Official Luthiers Forum!

So, I'm toying with the idea of going CNC. What are guys using for instrument building? I'm maybe looking at a 2X4 K2 desktop model. Anyone out there who switched that now regrets it?

2х4' work area is a decent size for guitar work. Mine has a 20x28 (24x36" rails) and I wish it was just a tad bigger. If I had it to do over again I'd get probably 3'x4'. Get as large and solid as you can get with your budget and space constraints.

K2 seems like a solid machine.

Don't forget that the machine part is the least of your worries when it comes to CNC. CAD and CAM software is where most people find the frustrations.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

Here's what I'm looking at:

http://probotix.com/FireBall_Comet_cnc_router/ $3295

or this

http://probotix.com/FireBall_Meteor_cnc_router/ $3995

The prices shown are COMPLETE with computer, monitor, etc.

Just add some software and a router (and mount).

_________________
P A U L B U R N E R
Burner Guitars

I'm kind of downtown cad wise...from a previous business so my point of reference coming into this is skewed to the fancy end of things. ...but I know what you need to do. I recently mentored a local luthier who bought a Laguna IQ Pro which, after my initial poor impression, turns out to be a decent little machine. But I'd advise you to focus on drawing models, making tool paths, and acquiring your equipment....in that order.

The machine is intimidating...but after the initial confusion it's really just a complicated printer that does what its told. Leave it for last.

Drawing is absolutely the most important aspect of using the machines. Without a perfect model....you will NOT proceed smoothly to step two...making tool paths. For the beginner I recommend not to think about the machine or the tool path software.....but simply acquire a cad program to make 3D models. I use Solidworks which is about $4000+ new....plus training. I also have Rhino and I think that would be your best bet since there are decent toolpath modules that can plug into it later. The plug-in aspect of Solidworks and Rhino is important. Once having learned to draw...you will have learned the software interface. Plugging in a toolpath module will be a new learning curve...absolutely...but the having the interface understood is a BIG part of any software package. Don't underestimate this. You'll need training on Rhino or any other decent modeling software...but you'll find it to be capable of doing virtually anything you'll ever need to make guitars.

Don't scrimp on your drawing software. If you buy something cheap...just to get your feet wet...you'll probably find later that it won't do what you want it to do and then you'll have to buy something else and potentially have an entirely new learning curve to deal with. Get the right software and get some training on it. That's the fastest path to productivity.

_________________
I read Emerson on the can. A foolish consistency is the hobgoblin of little minds...true...but a consistent reading of Emerson has its uses nevertheless.

StuMusic

Stuart gives good advice.

The breakdown of time spent learning is so very true. You'll pick up some things about toolpaths from experience (watching them run, seeing what gives you a good finish where, etc) and running the machine really is pretty much hitting the green button. There are tricks in toolpathing and fixturing, but they're usually not huge breakthroughs.

I'd say it took six months or so before I could model most everything I wanted to in a short time in Rhino, so six months to be able to model something like a PRS body with a lot of compound curved surfaces with good intersections. I'd say the hardest thing to toolpath on a guitar is probably a neck, and you'd need maybe a week of fiddling around in your CAM program to be able to put together a good series of toolpaths for that. I can train someone to be a machine operator, who's never seen a CNC, in about an hour.

On a machine you'd need to decide if you're going to work on or upgrade it later. I'd recommend different things to people who were planning to go 'under the hood' on their machines VS people who were going to leave them be.

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

Thanks for all the good advice, guys.

re: tinkering or leaving a machine be
I'm not sure. I'm inclined to leave it be. I had a 15 year career in software, robotics and electro-optics so the machine itself holds absolutely no fascination for me and I've no burning desire to tinker with it unless I have to.

We have a K2 3925G at work. Within its limitations, it's a decent machine. I wouldn't recommend the one we got, but maybe one configured differently.

