly tapered and have extra untapped depth in the holes on the slider which mounts it to the C-channel, such that the screws that mount them to the channel also push them into their groove.
- Use Delrin and coat the grooves in a little oil, for reduced friction motion.
- Connect all of the contact points on the sliders such that there are only two (a front one and a back one) and use the mill to match drill each slider and C-channel with precision so that they sit exactly where I want them.
Manufacturing
I fabricated my sliders by cutting them with a waterjet and then match-drilling them with the C-channel on a mill. I cut the extra holes for testing (eg. laser mounting and force application points aligned with the center of friction / center of stiffness) on the C-channel using the drill press. For the fixed end supports, I ended up finding L-brackets with appropriately sized holes for the optical bench, on which I would be carrying out my tests, in a bucket of scrap metal. The assembly of these parts is shown in Figure 2. Qualitatively, the slider slides very smoothly and it feels like there isn't much backlash.
Figure 2: Assembled linear motion slide mounted on an optical bench.
Testing
For testing, the wing nuts are used to constrain the four important degrees of freedom when mounting the laser pen. The wing nuts can be mounted in the orientations shown in Figure 5, determined by which test is being carried out. Screws can be attached at various points (the holes on the side are aligned with the height of the center of stiffness and the center of friction) such that a force can be applied using a spring scale (and some rope to connect two screws or apply a force along the axis of a screw where necessary).
Figure 3 shows the test setups with the laser employed for accuracy, repeatability and angular stiffness testing. For applying a force to measure yaw stiffness, the spring scale was clamped between two nuts and used to pull a screw mounted on one front side hole along its long axis. For applying a force to measure pitch, the spring scale was used to pull upwards on a rope attached to the screws mounted to the two front side holes. For applying a force to measure roll, a spring scale was used to pull upwards on a central side hole.
Figure 3 shows the test setups with the laser employed for accuracy, repeatability and angular stiffness testing. For applying a force to measure yaw stiffness, the spring scale was clamped between two nuts and used to pull a screw mounted on one front side hole along its long axis. For applying a force to measure pitch, the spring scale was used to pull upwards on a rope attached to the screws mounted to the two front side holes. For applying a force to measure roll, a spring scale was used to pull upwards on a central side hole.
Figure 3: Testing setups for this week’s testing. (a) Linear motion slide with the laser mounted for roll and yaw measurements. (b) Linear motion slide with the laser mounted for pitch and yaw measurements (the front fixed end support is missing from the picture).
For measuring straightness and radial stiffness, a dial indicator was clamped onto the top of the carriage, pointing both horizontally and vertically depending of the parameter of interest and set up next to an optical mounting rail to measure the deflection relative to. For measuring the static friction coefficient, a spring scale was used to test how much force had to be applied before the carriage could be moved along the rail.
Results and Analysis
Table 1 gives a breakdown of the measured results from testing and the predictions I made using the analysis and equations in my design spreadsheet. An appendix for the actual testing results is attached at the end of my full write-up document.
Table 1: Summary of predictions and testing results.
Conclusions
From my testing, I verified that all of my predictions were on an appropriate order of magnitude, but were not necessarily good predictions. In the future, I think a couple of things can be done to improve the validity of the tests that I ran, including: testing right after assembly such that no wear occurs to throw off the results; find a more reasonable method for predicting straightness, because averaging the tolerance does not always reflect reality and be sure of the surface roughness of the rail along which the dial indicator slides to test straightness.
My predictions in the pitch and the up / down direction were largely very good, possibly indicating that wear did not occur so much in this direction when playing with the carriage. Some wear should definitely have been managed by the preload, but if I did not preload one of the sliders well enough or was not precise enough in mounting the sliders to my carriage then this would have thrown off my results somewhat.
My predictions in the pitch and the up / down direction were largely very good, possibly indicating that wear did not occur so much in this direction when playing with the carriage. Some wear should definitely have been managed by the preload, but if I did not preload one of the sliders well enough or was not precise enough in mounting the sliders to my carriage then this would have thrown off my results somewhat.