Food trucks are all the rage lately. Good food that finds you; what is not to like about that? But where do you sit? Trevor and Brad McAllister are brothers on a mission to answer that question. They aim to provide a unique product to this growing food service segment. Their design is a very tidy one: a collapsible, yet sturdy design with integrated trash bins and even a spot for an umbrella.
During the design phase, Trevor and Brad wanted more than just capability to visualize their product on a 2D workstation display, they wanted to hold a scale model in their hands. That is where Chatt*lab enters the story and I enter armed with Trimble Sketchup, MeshLab, and a Makerbot Replicator.
Over the past couple of months, I basically have had my arse handed to me on this one folks so pay attention. This was the most challenging and rewarding 3DP project I’ve experienced yet. I knew going in it was going to be a tough one but once I got the STL from the machinist I had to take a deep breath. Lessons follow.
Lesson 1: Quality source files
Work with the designer to get a quality STL or whatever you can obtain. Sure, there are fantastic apps like NetFabb Basic that will fix your models preventing ‘why did that happen’ moments off the printer. Suffice it to say there are plenty of CAD programs out there that will generate non-manifold models for you to struggle with trying to manipulate and print. Quality check these and see if you can improve the source data to get the vertex count down to under a billion. Seems like I had about a billion point 5 to contend with and just remodeled a number of parts.
Lesson 2: Print orientation multiplies your work (often)
There are somethings that just cannot be printed on a 3DP. It is something we need to talk about more often but overhangs and bridges can be a real challenge. Often times the slicer/firmware will just print into space. After all these are just robots that happen to spew out ABS/PLA, etc. and don’t care about your print as they are just following orders! Reorienting, disassembling parts for later assembly will cost you time, lots of time and headache and get you into trouble with fitment of parts. Guaranteed. You can actually bridge some huge spans with ABS but it isn’t solid once printed.
Lesson 3: Scaling and assembly multiplies your work
If you enjoy model building and gluing your fingers together with SuperGlue then by all means printing smaller components for assembly into a larger object may be right up your alley. For this model I pushed the limits by scaling the largest part I could print on the given 6x6x8″ (rough) volume I had available. This required me to split specific parts into two which also caused more work.
Lesson 4: Large prints fail (often)
This is a statistical fact. These machines are far from perfected consumer appliances at this point. Don’t get me wrong, they are wonderful machines and you’ll have to pry mine from my cold hands. I really pushed the boundaries of size with this assembly and large, complex structures will have a failure rate that will frustrate you. One issue was maintenance (lightweight) maintenance that was needed on my Rep1. The other issue I ran into was how to handle a 1kg spool of ABS. I build a few different spool holders and finally found one that worked reliably. I blew a few major prints in that time.
Lesson 5: Kapton tape is the devil
I’ve turned into a real pro at applying two pieces of 4-inch Kapton tape to cover the print bed. This stuff is lightweight and sticks to air and anything else with a positive charge including fingers, Makerbot parts, itself. While I’ve mastered handling this material, I am looking forward to printing more PLA on glass with my new Prusa. Kapton is used as it is just so happens that ABS sticks to it very well. The issue is when you do a large surface area, actually removing the part can be a real challenge without damaging the part itself.
Lesson 6: Trying something beyond your capabilities is the fastest way to learn
Prior to this challenge I had some experience in TinkerCad (does that count?!), Sketchup and Blender but these tools were maxed out by this project. I found the limits of Sketchup, especially with regard to planes and the ability to do complex shapes quickly. Also, Trimble, the new owners of Sketchup, need to work on their support of STL out of the box. I was able to get decent results but always had to pull an STL into NetFabb where numerous errors would be fixed. I’m very pleased with AutoDesk’s 123D Design, which I’ve used to build new things as of late. It has an amazing amount of power and while the GUI is unlike much you’ve seen prior, it just works and actually shares a great deal of ‘motif’ with their more advanced suite’s such as Inventor.
I certainly wish I could have turned this around faster for my ‘clients’ but this was an all volunteer effort and mostly dark thirty work given other priorities. I’m grateful for the experience.