This shouldn't be hard for any good CAD and CAM software to handle. Using Rhino and RhinoCAD I project the 2-D fret line onto the compound-radius surface, then use the Engraving function which produces the 3-D toolpath for my frets.
Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
Any amateur builders using their own CNC?
- Thread starter Jeff Siddall
- Start date
700 passes? Yeah, I cut along the length of the neck, not parallel with the frets. So about 75 passes @1mm stepover. Per level.
When I do a neck profile that is functionally a compound radius, I still use the parallel passes and that will have some Z change in each pass.
Last edited:
A fishtail will still give you a flat bottom. Spinning at 21,000 rpm, which I use, and at those slow feeds you will get a flat bottom. The leading edge and trailing edge of that whirling dervish will prevent a raised center ridge.
For depth I project the frets onto the surface of the board instead of cutting down to a flat plane so I don't need to go as deep, and it's super hip and cool. Some fretwire have shallow tangs, I think around .080, I'd have to go measure again. Here's some of the mills I've bought, notice these have a .120 cutting length, and $4 each.
(10) 0.70mm (.0276") 2 FLUTE MICRO CARBIDE ENDMILLS 1600.0276.120 | eBay
Are you doing blind fret slots? I have just been cleaning the slot with a few strokes of a fret saw to get my final width and depth and make sure.
I'm also using larger bits .025 or .027. I've read CNCs tend to cut on the tight side due to a lack of wobbling you find in a handheld crosscut saw. I'm also a fret gluer so a looser slot works fine for me. I got 4 necks waiting for frets and I just made my new fret press for about $12 which feels like it's going to work great.
As an eBay Associate, TalkBass may earn a commission from qualifying purchases.
As long as we're on frets....
I don't remember where I got this tip but when you look at the profile of a fret, the tang and fret bottom don't meet in perfect square edges. There's a bit of a fillet in there. So the tip is take the edge off the top of your fret slots, maybe with a triangular file, to leave room for the fillet allowing the bottom of the fret to seat better on top of the fingerboard.
I don't remember where I got this tip but when you look at the profile of a fret, the tang and fret bottom don't meet in perfect square edges. There's a bit of a fillet in there. So the tip is take the edge off the top of your fret slots, maybe with a triangular file, to leave room for the fillet allowing the bottom of the fret to seat better on top of the fingerboard.
Ah, that makes sense. I think you can do that with "groove" in FreeCAD. I'll look into it.
Yeah, that sounds like a better way to do it.
That's all good info and makes sense. My original question about the fishtail and "flat bottom slots" was not well explained. I agree, the bottom of the slot will be flat but what I meant is that if I cut the slot at a fixed depth then the slot won't follow the fret radius and will instead be flat. To get the right depth at the edges, means it will be thicker in the middle of the board.
Anyway, I think I will just make the slots follow the curvature of the board. As pointed out by you and @gyancey it should be pretty easy to do that.
And yes, since I am going to all that effort I will also do blind slots.
That is a good idea. Files sound so old-school that I might get ambitious and try to add that to the CAD model also.
As an eBay Associate, TalkBass may earn a commission from qualifying purchases.
That would be very impressive if you can get that tiny fillet at the top of the fret slots off the CNC. I can easily make the model do it in the computer but in the real world things keep ending up different.
I've done a few blind slots by cutting normals slots, adding wood binding, then fretting over that after trimming the ends of the tangs. It's a lot of work. Very classy.
I've done a few blind slots by cutting normals slots, adding wood binding, then fretting over that after trimming the ends of the tangs. It's a lot of work. Very classy.
A technical note: The little fillet between the head of the fret and the tang is really tiny. It's not going to interfere with the seating of the fret into the wood. Remember, you are cutting the slot a little bit wider than the tang, so the walls of the slot are spaced away from the sides of the tang, and the fillet. And besides, crushing the top corner of the slot edge a few thousandths of an inch is nothing to the loads of hammering or pressing in the fret.
In the old-school (non-CNC) world of fret slots, there are two reasons why we push a little triangular file across the tops of the fret slots before installing the frets: To remove any wood fiber fuzz, and to make it easier to get the fret tang started in the slot.
The fuzz can be a problem however you cut the slots. It can fold down into the slot as you push the tang in, making the fit much tighter. Or it can fold back, keeping the head from seating on the surface. Little teeny fuzz will just get smashed into the wood, but bigger fuzz from tough, grainy woods can be a real problem, and must be removed. How much fuzz you get depends on the wood, sharpness of the cutter, cutter speed, and feed rate. It's something you are going to have to deal with. In a fully automated CNC line, they'll often have a pass by a secondary spindle with a little "defuzzing" wheel.
