Computer Aided Designs and Computer Numerical Controlled Machines

[Steve Graves]

There is no doubt that the CNC cutting machines with the designs coming straight from the CAD design softwares are here to stay. It will soon be as accepted as the chisel and bowsaw. Young workers in the music industry are taught using these tools and are expected to know them. I'm working on bracing systems based on a solid cnc modeled rigid brace cut from one solid mass. I'm also designing rigid frames for steel guitars. I'm looking for other designer/builders to join this process on this forum. Can we add this as a topic on the Index?

[Steve Graves]

Can CADCAM be taught to tap tune instrument soundboards ? Got ideas?

[Ron Belanger]

As a tool, I think CAD / CNC have a place, if the builder wishes, in the design of the instruments and in making of the component parts, particularly those that are standardized such as bridges, head-stock inlay, fret-board and nuts and saddles.

For me the important part of the Hand Building process is Hand Building. Speaking about acoustic fretted instruments I don't believe you can teach a machine the finesse and nuances needed to tune a top or a back or any part of the instrument. The pleasure and satisfaction comes from not doing everything exactly the same and knowing that each instrument I build is unique and different and excellent.

[Steve Senseney]

Also, cutting a solid top or back, with the brace as part of the top or back will have grain running the wrong direction in the bracing (for an acoustic instrument).

[Clay Schaeffer]

Hi Steve G.,
I don't think CNC discussions need to be separated from "tools and jigs". I like reading the threads on CNC and cad/cam and I am interested in the new developments that are happening in those fields, but they are just another type of "tool".
I think as the hardware and software becomes more affordable and user friendly (which is happening) more hand builders will adopt the technology.
For a computer to tap tune "plates" I would think you would need to make use of Chladni patterns or something similarly measurable and have it tell the machine where to remove material to adjust them to what is desired. Can it be done? Maybe.
I think most craftsman value the interaction with the materials we use. Our mastery of the materials is to some extent a mastery of ourselves. CNC challenges a different skill set than we use in traditional lutherie, so will appeal to some ,but not to others.

[Steve Graves]

Clay, I agree with all that you've said. The discussion of mankind's relationship with his tools goes back to when the first man used a rock or a lever. But all the passion in the world will not make a better guitar if a person does not have the skills and tools. I realized the possibility of having a machine tap tune my violin tops the first time I rode a Segway. Thanks for the reference to Chladni patterns. Do you know anyone working them into useable algorythms ?

[Andy Birko]

Can CADCAM be taught to tap tune instrument soundboards ? Got ideas?

If something is measurable and repeatable, a machine can be taught to do it....given enough time and money. However, the process of top tuning seems to rely on a lot of intuition and "soft" measurements that would be difficult to teach a machine. Part of the problem I think, is that the goal is somewhat vague.

If you can mathematically characterize what it is you're looking for, it is possible to engineer a solution that gets very close to that goal. I don't have a link, but somewhere there is an engineering grad student paper from a guy who characterized I believe a Yamaha acoustic in terms of tonal response and then used engineering tools to try and duplicate that response with a carbon fiber instrument and he was quite successful. But at the end of the day, if he got perfect characterization and perfect execution - it's only going to sound as good (or bad) as that Yamaha that he modeled. The behavior of carbon fiber is very well known and it doesn't have the natural variation in structural parameters that wood does. This makes life a lot easier when going down that path.

Theoretically, if you could model how you want a top to vibrate and very accurately measure the local structural parameters of a piece of wood, it should be possible to create a program which would adjust thicknesses of braces, top etc. to get the plate to vibrate the way you want it to but even then there are limits. If the piece of wood selected for the top is too far off the mark, it's not going to sound the same. Might be similar but it simply may not be possible to get a particular piece of wood with its natural variations to behave the same as your model is asking.

So to put it another way - without knowing exactly what you're looking for, it would not be possible to teach a machine to tap tune. A luthier's brain is a very powerful computer that seems to make leaps of intuition that are very difficult to characterize. This doesn't mean that we can't use aspects of these techniques. E.g. Many luthiers are already using measurements and spreadsheets to account for some of the variations in a top plate and help choose a thickness that will match the stiffness that they're after.

In another life, I used to work in the auto industry and a ton of money and man hours have been thrown at characterizing how the IC engine works. There are programs and algorithms that will automatically generate a "Base Engine Calibration" for the engine controller but at the end of the day, there's still a guy driving a car around a track or on a dyno to tweak those automatically generated calibrations to get the kind of performance that we're used to.

All that said, CAD, CAM and CNC is a very useful tool in luthiery.

[Simon Magennis]

... I realized the possibility of having a machine tap tune my violin tops the first time I rode a Segway. ...

OK, you got me.

Explain.

[Randolph Rhett]

There does seem to be a lot of either wishful thinking or misplaced fear around CNC machines. I was firmly in the wishful thinking camp a couple of years ago when I first tried to use a CNC machine. The idea that I could focus on dreaming up design elements and have them *poof* appear was intoxicating. I have a computer background and even studied CAD in school (back when dinosaurs romed the earth). Given a few hours I can model most anything luthier related in CAD. Unfortunately, CNC machines are not magic. The translation from design to wood is not smooth.

The truth for me was that the machine was full of limitations, and messaging the "g" code produced by CAM software so that it actually did something useful was extremely time consuming. If you don't know what "g" code is or the difference between CAD and CAM, CNC is probably not in your near future. Beyond that, short of spending the equivalent of a compact car on the machine, the machine itself had many limitations. First the footprint is much larger than the cutting area, so it takes up a huge chunk of your workshop. Secondly, the bits and spindles a hobbyist can afford will take tiny 1/8 to 1/4 inch passes --MAX! It can take hours to cut a neck, for example. In those hours it is almost guaranteed that a gremlin will get in there. Bits slip, motors overheat, dust clogs, etc. Even with good G code I ended up baby sitting the machine for the hours it took to cut anything; all the while anticipating the next weird problem to ruin the piece I was working on. Also, hobby level machines have a fair amount of flex, so precision work requires really tiny cuts (1/64"), adding to the time and frustration.

The further truth, in my opinion, is that unless you have steel fingernails NO ONE makes a guitar by hand. We all use tools, and most of us use electric band saws, table saws, and routers to get the job done. CNC is no different. It is just a tool. It has its place just like a table saw, but it will not build a guiar for you. Thinking you are going to make a guitar entirely by CNC is akin to thinking that an expensive chisel will build a guitar for you.

Having a CNC specific topic is certainly useful. It is such a difficult and frustrating tool to use, it would be nice to have a place dedicated to sharing ideas on how to make it work for luthiers. That having been said, it is just a tool. It shouldn't be dismissed because it doesn't tap tune, or bend sides, or rub out a finish, or glue in bindings, or .... You'll still have to use your hands for that.

[Eric Baack]

i may start building a small xy table and a z axis mount for a router that i can eventually fit some servos and controllers up to for a cnc set up. i am making a 3d model of it in solidworks now

[Jamie Unden]

CNC is not as scary as people make it out to be. If you can do CAD you can learn CNC. The problem is that there are no all-encompassing tutorials out there (That I have found) that lead you through, and if there were it would be for one development path, IE Autocad/CamBam/Mach3 or Solidworks/Mastercam/Mach3. If enough people were interested, a forum section would be great. If I had more time I'd write a book and make a fortune!

[David King]

For a lot of what I need to get done I wonder it it wouldn't be easier to learn and write straight up G-code rather than futse with the CAD programs and then CAM.
Lots of straight lines and right angles in my work. At least I'd have a better understanding of what the machines have to go through when the vectors and 3-d, curved surfaces come into play later on.

[Adam Savage]

Ralph - I think that your 3rd paragraph is an excellent summarry of what CNC is (and associated components). It is no more than a tool, and used well, a very efficient one (carving the majority of excess timber from necks and bodies for instance). And as a tool, like a chisel or backsaw, you need to know how to use it and its' limitations to get the best out of it. It isnt a bad tool, but as you say, it wont build the finished item for you.
I suspect, at least until the cost for 'decent' machines and software comes down, that the place for CNC in the hobby/small producer market might be limited to things like neck joints and pickup cavities. I have never used CNC (which might be a little obvious!), so if I am way off track, I understand....

As a slight extension to the topic, what about 3D printing and prototyping? I suspect that, like DVD's over Laserdiscs, this new technology will leapfrog CNC in price and accessibility in little time. It will be far easierr/cheaper to quickly draft up a 3D model of a new body/neck/etc design and fire it off to a company who produces the 3D printed items than to do similar with CNC. Ok, so you wont be able to work in wood....

Anyway, time for a cuppa.
Cheers,
Adam

[Mario Proulx]

If enough people were interested, a forum section would be great.

I'd welcome a section or thread taught or tutored by someone here, geared toward the absolute beginner to computer drawing(me!), IF it would make use of a free/shareware program. Doesn't Google have something? Sketchup or something?

For many of us, I suspect the first hurdle toward CNC is simply using a computer to draw out our most basic ideas. I'm great with pencil and paper, but totally useless on the computer; I can't even draw a square box!

[Andy Birko]

I suspect, at least until the cost for 'decent' machines and software comes down, that the place for CNC in the hobby/small producer market might be limited to things like neck joints and pickup cavities.

I beg to differ. Neck joints and pickup cavities are the lowest hanging fruit. I've attached a shot of some of the stuff I've made with my CNC machine and I don't feel I've even come close to the limits of my machine.

Everything in Ralph's 3rd paragraph is absolutely true but, just like a hand plane, once you learn the ins and outs, you can do some really fantastic things.

In the picture, you can see a bit of a scroll I carved with my CNC machine. Although drawing the part in CAD, programming the toolpaths with my CAM software probably took 20 hours, it's a much better job than I could do by hand in the same amount of time (I've carved about 7 scrolls by hand and although my latest ones are passable, they're still not nearly as nice as my machine made ones or ones made by an expert at scroll carving). The huge bonus is that now that I've got it done, I can tweak the design to achieve variations in a fraction of the time it would take me to design from scratch.

Here's another shot of a rather novel use for my CNC machine - I'm drilling about 30 equally spaced angled holes in a bandura bridge.

Each hole has a different x,y,and z start point and are at an angle to the plane of the soundboard (to create a break angle for the string). I've done this by hand many times and although I've gotten pretty good at it, the machining method does it in about 2 minutes flat and the positioning of the holes is just perfect!

Actually operating the CNC machine is pretty easy, the real issues are drawing your parts and creating toolpaths and it's something that at least today, is tough to get good at if you're merely dabbling. Like anything else, it takes practice and concerted effort to get good at it. I'm either drawing in CAD, creating toolpaths or using my CNC machine nearly every day these days and I'm happy to say I'm at a level where I can be pretty confident of the results I'm going to get before I start cutting. It's taken a lot of ruined parts to get to this point....

[Jamie Unden]

Ya, but Andy is a legend in the CNC world! Still, it's good to have goals, and someday I'll get there.

What Mario says about using shareware/freeware programs is part of the problem. There are a LOT of CAD programs, they are all different, and everyone has their preference. That's what I meant by the development path - Autocad/CamBam/Mach3 - There are so many choices it's worse than a Chinese menu! Having ONE CNC thread would be hard. Just have a glance at CNCZone and you'll see what I mean.

[David King]

Jamie.

That's exactly what I'm up against myself. The more CAD people I talk to the more it sounds like a CAD program isn't quite enough, you need it to be parametric if you are going to be generating new shapes or sizes from old ones. They all admit that Autocad is a very difficult and awkward program to learn because of it's 2-D deep roots. The newer programs like Inventor and Solidworks are supposed to be much more intuitive and therefore easier to get going on. The price and hardware barriers are very significant. It seems like Rhino3D has been the default for most luthiers due to it's modest price point and unique ability to modify surfaces by pulling or pushing on them. (Don't ask me how that works.)
I tried playing with Google sketchup but was completely stymied by the interface and haven't gone back there. There are other choices like Vectorworks, Amapi, Think-3 (gone bankrupt) etc.
I know people who do all their design work in Illustrator and then import their vector shapes into a cad program.
I wish there were an obvious starting point for all of us beginners and that we could all make the leap together but that might be asking too much. At least Sketchup is free and perhaps it's a reasonable starting point?

[Jamie Unden]

I think each has their strengths. I grew up on Autocad, and I still use it for 2D. I tried TurboCad - hated it. I tried Sketchup - hated it. For 3D I use Rhinocad. At my new job they use Solidworks. I tried it - hate it. I think you just have to pick one and stay with it.

[Mario Proulx]

I'd be happy if someone would just hold my hand and guide me through 2-D drawing, for a start. Sketchup would seem to fill the bill for that?

What stopped me from getting anywhere at all with the demo versions of the usually recommended CNC programs was the language barrier itself. It's a whole new language, and this 'ol dog already speaks enough languages and has little room in his feeble mind to learn yet another, on his own. Sure, there are dozens and dozens of tutorials out there on the various forums, but they were written a good while back, and asking questions when stuck on one of the basics gets you ripped apart and told to "read the archives!". Responding with "but I don't understand what I read in the archives" brings little help... Sound familiar?

[Eric Baack]

You can get into a solidworks educational license for much less then the full package. Or maybe look into getting an older generation of it. Once you make the part, you typically just save it as a more universal file format and import that into the CAM package to make tool paths I believe.