Pros... on paper it looks great: all aluminum construction with profile rails, ballscrews, and servos on all axes. Our servo drivers are Geckos. It does have the 0.001" repeatability they advertise. It performs pretty well going slow (< 80 ipm) and taking ~light cuts. When cold (60F) and before warming up, the Z sometimes runs at 2x displacement (that will get your attention!) but other than that it's been very reliable.

Now the not so good. We got the machine before knowing much about CNC routers, trusting K2 would make good design decisions (bad strategy). Now that I know more, I'm surprised by their design. Our machine only has 1 bearing per rail for the Y (short axis) and Z. That creates a lot of flex, stresses the bearings, and defeats the purpose of using profile rail. Today they offer 2 bearings per rail as an option. Our machine also has the 8" Z, which makes the flex worse. (Stiffness at the tool is inversely proportional to the square of the z height.)

The main gantry cross-member is 1/4" thick alum plate (not a tube), so it has low torsional stiffness. Combining that with the single bearings, the machine stiffness measures about 1000 lb/in at tool height (not including router flex). So, a 20 lb cutting force causes a tool deflection of at least 0.020". That's a lot, and the low stiffness plus the high Z lead to another problem: the noisy ballscrews excite a vibration in the flexible gantry/bearings. It starts around 100 ipm and gets pretty bad around 150 ipm. A 200 ipm rapid is possible but doesn't seem wise. The gantry vibration creates little dimple marks in the wood when cutting above 100 ipm, worst in the X (long axis), of course. For this machine, I make toolpath decisions partly based on how it vibrates when moving in X and Y.

If we had the 5-6" Z and the 2 bearing/rail upgrade, it would probably be a considerably better machine. If I were doing it over again, I'd use the $ to buy good components and build it myself, but that's a pretty long learning curve to do it well.

_________________
David Malicky

It sounds like you have the expertise for the type of 'tinkering' I'm referring to, which is mainly of the part-swapping type. Knowing how to read a datasheet and swap one servo drive for another isn't something >95% of CNC owners can do, but it's trivial stuff for someone with the training / know-how.

For you, I'd recommend focusing on getting a machine with a strong frame and decent ballscrews above all else. You'll be able to switch steppers for servos later, you'll be able to switch from Mach 3 to a real controller easily enough. Switching lead screws or beefing up a frame are a lot more work. I always advise getting a spindle, because the combination of negligible noise, longevity, and cleaner cuts will certainly improve your enjoyment and productivity. Spindles from SolarJean on eBay are well-regarded and extremely affordable, and so in many cases it's better to get one there than from the machine builder.

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

Hey, Bob, could you recommend a setup and maybe give me an idea what it may cost? I was looking at the K2, but I get the sense that you wouldn't recommend that. If I can't do this and get good results in the $10,000 to $15,000 range with software, I'm may have to reconsider.

Rhino CAD plus a good CAM plugin will not leave you much change out of $2500.....be sure to budget accordingly.

Dave F.

_________________
Cambrian Guitars

"There goes Mister Tic-Tac out the back with some bric-brac from the knick-knack rack"

John,

A budget of 10 to 15k is totally adequate to get a very good setup with something like Rhino and RhinoCAD as your CAD software. It just might not be possible along with a new top-o-da-line K2. K2 looks like a decent machine, even considering Dave M's experiences, but it is rather pricey for what you get. There are other options. If you scour the CNC board, there are several threads concerning machine selection. Here's one I started: viewtopic.php?f=10106&t=30245

When we consider the time-quality-cost triangle, it's obvious that if you want a top quality machine in a short order of time, it's going to cost you money. To bring the costs down, you're going to have to do some of the stuff yourself. I went the kit rout and I'm very happy with my machine. It also took me about 3 months to assemble everything together and get to my first cuts.

To get a good machine, I'm in total agreement with Bob: Get something with ballscrews and profile rails. The probotix looks ok but it has supported round rails which are o.k. but they're not like profile rails. ACME screws can work pretty well too but you'll never have the speed of a ballscrew machine. The Romaxx is a step up from the Probotix for not a lot more but it's also not super rigid. Flexier machines can get the job done, it just takes longer.