In your case, if you have a fuzz problem, you'll probably just get out the little triangular file, rather than mess with changing cutters and programming another pass over all the slots. And besides, making another pass with a bevel cutter can make its own new line of fuzz.
The other reason for slightly beveling the tops of the slots is to help make sure the tang gets started in the slot as you start tapping or pressing it in. Just to make the installation faster and safer. If the tang jams and kicks sideways as you smack it or press it, you can damage the fingerboard.
Some little technical details to think about as you are working out your process. As you are seeing, trying to make a machine do all the work for you gets complicated fast. When you make something with hand tools, there are a whole lot of little details going on that you may be taking for granted.
Thanks for the details Bruce -- all great info. I agree about "getting complicated fast" so I will save that for down the road.
Thanks everyone for all the great feedback!
Having never been one to avoid a challenge I decided get a small CNC despite the many warnings. This is one of those things I just want to try for myself, and they way I see it I have little to lose. If I fail miserably at getting it working I can just sell the machine and then take my files to a big CNC shop to have them do it for me.
So a week ago I placed my order from a seller in Hong Kong some 12,500 km (about 7,800 miles) away. A day earlier my son bought a laptop from Canadian store just 375 km (about 230 miles) away. Both orders were "free shipping" and his order shipped a day earlier than mine.
On Tuesday I got a call from DHL that the CNC was ready to be delivered but needed taxes and duty paid. I paid them the same day and the CNC arrived first thing this morning.
This afternoon my son's laptop also arrived. It was shipped Canada Post "Expedited". Sigh.
Anyway, it will take a while for me to get this thing unpacked and setup but I will post some updates once I get it going.
Having never been one to avoid a challenge I decided get a small CNC despite the many warnings. This is one of those things I just want to try for myself, and they way I see it I have little to lose. If I fail miserably at getting it working I can just sell the machine and then take my files to a big CNC shop to have them do it for me.
So a week ago I placed my order from a seller in Hong Kong some 12,500 km (about 7,800 miles) away. A day earlier my son bought a laptop from Canadian store just 375 km (about 230 miles) away. Both orders were "free shipping" and his order shipped a day earlier than mine.
On Tuesday I got a call from DHL that the CNC was ready to be delivered but needed taxes and duty paid. I paid them the same day and the CNC arrived first thing this morning.
This afternoon my son's laptop also arrived. It was shipped Canada Post "Expedited". Sigh.
Anyway, it will take a while for me to get this thing unpacked and setup but I will post some updates once I get it going.
Last edited:
very interested in seeing how you make out. I wouldn't use a CNC for major operations, but it would be nice for inlays, blind frets, neck pockets and maybe fretboard radiusing.
Update: I decided to climb that steep CAD learning curve up front, and it has taken long enough that I haven't even had time to assemble the CNC yet!
In actuality the problem wasn't really learning how to use CAD, which was covered off by a few tutorials, but rather how to write the python code to completely script the creation of a multiscale compound fretboard. I figured there was already a compound radius fretboard macro, so how hard could it be to add multiscale?
Five long evenings of work later I can confirm: it is very hard. I should have considered Bruce's assertion "trying to make a machine do all the work for you gets complicated fast" more carefully but, per usual, I made it about as complicated as I could. The math itself isn't hard (mostly just trigonometry) but visualizing how to make it do what I wanted took more brain function than I have used in a long time.
My initial attempt to tweak the existing parallel fret design quickly became unworkable as there were just too many things that relied on assumptions of parallel-ness. In a fretboard where it is possible that absolutely nothing is parallel I essentially had to rewrite from scratch.
Anyway, I am happy to say that I prevailed and created what is effectively FretFind3D. In addition to just adding multiscale, I also added the option to have it construct either a zero fret or a nut slot, added string margins with the option to use different amounts on each side of the fretboard, and made several other improvements to usability (ex:configurable perpendicular fret). Here is a screenshot of two bass fretboards, one with a nut slot and one with a zero fret:
Once I figure out the right place I will probably post this on the FreeCAD site in case anyone else is interested.
Having solved my initial problem of making radiused fret slots in CAD, I can move onto the CAM part now!
In actuality the problem wasn't really learning how to use CAD, which was covered off by a few tutorials, but rather how to write the python code to completely script the creation of a multiscale compound fretboard. I figured there was already a compound radius fretboard macro, so how hard could it be to add multiscale?
