CNC Making Discussion

[Don Noon]

Nathan,

You don't know how slow I am with a knife, and how crappy my hand-cut bridges turn out .  And any hand work that I can avoid will prolong how long I can keep making these things.

I would prefer to see a separate thread to (re)hash the aesthetic/artistic opinions on CNC, and keep this thread for the users to exchange ideas.

Edit:  OK, I started up another thread for the side discussion.

[carl1961]

here is a interesting Fusion 360 video doing the         Guarneri 'del gesu' Violin 1741                              

 

[nathan slobodkin]

1 hour ago, Don Noon said:

Nathan,

You don't know how slow I am with a knife, and how crappy my hand-cut bridges turn out .  And any hand work that I can avoid will prolong how long I can keep making these things.

I would prefer to see a separate thread to (re)hash the aesthetic/artistic opinions on CNC, and keep this thread for the users to exchange ideas.

Edit:  OK, I started up another thread for the side discussion.

Don,

My apologies for the tone of my response to Michael and my ungraceful criticism of your work. I will be  glad to move my further discussion of this topic to the other thread and leave the CNC folks to their technical issues.

[scordatura]

11 minutes ago, carl1961 said:

here is a interesting Fusion 360 video doing the         Guarneri 'del gesu' Violin 1741                              

 

Carl. This is awesome! It is really cool to see how he approaches the modeling process. Thanks

[carl1961]

here is the instructables of his work https://www.instructables.com/id/Fusion-360-Proportional-Design-for-the-Violin/

[Don Noon]

Thanks for the Fusion360 link.  I'll probably have to watch it carefully several times to try to understand all the tools and operations that go whizzing by.

But the arch he comes up with is nowhere near the shape I want to get.  So I'm not sure if the demonstrated method would be applicable.  At least there might be helpful hints on how to use Fusion360 for 3D work.   I see this as a typical fault with CNC use... the shape you can conveniently model with the software is not the same shape and experienced maker would be getting by carving.

RIght now, this is still looking like a very steep learning curve that I might not be able to climb for months.

[Fiddlemaker5224]

Very informative video, thanks for posting it Carl.

[scordatura]

2 hours ago, Don Noon said:

I see this as a typical fault with CNC use... the shape you can conveniently model with the software is not the same shape and experienced maker would be getting by carving.

I am not sure I agree with this Don. If you can recreate the arch from the templates that exist for a given model, your "carving" is your ability to create the curves with the software. Granted there is a lot going on in a curving 3 dimensional space. I have worked quite a bit with Photoshop, Illustrator and a bit of Autocad and feel that it is definitely possible to get good long and cross arching. It will be a bit tedious to execute. It is carving with the software vs. the gouges and planes. The other way would be to carve (with hand tools) the top and back then probe (in high resolution) with the CNC/software.

To me the real question is how to manipulate the arching for a given piece of wood that you are working with. I always remember what Luiz Bellini told me. He said that he was not too worried about thickness. For a given top or back he changed the arching rather than the thickness.  This is where experience with understanding the wood and what you want to do with the model and sound come in. If you have a decent model, one could experiment with moving points in the software and go from there. I can say I am on the outside looking in with this problem (amateur).

[Don Noon]

15 hours ago, Don Noon said:

 I see this as a typical fault with CNC use... the shape you can conveniently model with the software is not the same shape and experienced maker would be getting by carving.

13 hours ago, scordatura said:

I am not sure I agree with this Don. If you can recreate the arch from the templates that exist for a given model, your "carving" is your ability to create the curves with the software.

My point is brought out by the video... the shape is made by dragging a profile around the outline, and ends up with a distinctive shape that I have seen from a number of CNC users. It just screams CNC.  Making a shape similar to hand-carved or from historical profiles is more difficult... but not impossible if you have the motivation and knowledge.

[scordatura]

40 minutes ago, Don Noon said:

the shape is made by dragging a profile around the outline, and ends up with a distinctive shape that I have seen from a number of CNC users. It just screams CNC. 

