20. Project Development

Objectives:

Complete your final project tracking your progress.

  • [x] What tasks have been completed and what tasks remain?
  • [x] What has worked and what hasn’t?
  • [x] What questions need to be resolved?
  • [x] What will happen when?
  • [x] What have you learned?

Tasks completed vs remaining

The PnP system envisioned consists of two main elements

  1. The XYZ motion system to carry the Nozzle Revolver between pickup zone and workpiece. This has been fully completed mechanically with belts pulleys and stepper motors and all required fittings. We choose to use a commercial Ramps1.4 controller for now to save time on development.
  2. Revolver head that actually contains all elements for Pick & Place on board has been completed in the mechanical build.

What remains

  1. Revolver head adding on board vacuum system
  2. Revolver head actuators for the rotational calibration of components
  3. In flight camera holder/fixture design and printing
  4. Wire up the designated controller based around the ESP32 dev board created during Week 14 - Networking with the Week 12 - Stepper Controllers
  5. Hookup to a full OpenPNP control system (out of scope)

What worked vs what not

What worked

  1. Make in generic/Versatile: Making a versatile XYZ frame fit for many (experimental purposes) worked really well. Especially when needing to tweak things i never had to use any fittings other than the standard 3 types we had.
  2. Build for general purpose:Also building the Gantry by design for more universal use proved a big win. Initially we planned to use the lower X beam setup for the rotator head we did not populate with the nozzles at that time yet. We received only a single Juki nozzle 2 weeks ago and it was clear I did not have enough headroom for the head now. We just shifted the X-beam to the higher position and BAM we we ready to go! I even have holes prepared for running the Nema stepper on the Gantry itself (like the XCarve machines)
  3. Build around what you have: Using T-slots 40x40 i had in stock instead of waiting for arival of any V-slots etc from China would have killed this project.
  4. Copy paste from existing design: Copying the design/structure/dimensions of our Stepcraft840 as much as possible, also saved a lot of time (where to put what questions) and this became very clear when we hooked up the Ramps1.4 with Marlin installed. All directions for homing and control were already fine by default. Only the Dual Y stepper setup had to be made as both of the Y beams are equipped with separate stepper motor.

What did NOT (yet)

  1. Axial rotation gear:The gear for axial rotation for calibration of components was a fail. We did not test in a small design for proof of concept. The type of gear choosen (Beveled Crown Gear) seemingly rotates only in one direction which is NOT what we need for calibration. Furthermore it is very sensitive to even the smallest deviation on any of the planet gears making the gear almost impossible to turn as soon as you fit the second planet gear in the enclosure.
  2. Vacuum delivery: Vacuum delivery to the inside of the revolver we envisioned initially to use a holle shaft Nema14 for the main rotational drive and passing vacuum to the internal rail. I could not get one such motors in time so we switched to adding a rotational gear on the back of the revolver driven by 2 CD-Rom motors from our scrap-yard. But now we had no hollow shaft left for vacuum delivery.
  3. Pen-stop vacuum set/reset : For activating and deactivating vacuum pressure on each nozzle we envisioned the Pen Stop switch. Designed and all but on the current design it had to fit in a 6mm stainless steel tube. It is simply to small to get it properly 3D printed in that space.

What questions needs to be resolved?

So looking at the list above it is clear what we have to resolve to get the machine working fully.

Most critical 2 are:

  1. The axial rotation gears need to be redesigned and 3D printed
  2. Pen-stop vacuum set/reset will probably need to be partly made from stainless steel combined with rubber seats and seals. This is impossible to achieve within the remaining time frame so we will just hook up an external vacuum pump for the simulation of its behaviour and move production to 2nd iteration
  3. Redesign the main rotator shaft to be hollow and pass vacuum through it and feed the rail vacuum through an air slip nozzle.

What will happen when?

Before submission closing we will hookup the Controller for the revolver head to at least proof the concept without the actial axial rotation.

In the 2nd iteration of the product we will redesign and re-print the main beveled gear this time an beveled gear with curved teeth to minimize backlash an give a smooth operation. It takes about 24 hours total to print whole gear assembly.

What have you learned?

Main things learned is to work from the beginning with parts you have. Most of the delays caused in this project was trying to design around parts still in delivery and parts you actually never seen in you live before eg. the Juki nozzles and the hollow shaft nema14.

Second is to create some kind of small version prototype of your project key component early in the process simulating and fit and see where the constraints or barriers may be hit. This was so true for the rotator head assembly. One needs to iterate fast on critical parts.

3D printing is very versatile but extremely slow! Think well before smaching a 20 hour print on the printer. That contradicts the “iterate fast” statement so reality is somewhere in the practical middle especially when sharing the printer with 8 other Fabbers.