GuitarEng

I'm still semi-seriously playing with the idea of getting a CNC setup made, although probably a converted mini mill and not a big router-based unit, which I simply do not have the space for.

the thing I keep coming back to is this: I have yet to build two identical guitars, and I won't save any time (none at all, zero), likely waste a lot of time designing everything in CAD, transferring to G-code, etc. if I keep on doing - essentially - one-off instruments. And once a template's made, well, bandsawing and template routing a body takes, what, 10-15 minutes tops? Forget about 3D shapes like necks; lots of programming time and cutting time (yes, time I could do something else, I know, but I'm not doing this in a production setting) for something (making a neck) that takes little more than half an hour to an hour starting from a prepped neck blank.

The thing's I'd like to do with CNC: make metal hardware (bridges, headless systems), metal detailwork/small items (pickup rings, various hardware bits and bobs), inlay (logos in particular), but it's still a lot of money and quite a lot of time for someone without aspirations of large-scale building. For example, I can hardly think of a worse option in terms of prototyping.

I can build an average eletric, including a fair amount of inlay, in about 100 hours of actual work (I'm a bit slow, though, and have few consecutive hours to spend on much of anything). No way I could do that with CNC, which would still require a lot of hand work (the hard stuff) to get fitted together right and feeling right.

Thing is, it is very accurte, fantastic way to make precise parts, templates, jigs (if you don't want to cut the instruments themselves), so if/when I get a bigger place, I may well shell out the bucks and get a CNC. Thing is, by the time I get a bigger place, I should be able to afford a K2

If you are just interested in CNC as a hobby then build one yourself. If you are interested in it as a tool, then buy one.

If you can take a block of wood and turn it into a neck in 1 hour (profiling, headstock contours, tuning peg hole drilling, and carving), then don't waste time with CNC!

It would be hard to make a similar argument against CNC for inlay.

One-off's are pretty easy if they are truly one-offs. You have to design everything at some point...and if you have your CAD skills down then modifying existing drawings and making code from them is pretty trivial.

For metalworking, you need a CNC mill, not a CNC router. Unless you can afford something suitable for both (i.e. a used Fadal for between $25 and 50k) then you will need to chose one or the other. You definitely don't want to do metal working with a K2.

I use Fadal's at work and I have a K2 at home. CNC is the way to go for me...but I have 15+ years of CAD. If you aren't up to par on CAD then that is the first step.

I pretty much do CNC all day, most days, let me know if you have any questions.

G-Eng

I can also CNC cut pretty much anything you would want within reason.

Send me a sketch and I'll send a quote.

G-Eng

I thought some of you might get a kick out of my latest design. I'm looking for comments both positive and negative...lay it on me!

I designed everything using Rhino CAD and then cut the parts using my CNC machine.

Here are some pics:

Peace,

G-Eng

I use ebony...oh yeah...the CNC machine helps out a lot with this...

-G. Eng

You're really just inviting a headache if you make a single piece guitar...mainly due to wood movement. If you do attempt something like this, I'd recommend using something fairly stable such as mahogany or rosewood.

As for tonal differences due to adhesives...I think that's stretching it. There is absolutely no such thing as perfect tone..it's completely subjective. I would really doubt that a .002" thick layer of glue really affects "tone" (however you define it) to any appreciable degree. There are a TON of things that matter way more than how thick your glue line is. My advice is to worry about those things.

That's my .02

G-Eng

the harder and stiffer the wood the longer it will sustain. the specifics of the wood dictate which frequencies sustain the most (as well as the design an many other things).

I've always understood that you need mass to get good sustain. Pick the heaviest block of wood you can find and that will sustain the longest.

Not sure on the head angle (I seem to recall 11 degrees, but I'm not positive), and the neck angle depends on how you position your neck (raised, flush, where it joins the body), scale length, bridge height, shape of your carve, so unless you're building a PRS copy from exact plans (in which case you'll have the neck angle already), use the pinned tutorial in the tutorials section to figure out what the neck angle needs to be for your own guitar.

The PRS-style guitars I've made have needed neck angles of about 3.5 degrees, but since I do a slightly different carve every time, I always, always calculate the angle again before routing and fitting the neck.

Ditto on what Mattia said - take your time and do the design right. Draw it all out with the right angle & everything..

You'll probably see me advertising on here before too long. All of those things have been in the long term plan...

In the meantime, let me know if I can help you out with anything!

Nifty body. I like bookmatched and slip matched. If the grain is cool looking it works(well unless you are talking about a functional plate or soundboard, then I have a little different preference). Looks like you are getting some good use out of your machine.

Peace,Rich

Yeah...it's cool stuff. I'm looking to start running a lot more stuff soon. The wife tells me I have to actually make money...apparently I promised something along those lines at one point...

Thanks for the compliments!