If / when I upgrade my machine, it will definitely be a "home made" machine but, I'm also very happy that I went with the kit for starters. The kit saved me probably around a year's worth of fiddling with stuff and now that I've got over a year of this under my belt, I know what needs to be done to make the next machine.

The other option which would also be fantastic is to seek out a used machine. Ken C. on this board picked up a used K2 3925 for about half price. Others here, I think Shane, picked up an industrial (4x4 I believe) machine for I think what I paid for mine but he too had to do some re-works to it.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

I just wanted to thank Andy. He's been going back and forth with me via IM and trying to get me pointed in the right direction.

Andy, That XZero Raptor looks like a very nice machine, and a good bargain. I'd say you did your homework! I'm curious if you've measured its stiffness? The 3/4" gantry plate should help a lot. Also what kind of rapid speeds do you get?

_________________
David Malicky

Well, thanks for all the input and help, everyone, especially Andy. I've been looking at these options for a LONG time, but never came across the XZero. Andy mentioned it to me, I checked it out and badda bing, badda boom....it was a slam dunk for exactly what I need. Banks are closed today, but payment going to Canada tomorrow.

Now all I have to do is decide on stepper driver or motion controller, spindle etc. I'm vaguely leaning towards a Gecko Drive but we'll see. I'm wide open to suggestions here as well. I don't need industrial strength, but I don't want to be upgrading in 6 months either. The XZero is at such an attractive price that I have some room to play with a bit better than entry level on some of this other stuff.

Here's a good faq on steppers and servos: http://www.cnczone.com/forums/faq_cnc_m ... ervos.html
Gecko's stepper guide: http://www.geckodrive.com/gecko/images/ ... 0Guide.pdf

For steppers, the Gecko G540 seems to have universal positive reviews. The 380/381 oz-in Nema 23 motors are most commonly matched with it, with a 48V supply. You may know these already: match the motor inductance with the supply V (there's a formula in the Gecko pdf above), select a supply whose amps are at least 2/3 of the combined motor amps, and that 4 wire (bipolar parallel) gives the best torque and speed. I believe Keling matches things right in their kits: http://www.kelinginc.net/CNCNEMA23G540Package.html
CNCrouterparts has the motors and cables with DB9 connectors presoldered, which saves quite a bit of time for a bit more $.
The G540 minimizes many of the traditional stepper disads (heat, noise, resonance, resolution vs torque). Mounting the motors on a plastic subframe (around the flex coupler) softens the noise, too, especially the buzzy harmonics.

Servos give more high speed torque and are very quiet, but also $$, more complex (encoders), and usually need a pulley reduction to get them in their best speed range for the ballscrew.

_________________
David Malicky

I haven't actually measured the stiffness other than pushing on the nose of the spindle with a dial indicator on there. As with everything, it does deflect but by hand I could only get it to deflect a few thou, nothing near the 20 you were experiencing. In my experience when cutting, speed does make a difference and what I'll usually do is e.g. rough cut profile at around 250 to 300 ipm at maybe .5x of the bit diameter and then do a cleanup pass at full depth at about 150ipm. I pretty much always climb cut. This is fast and accurate.

On my machine, I'm getting reliable 400ipm rapids on the X (long) and around 450 or something on the Y & Z. Accels are I think around 40 on the X and 60's or so on the Y & Z. I'm using a G540 with 381oz-in motors and a 48v linear power supply. on the X axis, I can cut at 300ipm no problem and for some operations (e.g. radiusing fretboards) I even cut at 400.

The downside is that with his design, you can only access a fraction of the table. My rails are 36"x24" but I can only reach about a 20x28" work area. It turns out that this is enough but if I could go back in time, I'd go for the 32" wide version.

Xzero is a great machine but as mentioned in the other thread, he's got some delivery issues but as far as I know, he's never actually not delivered a machine, he's just been quite late at times. Now that I have it though, it was worth the wait.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

On that 24" machine, it's more like 18 when I spoke to him earlier today....at least on the current model. That was dangerously close to 17" so that pushed me to the 32" machine.

Thanks again for the help and the chit chat, Andy. I won't blame you when I'm sitting here in April, waiting for my machine.

I checked my files and I have exactly 19.4" of usable area on my machine, call it 19 if you count the bit width but yes, I'd really like to be able to stick a 24" wide piece of something on there. I think 32" wide would be about perfect.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

He mentioned that he had beefed it up a bit. Maybe that knocked off an inch? I don't know. He may have misspoken or I may have misheard. I do wish he would put those specs up on his site.

Very well could be. If the bearings are a little longer that would eat into travel.

Mine has Hiwins on it which aren't the best out there but they're more than good enough for our purposes.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

I have two of the Xzero Raptors. The first one I bought on ebay--it is one of the lighter built first generation models, but it's still built heavier than most machines out there in its price range with 3/4" uprights, hiwin square rails and beefy ballscrews. It's the same size as Andy's, but not as heavily built. I have been able to throw some pretty heavy work to it, and have never had a problem with stiffness issues. If you go to youtube and look for cnc banjo neck, you can wee this one in action.

I liked the Raptor well enough that when I saw Andy's pictures on this forum of the upgraded machine, I ordered a 24x48 version from Xzero. I didn't have room or the need for a wider machine, but did need the extra 12" length. It took quite a while to get all of the parts, but it was definitely worth the wait. For the money, I haven't seen anything on the commercial market come close. But patience in delivery is required. If you look on Youtube for cnc archtop ukulele, you can see the newer machine in action.

Software and learning how to use it effectively is still the biggest issue, but a well designed aluminum and 8020 router can be a very effective and sturdy machine for the kind of work an instrument maker does.

Dave

What model spindles and VFDs are you guys running? I know Andy has a tekno, but not sure which one or even where to buy such a thing.

re: software
I hear you, there. Coming from a software engineering background I well know the pain of learning how to use complex tools. I've pretty much settled on Rhino for my CAD. I'm going back and forth between madCAM and RhinoCam. madCAM seems like a more complete package for the price point I can consider.

I also got my Teknomotor from George at Xzero. It's the single nose bearing version good for 2kW @ 18k. You can find it here: http://xzerocnc.com/spindle%20cnc%20router%20xzero.htm it's the last one in the top box. He offers a pretty good discount if you get it with the machine.

For a VFD, I'm running a Hitachi WJ200 from driveswarehouse.com with a line reactor. A lot of people are using the cheaper Hitachi, X200 I think but the distributor at Tekno and the drive place really recommended going with the WJ. Apparently it's got "High starting torque and dramatically improved speed stability thanks to advanced sensorless vector control" and it wasn't that much more expensive. The Tekno guy pretty much insisted that I get a line reactor as well. My spindle runs about 5 degrees C cooler once I installed it.

As to RhinoCAM vs. MadCAM, I can tell you that RhinoCAM is pretty darn good but, you have to go to the pro version to get all the cool stuff. I picked up a used licence for the 4-axis which saved me a bunch of cash. I have no experience with MadCAM.

_________________
http://www.birkonium.com CNC Products for Luthiers
http://banduramaker.blogspot.com

Thanks, Andy. Wow, I could only dream of those numbers with our K2. Sounds like a combo of rigidity and the 10mm lead screws to get the 400 ipm. It's great to know there is a company making that kind of hardware for reasonable $ -- maybe others will up their game.

_________________
David Malicky

I use a Colombo spindle with a Delta VFD. No problems so far--seems like a good combination.

On CAM, I've only tried Rhinocam in a test version, but have been very happy with MadCAM, which I bought. It's not perfect, but it has a lot of capability for the price, and generates very good toolpaths very easily.

Dave