Agreed. I did not like the look of the top of the arch. The contour lines he created were too pinched in the center bout. The way he did the corner areas was also was concerning. In his defense, he said that he was trying to speed up the video. By the look of his videos, he also does other instruments. He may not be that knowledgeable with arching contours on a violin. I am in the process of developing a procedure. I think the way to do the outline and main cross arches is to use a function in Illustrator or Inkscape (freeware I believe) that "vectors" the bitmap. Fusion 360 does not have this vector draw feature. This should be much less tedious and a better outline if done correctly. Starting from a .tiff scan for instance is better than .jpg or .gif. 

[Michael Darnton]

Don's attitude is exactly the reason I was once considering CNC: though letting math or some other production strategy like dragging a profile around an outline will, I think, lead inevitably to the wrong thing, the CNC has incredible possibility for writing the right thing into it and then doing that 20 times identically to see what the overall result is. To my eye, some of the things I'm most interested in are things that software wouldn't do, but that's a limitation of the software and the operator, not the CNC machine. Garbage in, garbage out, as they say.

[scordatura]

14 minutes ago, Michael Darnton said:

To my eye, some of the things I'm most interested in are things that software wouldn't do, but that's a limitation of the software and the operator, not the CNC machine.

This is my point. I feel that with the right approach you CAN do what you want. It is all how you execute in the software. This is not rocket science (sorry Don) it is just another skill. For me and possibly others, time is the enemy here. I have so many irons in the fire that I have trouble getting it all done let alone striking while the iron is hot!

[scordatura]

Does anyone own a touch probe for their CNC? If so what brand, cost, etc.?

[Don Noon]

23 hours ago, scordatura said:

It is all how you execute in the software. This is not rocket science (sorry Don) it is just another skill.

It's a lot closer to rocket science (where the goal is known, and the method to get there needs to be worked out) than figuring out how to make a "great sounding" violin (undefinable goal, infinitely technical complexity in getting there even if you had a goal).

[Fiddlemaker5224]

I own a Touch plate. Makes thing Like setting the 0 height quick.

https://www.rockler.com/circular-touch-plate-for-cnc-sharks

[carl1961]

7 hours ago, scordatura said:

Does anyone own a touch probe for their CNC? If so what brand, cost, etc.?

I have a touch probe made by DrewTronics, but mine was one without the led (one of his first models 130.00) and I use the Prob-it  software for Mach3 (35.00) so far I only used it to find center and out side, but actually bought it to probe a violin plate, but never made it that far yet   picture is mine with a case I cnc carved from Plexi-glass.   

 

https://drewtronics.org/

http://www.craftycnc.com/probe-it-wizard-mach3/

 

 

 

[carl1961]

2 hours ago, Fiddlemaker5224 said:

I own a Touch plate. Makes thing Like setting the 0 height quick.

https://www.rockler.com/circular-touch-plate-for-cnc-sharks

Nice. Touch plate, I use the 5$ ones on eBayand they work great . I have a script that you put in mach3 to set offset and to do the zero, if anyone needs it I'll upload it.

[Michael_Molnar]

All I do to reset Z Is use a strip of paper to feel the endmill tip as I slowly jog it down.  I know this sounds crude, but it works for me much like working a feeler gauge.

[carl1961]

58 minutes ago, Michael_Molnar said:

All I do to reset Z Is use a strip of paper to feel the endmill tip as I slowly jog it down.  I know this sounds crude, but it works for me much like working a feeler gauge.

Michael, to me that's the best sure way of doing it. Thanks for the tip.

[Michael_Molnar]

14 hours ago, Michael_Molnar said:

All I do to reset Z Is use a strip of paper to feel the endmill tip as I slowly jog it down.  I know this sounds crude, but it works for me much like working a feeler gauge.

Zeroing Z with the Mach 3 button always works, but entering a new Z value must be followed by hitting enter (return button) to override the old value. 

I ruined a lot of wood by not striking ENTER after typing in the new Z offset. 

Duh!

 

[Don Noon]

1 hour ago, Michael_Molnar said:

I ruined a lot of wood by not striking ENTER after typing in the new Z offset. 

Duh!

My trick (so far) is hitting the "run program" after getting Z set, and then discovering that it was running the previous program and I forgot to load the new G code.  I usually "air run" a program with the Z far away from the work, but overconfidence and laziness sometimes win out.