I cut it on my CNC - I have one of the K2 Machines. I did all the design & surfacing work using Rhino & ran my toolpaths with RhinoCAM.

I didn't have any bookmatched wood laying around...and I thought it looked cool having the pieces flipped (the top wood pieces are sections from the same board)...the swirl on the top opposing the swirl on the bottom, etc.

I was going to go with much more standard F-holes and did this one on a whim. I've seen some other f-holes done per the grain lines..and this is my first go at doing it with the CNC.

Here's the latest body off my CNC machine..it's a Semi-Hollow Singlecutaway design with the F-Hole designed to mimic the grain pattern. The body has a 2 degree taper between the top and the back and it has a nice rounded carve transitioning from the top to the sides. Oh yeah...the back is a single piece of Mahogany and the top is Black Walnut.

Let me know what you think..here's a pic of the body with the top wet down...

I would think that rosewoods and ebony are used for their stability and strength. It probably doesn't hurt to have some weight to the bridge either. Rosewoods and ebony are pretty stiff...if they were more flexible, this would probably steal some energy from the strings instead of transferring the energy right to the soundboard.

I think there would be some obvious problems with using a softer wood for a part like this...over time & under load, the part would deform due to the string tension & load across the saddle.

I'll CNC cut them if you want....I can do rounding, slotting, profiling, and do any inlay routing you may want as well....Message me if you are interested.

I've got a K2 3925 machine and it is repeatable to within about .002. This number includes all sources of error from all 3 axis. I highly recommend it...it is good for "slower" production work. The main limitation is the size of the spindle (which is a 3.25 HP porter cable router)..and the flexibility of the frame. Even so, it is definitely faster and more accurate than doing things by hand.

The compucarve thing is not geared towards inlay...or really any other parts of guitar making.

A machine like the compucarve is controlled via stepper motors whereas higher quality machines make use of servo motors. Stepper motors only move in certain increments..and when they wear out they miss steps. The machine controls position of the axis by counting how many pulses it sends to the motor. The servo motors on my K2 have optical encoders...they constantly feedback their position to the controller PC.

For anyone who is interested in getting into CNC, I highly recommend you download a copy of Rhino CAD software. If you get to the point where you are comfortable with CAD, then you are ready to go to CNC

one last thought to add to fryo's would be that you probably aren't going to drastically change the way the instrument sounds if you 1) build it out of normal woods 2) build to pretty standard dimensions 3)brace it using a standard pattern and 4) put the sound hole in a normal spot.

You gotta play big to win big..if you're only making $1 bets, it's hard to strike it rich! And even then...there's always the possibility that the time tested designs we all play today are pretty darn close to an optimized design...

Soundholes are going to set up your Helmholtz resonant frequency(like when you blow on a soda pop bottle and it creates a tone). The size of the soundhole opening and volume of air in the box will give you your resonant frequency. Do a quick search on the subject and you will get a better idea as to how it works, and you will also find formulas if you would like to do the math.

Placement, size, are all personal choice based on what you want to achieve, depending on the size and design of your guitar.

Peace,Rich

Here's a cool pic - this is from a very simplistic Finite Element Analysis model that I ran for a prototypical acoustic soundboard. Without going into the details, I think it is indicative of the deformed shape that an acoustic takes on when it is under load. I'd also propose that as the strings vibrate, they basically form a bellows...with the back portion of the body "pumping" air out the soundhole...and the "tone" of which is probably dependent on the Helmholtz frequency. This tells me that you should put your soundhole wherever you think you'll get the most deflection...and that this should give you the most volume...

Cool, congrats on picking up a seat of Rhino.

I also have Rhino 3.0...

I'll reply first to your 2nd to last post: Once you have done all the surfacing, you program the CNC so that it cuts the whole thing all at once. You don't cut the arm carve, then the binding, then the rest...you cut it all at once using either a spiral toolpath or a simple back and forth toolpath (depending on the part and your CAM software).

With regard to how to model a carved top....I would recommend starting with a neck carve. Carved tops are definitely more difficult. In both cases, you will be using the following commands:

*surface from 2, 3, or 4 edges

*Lofts

*Surface from network curves

*Sweep 2 rail

*Patch surfaces

Each of these tools have their place. The general order of operations is to build a wireframe grid..and then to use those splines & curves to create your surfaces. Make sure when you do the grid that if you have two curves coming togethor, they are are tangent at a minimum...and preferrably curvature continuous.

I'd recommend reading the book and then doing a lot of playing...you will need to spend a good bit of time on this in order to get good at it. Good luck!

Just to give this a bump update. I was able to find a dirt-cheap v3.0 of Rhino on Ebay, and I've got the 2D shape down using polylines, no problem. All x and y axis so far.

Now where to even begin to shape a carved top onto the 2D outline I've got so far?

I can use Surface | Extrude curve | Straight to bring it into the z axis and give it thickness, but that's not really a carved top.

