7. Build & Test Entire Machine
Build & Test Machine
2016
This week I've completed the first build and assembly of John SteinBot right before the showcase! I spent the majority of the week machining my Y axis and bed. Then I began assembling and arrived at a working John SteinBot. I tested his performance and compared with predicted values. My measurements were within the error for the predicted values. I can say with confidence that John SteinBot is accurate to 1.5 mm over the entire workspace just like he was designed to be, Unfortunately, my team and I were too busy manufacturing to peer review this week.
"First, I made the easy parts like the motor mount, Y bearing holder, Y leadscrew nut, and updated bridge. Then I machined my Y bearing blocks from one piece of aluminum. However, I made a speed conversion mistake and set the CNC speeds way too high, so my finish was pretty bad. I also decided to get rid of the fillets on my clamps, which would have dire consequences later...
"The McMaster order arrived late and they refunded my shipping so I will be coming in under bugdet at $97! Nice. Next I machined a Y riser block because I had to make my X axis clear my Y bearing blocks. I also machined my leadscrew, but missed my press fit tolerance. I am trying to use super glue to artifically increase the diameter, but it is dubious. I also CNC machined my carriage, but made a zero-ing mistake which resulted in my holes being off center and having huge misalignment errors. Finally, I lasercut my bed. I also made supplementary features like wire guides and paper clamps from sheet metal.
"Then I began to assemble my machine. Quickly, I realized my bearing block fillet removals caused the bearing blocks clamps to break off when tightened. Ouch. I had to remachine the parts. Also, my rails didn't fit because of carriage misalignment due to a machining mistake. I had to make my own sleeve bearings from scrap because my old ones were destroyed. I noticed my machine was super loud at low speeds and concluded it had a low resonance frequency. Turning up the maximum speed solved the problem.
"After completing John SteinBot, I ran one last review to see if I had met my mission requirements. The cost was less than $100 at $97 and the workspace was big enough at 9.25" x 6.75". However, the final envelope was a little bit bigger and heavier than I wanted at 17" x 13.75" x 4.25" by 4 kg. The writing utensil requirement was satisfied with a fit of utensils up to 5/8" diameter. Safety was also good with nice wire management.
"Lastly, I tested the performance. The maximum speed of 25 mm/s was better than the requreiement of 20 mm/s. I measured the load capacity at 20 N which satisfied the 10 N requirement. The predicted resolution was 1.59 +/- 0.08 mm and the measured was 1.86 +/- 0.39 mm, within error. The measured accuracy across the workspace for the 10 N pen and 1 N pencil cases were within the 1.5 mm as predicted and designed for. Finally, the carriage yaw/pitch/roll errors were measured and found to be significantly less than originally calculated. My test is presented here for anyone to validate if they wish."
Its awesome seeing deterministic design at work. I designed for a certain performance and my measured results were consistent with my prediction (except for the large carriage erors... hmmm)! It feels good. Next week, I will talk about the future of John SteinBot...