Five long evenings of work later I can confirm: it is very hard. I should have considered Bruce's assertion "trying to make a machine do all the work for you gets complicated fast" more carefully but, per usual, I made it about as complicated as I could. The math itself isn't hard (mostly just trigonometry) but visualizing how to make it do what I wanted took more brain function than I have used in a long time.
My initial attempt to tweak the existing parallel fret design quickly became unworkable as there were just too many things that relied on assumptions of parallel-ness. In a fretboard where it is possible that absolutely nothing is parallel I essentially had to rewrite from scratch.
Anyway, I am happy to say that I prevailed and created what is effectively FretFind3D. In addition to just adding multiscale, I also added the option to have it construct either a zero fret or a nut slot, added string margins with the option to use different amounts on each side of the fretboard, and made several other improvements to usability (ex:configurable perpendicular fret). Here is a screenshot of two bass fretboards, one with a nut slot and one with a zero fret:
Once I figure out the right place I will probably post this on the FreeCAD site in case anyone else is interested.
Having solved my initial problem of making radiused fret slots in CAD, I can move onto the CAM part now!
Last edited:
That's awesome Jeff. Thanks for sharing it with the world.
Update: I got the CNC built and did a some tests including the first few inches of fretboard from my multiscale 5 string build:
and the last few inches of the back of my planned asymmetric P-bass neck:
There was good news and bad. First the good: It worked! I ended up with 2 parts that, had they been full length, would have been usable after some filing/sanding. I also found a workable toolchain and, after spending most of a day fiddling with things, ended without a single "electronic" problem. None of the steppers missed a step, I never stalled the spindle, and my aluminum table is still unspoiled. This despite doing most work within a few mm of the table and without a spoiler board!
But it wasn't without other problems. By far the worst was the spindle rigidity. I could deflect the bit several mm with my pinky finger, and the result was that it was nearly impossible to avoid chatter no matter how slowly I fed the wood. Granted, I started with hard maple and walnut rather than soft woods, but since I only want to mill hardwood there really wasn't a point in testing less.
Somewhat counter-intuitively, a major issue seemed to be the X-axis linear rails (the ones in the gantry that the Z-axis attaches to). At 10 mm diameter and 360 mm long I expected they would be pretty rigid but it turns out they deflect significantly with any Y-axis load on the spindle. The cheap 775 motor also had detectable play that probably started the vibration. I tried reducing the feed rate to near zero, which only resulted in a period of no cutting, followed by bite and severe chatter, then another period of no cutting. Sigh.
I ended up taking extremely thin passes to have any success at all. 0.3 mm (about 0.010") seemed the most I could do without chatter and that meant even a few inches of finished work took hours to produce.
Other issues then occurred because of the vibration: several screws loosened, including the collet which I had already tightened pretty well, and some of the linear bearings started moving out of their seats.
I also found some oddness with the toolpath on the slicing work which I had generated with PyCAM. It seems like it put some cuts well beyond the margins I had requested, resulting in rough cuts ending up below the final profiling pass. This is evident on the walnut piece, though on the fretboard, where I didn't do much slicing, the profiling there turned out pretty nice. I also discovered that the surfacing algorithm appears to only follow the profile along the X axis, and follows a straight path in the Y direction when using the circular tool path. So many quirks to learn.
Despite those issues, I have seen a glimpse of the results I could achieve if I remedied the problems. I made some design changes, and ordered some parts, and will post an update on the performance once I have them installed.
The biggest changes are the switch to fully supported rails for the X and Y-axis, an all-metal Z-axis assembly, and a better 0.3 kW spindle motor. Since I was replacing X-Y axis rails I also ordered parts long enough that I could do fretboards and necks up to 700 mm (28"). I should end up with a 700x300x72 mm cutting volume adequate for most guitar work.
Those changes will take several months as some parts are coming from China and many there are on holiday until mid-February. Then they have to sail around the globe. Hopefully next post will include a finished compound radius fretboard!
and the last few inches of the back of my planned asymmetric P-bass neck:
There was good news and bad. First the good: It worked! I ended up with 2 parts that, had they been full length, would have been usable after some filing/sanding. I also found a workable toolchain and, after spending most of a day fiddling with things, ended without a single "electronic" problem. None of the steppers missed a step, I never stalled the spindle, and my aluminum table is still unspoiled. This despite doing most work within a few mm of the table and without a spoiler board!
But it wasn't without other problems. By far the worst was the spindle rigidity. I could deflect the bit several mm with my pinky finger, and the result was that it was nearly impossible to avoid chatter no matter how slowly I fed the wood. Granted, I started with hard maple and walnut rather than soft woods, but since I only want to mill hardwood there really wasn't a point in testing less.
