In the CAD world, there are all kinds. Some people just use CAD as part of their job—in the same way as they use Word, Excel, or Outlook. Others become power users, investing themselves in becoming experts at using one or more of the mainstream or high-end CAD systems. Then, there are the CAD hackers. People who see computer programs as raw material, to be used and abused in whatever ways are necessary to achieve their creative vision.
Shane McKenna is a CAD hacker. He doesn’t believe in following the rules, or using his software tools in the way that their creators envisioned. Here, he talks about his process, and how he uses tools.
Most of the furniture I built using 2D CAD. To be honest, while 3D CAD is powerful, I can draw and project it out in my head for all the sides, faster than I can fiddle out the panel details in a 3D platform. I have tried many times to bend the time curve with 3D tools, but I still end up going back to my simple little $40 DeltaCAD program to draw it up. Time tested projected drawing techniques, used back in the days of the drafting table and T-square, are still faster for me.
If I am doing a lot of interlocking joints such as boxed corners, mortise and tenon, or shaped joints, then I will build it in 3D to insure fit before cutting. One of the issues is that a CNC router cut panel leaves material in the kerf. Wood compresses and springs back when cut. That spring-back amount varies by feed rate, spindle speed, grain direction, and type of wood or panel that you are cutting. You have to design out this spring back, and simply changing your offset in your tool path program is not sufficient for most of my applications. To avoid re-cutting parts, I do this kerf compensation in the design phase, and check for it in the 3D model at all the joints. This takes a lot of time, test cuts, and fiddling with offsets in part mating. The payoff is my fit is exceptional, often not perceptible, even with close scrutiny. It is not uncommon for me to spend a week to 10 days just on the fit in the design phase on a complex piece. I have a sign up in my shop, “Perfect is close enough.”
As and artist and designer, my primary skill is vision, and as far back as I can recall I could see an object in 3D, and rotate it in my head. This spacial awareness is mostly innate for me, but obviously honed with practice. Once I have a design concept in my mind, I begin to flush it out with hand sketching, hand sculpting, and 2D or 3D software. I don’t use the software to see what something is going to look like: I try to get the software to shape what I already see.
As far as CAD/CAM, I go through my bag of tricks and choose the one that works best to work out the shape. Many times I don’t have an known path, and I have to figure out how to get to the end result with what I have, or what I can create as a tool. The ship bed was a challenge to program the base shape. I could not get the software choices I had at the time to make the shape I wanted, so I designed and built a laser line scanning system to scan a hand sculpted 1/8th scale model. The system was crude, but it allowed me to create the shape I wanted. ArtCAM has the ability to create a surface from a grey scale raster image. So I have used Photoshop to create a shaded image then used ArtCAM to model a relief panel. The carved rearing horse doors are an example of this. I used hand sketched topography, scanned into Photoshop to create a level map of about 12 depth layers. The result gave me a quick way to hog out most of the material from the doors so that I could go in and hand carve the final shape and details. I did that project in record time, under a super tight 1 week deadline. (Another carver had dropped the ball, and they came to me last minute). That method can only work if you understand the shape. In my case, I grew up around horses and have owned many. If asked to do the same kind of project with elephants, I might choose a different way, since I would have to study the shapes more, and figure out how to compress the relief for their shape. Possibly sculpting the relief in clay and scanning it, so that I could add and take away until it felt right.
Sometimes my process can be decidedly “redneck,” like hot gluing up a bunch of layers of acoustic ceiling tiles, and grinding away the shape. It is what I had in arms reach, and in 20 min, I had a fiberglass mold. I will use cardboard, chicken wire, duct tape, or whatever I have laying around to make a shape. The scanner does not care what it is, only the shape counts.
I design longboards, and none of my software is perfect for creating the subtle shape changes for the deck molds. I use SolidWorks to create multiple shapes, then use ArtCAM to morph the shapes together. Maybe there is a way to get to the exact shape I want with just one program, but I have not found it yet. Software does not have to be spendy or have a bajillion features to get used by me. I used SketchUp to design a production machine for processing large volumes of urethane film. I engineered the machine, and built it in less than a week, which got my client a huge contract. I often use SketchUp when I need it fast, and it is all in-house work. I don’t need polished drawings to build my own design. I used it to design a large project for another shop, and it turned out perfect even if the drawings didn’t have all the polish of SolidWorks or AutoCAD drawings.
My method is constant experimentation with all my tools, and being willing to see it through, no matter how long it takes to figure out. In the long run, you learn how to use your software in ways and combinations unique to your projects. You gain skills that translate to other problem solving areas, and you step outside the box. My goal has never been to perfect a method. If I did that, I would falter from boredom alone. My goal is to accomplish the work, learn, and improve. A perfect result is far more satisfying to me than having the perfect tool.
Shane shared a list of the tools he uses for his artistic work (he also does more conventional engineering work.) Quite a mashup:
- DeltaCAD (his favorite 2D CAD software)
- David Scan
- A 5ft X 10ft CNC (that he built)
- A 3D scanner (that he built)
- Many g-code programs, and toolpaths written in spreadsheets.
- A well stocked wood shop.
So, what are your favorite CAD/CAM hacker tools?