At an abstract level, every new project is composed of three elements: The known knowns (the things that you know you know), the known unknowns (the things you know that you don’t know), and the unknown unknowns (the things you don’t know that you don’t know)*. Stay far far away from the known gnomes. Given any opportunity, they will steal a single sock from your hamper.
The Known Knowns
After initial brainstorming I begin 3D modeling, starting with the pre-selected commercial off the shelf (COTS) components. These are the agreed upon components that will be bought, not designed. Connectors, motors, switches–all that McMaster-Carry goodness.
You can, with a few potential exceptions, consider these items as a fixed and unchanging piece of your larger overall design. Much like an unpaired, sweaty dress sock needed as part of an incantation for gnome dark magic.
Gather any information you can, including drawings, 3D models, data sheets, and photos. However, get the physical COTS items in hand, if possible, as they are the ultimate truth–Drawings and 3D models are idealized representations that may not reflect some actual details that are, for example, by-products of the manufacturing processes (e.g. flash, injector pin dimples, warp).
Sometimes, access to the real items is difficult due to cost, limited supply, or shared need. However, politely requesting a sample or loaner item from the manufacturer can help build your own stash.
I yearn for the utopian society where all CAD models are open, accurate, and accessible, freeing designers up to do the really important work of creating awesome, unique designs that add value for others. Until then, we’ll do what we can to get that data in hand. To start, here are four options to help you along your way.
Approach 1: Request a 3D Model From the Vendor
The best option is to get a 3D model directly from the vendor. Some vendors will make this easy and provide models directly through their webpage. If models aren’t directly offered, e-mail the technical department and ask if they would be willing to send one. I have a stock e-mail I send that I have tweaked over time.
Subject: 3D model request
Hello, my name is Dan Slaski with Renegade Prototyping. We are an OEM planning to incorporate your [brief item description] into our new [brief assembly description]. I need to design custom mounts for your [brief description of item]. If you have a 3D model available, it makes my life much easier. Otherwise, I have to do my best to reverse engineer the item whcih is very time consuming. A SOLIDWORKS (.sldprt) file is preferred but an .igs or .step file is also fine.
Thank you in advance.
I like to keep it brief at first as it may go to a non-technical person before it gets to the right department.
The reason companies may hesitate is because they are concerned about protecting their intellectual property. If you think there is something proprietary about the item, add a line saying: I understand you may want to protect your intellectual property, therefore, a model of the exterior surfaces or simplified model would be adequate. Coming from the vendor is the best option:
1. When you are hyper “selfishly” protective of your time, i.e. efficient, you can focus on higher value tasks and it’s better for everyone.
2. It “should” be the most accurate.
3. Too much detail can’t hurt but too little can.
I always measure critical dimensions on the actual part and compare. I have received models from vendors that ended up being wrong in some important way. Vendors are fallible and keeping up with the configuration management of the CAD models they provide is likely not at the top of their priorities. Sharing models is a smart thing to do from a business perspective. Given the option of two equivalent parts, I will choose the one with an available CAD model available for selfish reasons and recommend that one to our purchasing department. Overtime, I have developed a database of go to companies that I know make high quality products (and have available CAD). I believe there is a correlation between companies that provide CAD models and make high quality products as it is symptomatic of a high level of attention to detail at all levels. If they come back with a no, politely ask for a 2D drawing. These models can be a design placeholder but be very cautious with the provided models! As distrusting as you would be of a gnome that knocks at your door who says their car broke down out front and is wondering if they could use your phone.
Approach 2: Seek a Third Party
Online, there are user supported 3D model directories–3D Content Central, Thingiverse, and GrabCAD to name a few. These sites are the wild west. There are some rating systems but no real checks or accountability. The modelers’ intent is unknown, and the model may have been made for show or practice, and may not contain the detail or accuracy in the areas important to your project. All that aside, it’s worth checking to see if some rockstar CAD ninja like yourself did grind out the model you need and paid it forward by giving your to do list a big ol’ easy check mark.
Approach 3: Reverse Engineer the Item
Dust off the calipers, put in a fresh SR44 battery, roll up your sleeves and get to work. Start taking measurements and with any dimensional drawings put your CAD skills to work. Here are three hacks to help you along the way.
Hack 1: If the vendor provided an unlocked drawing in vector format (pdf, dxf) you can import that into SolidWorks and then scale it to the correct size for gathering measurements. There is a dance to reverse engineering. Knowledge of the manufacturing process/es used to make a part can help to deduce dimensions. For example, if you know a part was cast or molded versus extruded or sheet metal you know to account for draft angles. In reality, designers often use even numbers and once you start measuring, patterns emerge.
Hack 2: Check to see if the vendor (or others) make mounts. This can be another way to get at critical dimensions or simply double check dimensions. We want to be creating our own try to put a positive spin on REing. Use the opportunity to research and learn about manufacturing processes and hone your CAD skills. Embrace the suck. As I am REing I am often thinking about the next design steps of how the heck to fit that square peg into a round hole. The process of intense focusing often results in seeing past implementation as intended with undesirable implications to that light switch moment of an idea that is crazy enough to work and with loads of awesome implications. I wish there was an easier way to get to that ah-ha moment, but for me it always has to come from a prolonged and awkward staring competition with that inanimate object where I refuse to back down.
Hack 3: Googly eyes can lighten the mood. What first sparked our internal design fire was studying, disassembling, and fixing. If you wait patiently, oftentimes someone (electrical engineers) will zap, mishandle, or generally break the part and then you have carte blanche to disassemble and really study the internals. Studying design allows you to notice clever design details for your own repertoire. RE’ing circuit boards is its own can of worms. Certain dimensions are difficult to measure like hole-to-hole, radii, and curved molded parts (more tips and tricks on these to come).
There are a variety of expensive measuring tools and gauges that can be used but this is often overkill. Simple 3D and 2D printers are effective for guessing and checking. Reverse engineering can quickly reach diminishing returns. I do it in stages, often starting with a few basic shapes based on just overall length, width, and height dimensions from a datasheet. What is the likelihood that the part will get implemented? Is the part external and will images be used for marketing? Will the part be reused across products? What is your personal level of obsessive perfectionism?
Approach Meh: 3D Scan the Item
I have experimented with a few low to mid-range 3D scanners on the market. In particular, the Matter and Form and NextEngine 3D scanners, with mixed results. 3D scanner options range from hobbyist to pro and hundreds to thousands of dollars. The first potential issue is you have to possess the part you want to scan (see issues above). Then the difficulty in finding a way to fixture the item in order to scan all the different sides and reflective surfaces like metal or glass without surface preparation. The preparation aspect has been an issue when I am working with electronics that I am concerned I will damage. Then there is the issue of converting the resulting point cloud into a usable model. The technology will continue to improve over time and may be right for your needs depending on your goals and patience level. If someone wants to prove me wrong and send me a 3D scanner, I will be glad to test it out and do an honest review.
With accurate models to build and design around, your project now has a real foothold with roots in reality.
Please help support the open source gnome trap campaign. The gnome dark magic is growing stronger and what they intend to do with it is a known unknown.
*Adapted from Donald Rumsfeld if you can believe it