Somewhat counter-intuitively, a major issue seemed to be the X-axis linear rails (the ones in the gantry that the Z-axis attaches to). At 10 mm diameter and 360 mm long I expected they would be pretty rigid but it turns out they deflect significantly with any Y-axis load on the spindle. The cheap 775 motor also had detectable play that probably started the vibration. I tried reducing the feed rate to near zero, which only resulted in a period of no cutting, followed by bite and severe chatter, then another period of no cutting. Sigh.
I ended up taking extremely thin passes to have any success at all. 0.3 mm (about 0.010") seemed the most I could do without chatter and that meant even a few inches of finished work took hours to produce.
Other issues then occurred because of the vibration: several screws loosened, including the collet which I had already tightened pretty well, and some of the linear bearings started moving out of their seats.
I also found some oddness with the toolpath on the slicing work which I had generated with PyCAM. It seems like it put some cuts well beyond the margins I had requested, resulting in rough cuts ending up below the final profiling pass. This is evident on the walnut piece, though on the fretboard, where I didn't do much slicing, the profiling there turned out pretty nice. I also discovered that the surfacing algorithm appears to only follow the profile along the X axis, and follows a straight path in the Y direction when using the circular tool path. So many quirks to learn.
Despite those issues, I have seen a glimpse of the results I could achieve if I remedied the problems. I made some design changes, and ordered some parts, and will post an update on the performance once I have them installed.
The biggest changes are the switch to fully supported rails for the X and Y-axis, an all-metal Z-axis assembly, and a better 0.3 kW spindle motor. Since I was replacing X-Y axis rails I also ordered parts long enough that I could do fretboards and necks up to 700 mm (28"). I should end up with a 700x300x72 mm cutting volume adequate for most guitar work.
Those changes will take several months as some parts are coming from China and many there are on holiday until mid-February. Then they have to sail around the globe. Hopefully next post will include a finished compound radius fretboard!
Thanks for the update on this. I feel like CNC is in my future, but I know nothing, so posts like help me (and I'm sure many others).
I see an awful lot of novice builders who jump into cnc machines thinking it is a way to bypass knowledge and skill- dozens a year at a minimum visit my place and show off their hard work. 95% of them get terribly distracted and spend a $#!@%^&* of money..... and in the end they build a pretty lousy instrument. They compromise good instrument design with easy to write CAD and coding. Arches and recurves and radii that are centuries old and proven in instrument design are not the same arches that beginning techs like to write.
A cnc will save your hands some carving time BUT it doesn't do you any good if you don't know what a good arch on an acoustic instrument is, what a great playing and functioning neck feels like, how a good neck joint should work and function. You are much better off learning to build a simple, traditional instrument that is well executed by hand. It takes me under and hour to compound radius and fret an instrument by hand. I'd never give up the ability to carve a neck by hand- in the white while strung up is one of my favorite tasks. The same thing with making carved archtop instruments. There is no quick and easy substitute to the decades of time it takes to become a skilled luthier.
Once you've built a few dozen or few hundred of those and truly understand what it is you want to build and what you want to say to the world, then it is a good time to consider more production style machinery. Without that knowledge, you'll just spend a lot of time and $ making mediocre instruments with fancy tech. I enjoy being a woodworker over a computer tech and my customers all dig that I build vintage style instruments with the same tech that Leo used....
Do you want to be a cnc builder or do you want to be a bass builder? I know a couple of folks who are so in debt from their $75000 machine investment that it controls their life and production. I'd much rather have lower output, drive and old truck, and have my afternoons free to go mountain biking than always be stressed about that giant machine. I've used cncs for over 20 years and currently do not own one, but I do have a very good relationship with four different folks who do that work for me. Today we have in incredible luxury in that for a few hundred dollars, you can find dozens of people who are good cnc techs that you can sub out parts to while you focus on the bass building.
In my own shop, I much prefer the smell of rosewood sawdust and carving the neck with a rasp over the sound of a F#@!$%^ router all day long.......
A cnc will save your hands some carving time BUT it doesn't do you any good if you don't know what a good arch on an acoustic instrument is, what a great playing and functioning neck feels like, how a good neck joint should work and function. You are much better off learning to build a simple, traditional instrument that is well executed by hand. It takes me under and hour to compound radius and fret an instrument by hand. I'd never give up the ability to carve a neck by hand- in the white while strung up is one of my favorite tasks. The same thing with making carved archtop instruments. There is no quick and easy substitute to the decades of time it takes to become a skilled luthier.