[HoGo]

I forgot to show a rendered pic of the model I created. It's just the basic crappy render without any smoothing of the surfaces I got during the fast modeling. Shows a few imperfections in chanelling at the lower corners but I guess the roughness of the cut would hide it.

[Fiddlemaker5224]

Impressive, my question is, if this is in the rough, did you allow extra material to cut the Profile? The profile edge can be a straight vertical cut that can be finished after roughing the plate. There will always be some variance with the overhang and slight adjustments may be needed. Just a thought.

What I have been doing is trying to get a good rough blank, for hand finishing. I will hand cut the purfling in after the plates are finished.

[Michael_Molnar]

I received an email asking about how I do fingerboards on my CNC. Below is a photo of the fixture for securing a fingerboard blank. There are 3 positions running top to bottom. The fingerboard in the photo is in the second position. 

The fingerboard is secured in the first position bottom-side up. It is clamped down in the middle. The CNC drills a couple of 1/8" pilot holes about 3 mm deep on the left half. The fingerboard ends (left and right) are then cleaned up to the finished length. It is important to center the blank because some come close to the final length. If they are off center, you will wind up with a fingerboard that is too short. 

In the second position (pictured), the fingerboard is flipped topside up. The pilot holes are pressed onto 1/8" alignment pins. The fingerboard ends (left and right) are clamped tight. The CNC now cleans up the sides and top surface leaving some material for finishing once attached to the neck.

The fingerboard is flipped to the underside up and clamped down in the third position. The CNC does the scoop and puts a taper on the lower edge of the scooped hollow.

Keep in mind that this is NOT a finished product. The dirty grunt work is eliminated but does not replace the need for a good eye to finish the surfaces (e.g., the shallow dip and rounded sides) and make a good fit with the neck. The advantage of my process is that I can clean up a bunch of fingerboards for future violins. 

[Don Noon]

This thread trailed off a few months ago, when I was just getting started using my CNC machine.  A recent question made me think of it, and there are a few updates I can add based on my experience of the last few months.

1) MAKE SURE THE MACHINE AXES (especially the X-Y) ARE ALIGNED PERFECTLY!!  My machine was (and still is, for now) misaligned by 0.3 degrees.  That might not sound like a lot, but if you're using the machine to cut precise outlines of the plates, the top and back plate corners won't line up, and any bolt and tooling pin patterns will be skewed so that you can't flip over a supposedly symmetrical thing and have the holes line up.  This has been a huge pain and time sink for me, solved (sort of) by making a separate model for the back plate, with a reverse skew.

2) There may be flex in the structure.  Mine uses round rail guides for the X axis, and at the middle of the travel, the flex is the most.  With the 12-pound spindle on my machine, this can cause a sag of .1mm or so, which also translates into a slight curve in any X axis movement.  If you want something perfectly symmetrical, it may be best to set up along the machine Y axis, along the centerline of the machine.

3)  Design and quality flaws to watch out for... Yes, mine is a Chinese machine, great for the price, but not industrial quality.  Small set screws hold the linear bearings in the structure, and they can (and did) come loose with vibration.  Hand-flexing the spindle every so often will show this up.  On my Z axis, the set screws were especially tiny, and the bores for the bearings were nearly .1mm oversize.  I had to disassemble the machine, add new (larger) set screws, and add shims to keep the problem from recurring.

4)  Coordinates.  There are design coordinates, G-code coordinates, and machine coordinates to deal with.  I think it simplest to keep the design symmetrical about the Y (long) axis (X coordinates will be positive and negative), and make sure everything is positive in the Y axis.  And use the same coordinate system to generate the G code.  Then there's just the machine offsets to deal with.

The machines are amazingly cheap, but getting all the software sorted out and models and fixtures designed is a huge hurdle.  I'd say a solid 6 months or more has been eaten up by this, and I still haven't gotten around to the 3-D modelling of the plate arching (for now, I just use the machine to cut key profiles, and carve most off by hand). 

And there's an upgraded design machine on the market now, with linear guides and much more rigid... so I'm going to get one in the near future.  Then I get to see how bad the market is for my used machine.