How to start to create something like the carved back in post #11? (obviously Rhino) What menu items?

Hey Dave, cool site. I have a similar background & similar CAD experience (less UG & ProE). I have done a lot of CAD automation/Design Automation programming using VB, Delphi, and Lisp though - you can do some cool stuff if you get into that.

Very nice web layout as well, it has a great look to it.

I do have to take exception with your idea that Rhino is only good for making swoopy curvy things. It is especially good at this, but I feel it is as good as AutoCAD at the more mundane CAD tasks. With that said, it really doesn't compete at all with SE, SW, or Inventor as a solid modeler. Rhino simply calls watertight surfaces "solids". Yuck.

Here's my experience, but note that with all of these programs if you're not stealing it, they're not cheap.

- SolidEdge: Cheap for schools so it was ubiquitous for a while. OK interface, crashes too easily, hard to get complex parts down easily.

- SolidWorks: Most user-friendly CAD package. Very powerful once you learn how to use it. Decent surfacing capabilities, not great for parametric based models if you need mathematical flexibility. This is my usual go-to. Built in FEA tools, decent drawing tool, and assemblies that can easily handle moving parts are all very nice and accessible features. Probably also has the easiest learning curve of most CAD packages.

- AutoDesk Inventor: CRAP. Sorry, but I hated it. Cheap, and more like a kid's drawing tool with it's snap-to-grid setup. Renders very well, but the drawing, modeling, and analysis tools are all very underpowered and not very versatile. The fastener library, though, is very nice.

- PRO/Engineer: Depends on what you're using. Wildfire is slick and tries very hard to keep the raw power of PRO/E while gainig a SW-like interface. Older versions are terrible from a UI perspective, but this is a very capable program. Very difficult to do surfacing work, but the parametric kernel makes for a very flexible model once you learn to set it up well.

- UniGraphics: Powerful tool, horrible interface. Unix based versions are the most common and the hardest to use to boot. I know some people who love this and can do great thigns with it, but these are invariably people whohave not used other packages much.

- Rhino: limited experience with this, but it seems to be more of a 3D modeler's program for making pretty swoopy surfaces than making precision engineered parts. There is a definite market for this, though, and guitars may overlap Rhino better than SW. That said, only real serious CAD monkeys can make SolidWorks do the surfaces that Rhino lets even rookies pull-off, so if you want to setup some surfacing for carvetops this is the easiest way to start.

- Cattia: heavily used in automotive, so it must have the chops. Only used it for a few hours, didn't feel super-simple to me, but half a day is not a reasonable amount of time for an evaluation.

As for grids, don't ever use grids for final layout so forget you even asked. You absolutely MUST use dimensions and constraints to fully define your parts if you want to have any kind of accuracy in the final product. That means if you use the grid for a quick layout immediately afterwards you use dimensions to set the actual down-to-the-thousandths spacing of your frets.

A quick note on CNC machine selection. Envelope will probably be the most relevant feature for you as accuracy and speed aren't nearly so critical in woodwork as they are in machining titanium and aluminum alloys for aerospace. HAAS and Fadal have a very strong presence in the job-shop market across the US, and seem to be in a lot of schools (excellent marketing) so many people are familiar with them, but widen your search and there are a significant number of VMC manufacturers to consider.

-Dave

The grid itself is inconsequential......when I'm laying out a part to show fret slots, or something like that, I'll draw the nut line and then use the offset command and then trim the slot lines to the edges of the neck. Most CAD systems allow you to enter coordinates/measurements so that you don't have to "eyeball" everything using grid & snap options. That would be a truly painful way to draw..!

Whatever software you buy should come with tutorials & a manual...if you are new to the software you could probably get a lot of mileage out of reading through that stuff.

Another CAD software question - what grid dimensions are most commonly worked in?

For instance, fret slots are usually done in 100th's (IE: 6.397" for the 5th fret on a 25.5" scale) whereas the most common nut width is 1 11/16" which would imply working in 1/16ths or 1/32nds (1 11/16" converts to 1.6875").

So is 1/100ths or 1/32nds for the grid?

I just saw the post about Solid Edge. I think that in the mid-level 3D CAD market - this means Solid Edge, Solid Works, and Inventor - SolidWorks is the best...and the other 2 are trailing. We use SW at work and it's a great package.

I gotta get my jabs in for Rhino though...Rhino does some surface stuff that SW just doesn't do. SW is a much more well rounded package, but Rhino is a fraction of the cost...

they cannot accept drawings with splines on them.

Tonemonkey - could you explain? Do you mean something like an LP carved top?

GuitarEng - where are you located?

Also (anyone) this is a good animated demo of Rhino:

http://www.k2cnc.com/videos/GuitarProjectw...thCallouts.html

Bottomline: if I were to invest in just *ONE* software package, would Rhino be the one? ...that would allow me to model the guitar AND then bring the dwg file (and the wood) to a shop w a K2/Haas/Fadal?

Tonemonkey - most shops with decent software can handle just about any curve that you give them. Splines are a pretty basic drawing element for most CAD/CAM programs...I'm surprised to hear that they are giving you a hard time about that.

P90 - I'm in Maryland (I used to post under the name "TechArt", but I lost that login info!). A bit more on what TM was talking about...splines are 2D or 3D curves...if you were going to model a LP top, that would be done using surfaces (which are usually creating using a bunch of splines). Rhino is very capable of doing surfacing tasks like this...but surfacing is an art and it will probably not be a quick thing to learn.

I chose RhinoCAD & RhinoCAM as my software and couldn't be happier with it. I feel that it is by far the most competitive package out there.

I think the best way to get into CAD/CAM is to get into CAD first. You can only machine what you can draw...and even if you never grow to the point where you are ready to jump into CNC, CAD is an extremely useful design tool.

Here's a screenshot of a surface I setup to carve a carved back for an OM style acoustic that I have been working on occasionally (it's basically an archtop-type surface):

Here's a pic of the actual part, carved with that surface from Black Walnut..the outline is a bit rough because I bandsawed around the outside. After that, I glued it right to the kerfing and used the router to trim the back even with the sides (as you would normally do). I'm really happy with it, I think it's noticeably more comfortable than a regular acoustic back.

Thanks! I went to the K2 website and they seem to be marketing their KG-3925 toward luthiers. Here's a FAQ page with links to a video and software demo near the bottom:

http://www.k2cnc.com/FAQ_GuitarMaking.asp

Also - moderators - should this be moved to the tools & shoptalk forum? (I dont mind)

That's the machine I have - I got it with the 8" Z and the limit switches. My only complaint is that they didn't provide hardly any info on the machine itself & basic operation. It wasn't hard to figure it out (I called them)...but it would have been nice to have something to go by!

Yep, get something that will export dxf files. AutoCAD is not really a standard anymore...but their file format is a standard. Most engineering/design places are now using something like SolidWorks, SolidEdge, or AutoDesk Inventor. These are mid-level parametric modeling packages that work great for building parts & assemblies in 3D....cost on these is around $5000 per seat whereas I purchased Rhino & RhinoCAM together for $1600. MasterCAM is nice, but it also runs expensive...about $10k to start.

Fadal 4020's are a popular model, a lot of makers use these machines. They also have a 15,000 rpm spindle which helps when cutting inlay, etc. A fadal will run you about $85k. I think the real justification for a machine like a Fadal is the quantity of parts that you are running. If you are only making one or two guitars a day, a Fadal is a much bigger machine than you probably need. A good starter machine (such as a K2) is available for about 10% of the cost of a Fadal....you just have to take smaller passes with the smaller machine. (I have a K2 at home)

Let me know if I can help you out.

Most any CNC shop will be able to run toolpaths off of dwg or dxf files.

Thanks! So what I need to do next is get a software app which will produce .dwg/.dxf files? Is AutoCAD a defacto standard? (You can get older versions of AutoCAD cheap on Ebay)

BTW, which Fadal/Hass machines are the most popular for luthery? What working envelope? I'm guessing the VMC2216 or VMC4020 for the Fadal or the VF1 or VF2 for the Haas?

http://www.fadal.com/?id=711

http://www.haascnc.com/VMC_MODEL_20IN.asp#VMCTreeModel

Most any CNC shop will be able to run toolpaths off of dwg or dxf files. I use Rhino CAD/CAM with my CNC machine and have been very happy with it (I use MasterCAM, Rhino, SolidWorks, and AutoCAD at my day job). I've found Rhino is better than AutoCAD for artistic designs...AutoCAD is more for the mechanical/architectural design. Rhino has great surfacing capabilities...I'll use this for doing carved surfaces like necks & bodies.

Haas and Fadal are both popular machines because they have decent spindle speeds (up to about 15,000 rpm), tool changers, and the accuracy of a low end metal working machine (they will generally hold .001"...which is very good for woodworking).

You could probably pick up a very basic CAD program at Best Buy or something for around $100 (I believe Rhino is about $600).

If you have a design (digital or not), I would be happy to work something out with you - just let me know.

I'm a total noob at CNC. I've read through all the threads here and the one linked from here to the Gear page. I had some experience with AutoCAD at school years ago.

Is there a piece of software where I can layout on my home PC in 2-D the exact dimensions of what I need, then bring a file on a disk (CD) to a shop with a CNC, insert the disk and have it carve out what I need on a piece of lumber (which I would also supply). I would presume the most popular CNC's for this seem to be Haas or Fadal?

What software might allow me to layout the 2-D cuts at home, and bring the disk to a CNC shop?