Once you've built a few dozen or few hundred of those and truly understand what it is you want to build and what you want to say to the world, then it is a good time to consider more production style machinery. Without that knowledge, you'll just spend a lot of time and $ making mediocre instruments with fancy tech. I enjoy being a woodworker over a computer tech and my customers all dig that I build vintage style instruments with the same tech that Leo used....
Do you want to be a cnc builder or do you want to be a bass builder? I know a couple of folks who are so in debt from their $75000 machine investment that it controls their life and production. I'd much rather have lower output, drive and old truck, and have my afternoons free to go mountain biking than always be stressed about that giant machine. I've used cncs for over 20 years and currently do not own one, but I do have a very good relationship with four different folks who do that work for me. Today we have in incredible luxury in that for a few hundred dollars, you can find dozens of people who are good cnc techs that you can sub out parts to while you focus on the bass building.
In my own shop, I much prefer the smell of rosewood sawdust and carving the neck with a rasp over the sound of a F#@!$%^ router all day long.......
Looks like a good start Jeff. Odd how it seems to Z dive a bit at a regular spacing in some of the pics.
One possible improvement to the results would be to run a few finishing passes with the stepover reduced to half on the last pass.
The slop in the rails will be most prominent in the roughing passes and the finishing passes won't be pushing back so hard and allow you to get a bit tighter tolerances.
Are you using a 1/4" ballend? A 1/2" will smooth out those cusps a lot especially if you use the same feed rate you do for a 1/4".
And yet another point, if your machine has that much slop make sure your fret slotting passes all run the same direction to try and keep the errors all pointing in the same direction and not compounding each other.
These days I'm trying to get closer to finished on my projects. I've started by leaving a good deal of hand work at the end to make sure I don't go too far on the CNC but now I'd rather spend 20 more minutes on the CNC to save 1/2 an hour of sanding.
One possible improvement to the results would be to run a few finishing passes with the stepover reduced to half on the last pass.
The slop in the rails will be most prominent in the roughing passes and the finishing passes won't be pushing back so hard and allow you to get a bit tighter tolerances.
Are you using a 1/4" ballend? A 1/2" will smooth out those cusps a lot especially if you use the same feed rate you do for a 1/4".
And yet another point, if your machine has that much slop make sure your fret slotting passes all run the same direction to try and keep the errors all pointing in the same direction and not compounding each other.
These days I'm trying to get closer to finished on my projects. I've started by leaving a good deal of hand work at the end to make sure I don't go too far on the CNC but now I'd rather spend 20 more minutes on the CNC to save 1/2 an hour of sanding.
Before I joined my makerspace and got my hands on a CNC I spent months contacting people and they wouldn't do custom work outside of the norm and wanted up to 4 figures for a body. Do you have any faves you can give us links for?
Thanks James for the well laid out thoughts. Having access to a wide range of experience and opinion is one of the things that makes TB great.
I have a few comments after reading your posts that will add some context to what I am doing, and perhaps what some others here might be thinking of doing.
First off, I fully agree with your point "There is no quick and easy substitute to the decades of time it takes to become a skilled luthier." And this is where I need to be pragmatic. I am an amateur. Very amateur. I am just now working on building my first bass from scratch and can probably count on my two hands my complete collection of woodworking tools. I typically have only a few hours a week to put into building things which means, short of a career change, I will never achieve the skills that you, and many others here, have.
Then there is your other question "Do you want to be a cnc builder or do you want to be a bass builder?" and for me the answer is essentially "both".
So my dilemma, then, is how can I combine my interest in making guitars with some existing tech skills, plus a decided lack of lutherie skills, to somehow produce a few nice instruments for my own use.
Which is why I started looking at CNCs. Most shops get a CNC because they can already build great instruments and want to scale production. But my goal is nearly the opposite. I am looking at what tools I can obtain to help me to create a great instrument with the least investment. Certainly it is easy to spend a ton of money to buy a CNC, but again, my goal is the opposite: how little can spend to get a tool to do an assortment of cool things that would otherwise be hard/expensive.
So far I am in $161 for the CNC, plus a $25 assortment of mills, and I already have some encouraging results. Honestly I feel more excited than stressed at this point, but I think that is probably because I am not "so in debt from [my] $75000 machine investment that it controls [my] life"!
As to whether I can turn all of this into something greater than "...a pretty lousy instrument". Only time will tell, but I think I can.
Last edited:
