Originally Posted by
Marty Jacobson
I built my first CNC machine using 80/20 and plans from Nate at Fine Line Automation. Made some modifications, made some parts for cable carriers, etc. Used Mach3. Total cost all in was well under $5k, including all software (much of which is inexpensive, free or open source these days) and two Porter-Cable routers because I used one up.
I recently rebuilt it to be much more capable and now it runs off of Mesa hardware and LinuxCNC. It's now a 27.5" x 49" x 11" machine that can rapid at 1200 inches a minute (HAAS speeds). Total cost is around $6k now.
Of course that's not counting my time. Time falls into a few categories:
1. Learning how to use CAD: I had university courses in this, and spent about 60 hours a week in university doing CAD for school and freelance projects for 5 years. Then I did architectural design and product development, which was CAD-heavy, for five years. Then I taught biomedical product development and ran a CNC machine shop for research and education for ten years.
After all that, I can genuinely say that making a guitar neck to get a good result 100% by CNC is the hardest thing I've ever done. So hard, I would argue that you shouldn't bother. You can make a lot of necks in the time it'd take you to learn CAD alone.
2. Learning how to program: this is another couple thousand hours, and it's not fun when you are learning because your work, your machine, and your health are all in danger until you really know what you're doing. I can now make parts 100% of the time which come out "right" on the first try, but it was not always like that. Until fairly recently. This is not like 3d printing, programming is not trivial.
3. Learning how to machine: If you are not a pro machinist, a lot of the concepts inherent in workholding and work rigidity will be foreign. A lot of it is common to woodworking, sure, but a lot of it is not "Home Depot" stuff. Things like datums, indexing, degrees of freedom, locating parts for second operations, etc. It's all High School geometry, sure. But was High School geometry all that fun?
So... you have limited time here on Earth. Do you want to spend your time building instruments? Then build instruments. If you want to get into CNC without investing too much time, buy a CNC shark or other bogus POS for too much money and make inlays. It will still be better investment towards your goal of making instruments. Unfortunately, making plates is one of the more complicated things to do on CNC. Some of the cheaper machines have probes you can plug in and it'll scan, sort of, your existing plate, and then it'll make a copy for you, sort of.
If you want to spend 2-3 years of your life, and have a machine that does everything exactly the way you want and ain't nobody's going to do it like yours does, build yourself a LinuxCNC machine from scratch and revel in not ever having to call someone else for help (but also realize, there IS nobody else who can help - lots of forum posts from people trying to be helpful, but 99% of those are people mutually scratching their heads). I'm the kind of guy who doesn't call the IT guy at work, I just fix my stuff because it's faster. If that's how you like to work and that's how you want it to be, a scratch build may make sense. Just be aware that it will NEVER be cheaper if you count in your time, and it will never be faster. You're just choosing to build machines instead of instruments for some portion of the time you have available to work on this stuff.
Fusion360 has full parametric CAD and CAM system which used to cost $10,000 separately, free if you make less than $100,000 (so free for all luthiers, muhahaha).
Bookmarks