3D CNC routing of plywood sheets

[wet feet]

Can you please tell us specifically which machines will generate a 3D toolpath from iges,sat,dxf or dwg? I'm beginning to suspect a lot more opinion and waffle than specific information coming out on this thread.I have generated 3D toolpaths on a number of CNC machines with various software packages and there isn't too much to convince me that nobody else has actual experience of such a procedure.

 

[TANSL]

What I can tell you is that I have sent files in iges format, with a 3D model of a 7 m boat hull, to a router that created, without further information, the 1/1 scale model of the hull.
I have also sent many files, in dwg format, with the nestings of the various pieces of a structure, which have been cut without more by cutting machines of 2.5 axes (those would be enough for your pieces of plywood).
I can not tell you what machines they were, their make and model because I do not know, it was not my business. My client, with my clichés, took them to his machine and cut the pieces from a flat plate. I do not know anything else, I'm sorry.
I suppose if you contact a specialized workshop (I think it's the best you can do), they will give you information about the machines they use and how they should receive the cutting information. This will probably avoid many discussions and misunderstandings.

 

[Mike Inman]

The cutting was actually done with a CNC waterjet cutter, I got a comparative quote for CNC routing but that was more expensive.

Last time I dealt with a waterjet cutter (about 18 years ago), the metalworking shop I was working with charged it out at $100 per hour, while the traditional mills, lathes, etc. were charged at $50 per hour. Not saying that waterjet might not be cheaper overall, just that it does come with some non-production time labor load to muck out the used cutting media... For thin wood, laser cutting _might_ be a similar option to the waterjet, and, of course, woodworking is different from metal. I used that waterjet to cut some kydex (plastic sheet) and it went really quickly, even as compared to 5052 0.90 aluminum sheet.

Yes, the thinking with 3D routing is to bevel the edges to get flat fits between parts joining at non-right design angles. You're right that epoxy can fill those V shaped gaps, but I believe that assembly of the parts can be a lot more efficient and accurate if they are cut to make flush fits rather than the assembler(s - including myself) expecting everything to be a bit off and never knowing if they've got the right fit or not.

 

[John Perry]

Hi Mike - even with 3D router control you wont get 'flat fits' between parts - you would have small steps on the edges of parts. Perhaps that would be a bit better for assembly than just having square edges but at the expense of more router time. In any case, I have heard of a trick to make assembly more accurate where square edged panels meet at an angle. Short length of round dowel (or perhaps short lengths of plastic bar stock would be even better) are fitted into the internal angle beween the panels and trapped in position by the wire ties (or cable ties) that stitch the panels together. This ensures that the inside corners of the panels mate - the edges cannot ride over one another. Then dabs of epoxy and glass are applied at intevals betwen the dowels so that everything stays put when the dowels are taken out, then more epoxy and glass are laid along the whole length of the seam. But my own experience is that a reasonably accurate assembly is not too hard to achieve, so I havent actually bothered to try the method with dowels.

For the small boat I mentioned I used strips of glass to strengthen the seams internally whereas externally I just applied an overal glass sheathing with no extra glass reinforcement along the seams. If you want a more robust construction with strips of glass (or glass tape) layed externally along the seams plus an overal sheathing then there is a problem in that these strips cause ridges in the sheathing. My experience with an earlier project was that it took a lot of careful filling and sanding to taper out these ridges. If you are going to use a 3D router to cut the panels then perhaps it would be worth routing shallow rebates (maybe only a mm or so deep) to accomodate external seam reinforcement without the need for too much later filling and sanding. To my mind this would be more useful than machining steps in the panel edges.

 

[DCockey]

even with 3D router control you wont get 'flat fits' between parts - you would have small steps on the edges of parts. Perhaps that would be a bit better for assembly than just having square edges but at the expense of more router time.

True with a 3 axis CNC system such as hobbyist and other lower cost systems. A 5 axis CNC system can cut plywood with bevels.

 

[TANSL]

A "two and a half" axis CNC system can cut plywood, or any other material, with bevels.

 

[Barry]

What is the difference between a 2.5 and 3 axis machine?


Response:
An axis is a direction of motion controlled by the CNC machine control. It can be linear (motion along a straight line) or circular (a rotary motion). The number of axes a machine has determines it's machining capabilities. A 2.5 axis machine really has three moving axes, but only two axes can move together (most machines sold today are full three axis machines). For machining centers, a three axis machine will have three linear axes. A four or five axis machine will have three linear axes as well as one or two rotary axes.

Note that 2.5 versus 3 axis has yet another context. 2.5 axis machining requires that the machine have three axes, but only two axes must be moving simultaneously at any one time. (Simple operations, like drilling and most milling, fall into this category). On the other hand 3-axis machining requires that all three axes be moving at the same time (More complex operations, like the machining of sculptured surfaces required in molds and airfoils, fall into this category.)


The term “5-axis” refers to the number of directions in which the cutting tool can move. On a 5-axis machining center, the cutting tool moves across the X, Y and Z linear axes as well as rotates on the A and B axes to approach the workpiece from any direction.

The above taken from various manufacturers of cnc machinery

DCockey has it right, for cutting bevels (unless it is a fixed bevel over the entire cut and the tool profile produces a bevel), you require a rotary function to tilt the tool to produce a bevel. The most common would be a tool head that rotates around 2 axis and hence a 5 axis machine

 

[wet feet]

A good summary Barry.I actually learned CNC machining on a 5 axis machine,but when I had to bevel bulkheads with a 3 axis machine it wasn't such a huge deal to use a series of 2mm steps on the bevels and just plane the ridges away by hand.It was certainly a lot faster than cutting square and having to plane it all.I will add a word of caution though,it does add a bit of time at the computer to extract the surfaces for bevelling compared to just finding the largest face and cutting square.On balance I consider it time well spent.

 

[John Perry]

Wet Feet - I wondered if you meant to say 'finding the smallest face and cutting square', at least for the hull skin panels. If you use the largest face for the skin panels the hull will end up slightly larger than you drew it with your 3d software but the bulkheads will still be close to the correct size so will be a loose fit in the hull. But perhaps thats what you intended, to leave space for the epoxy to fill?

I think this thread has beem taking a bit of a strange turn with this talk of 5 axis machine tools. A 5 axis mill with a bed large enough to take a full sheet of plywood is a hugely expensive precision machine which would seem rather extravigant for building a plywood boat! Although I havent operated CNC routers myself, I would have thought that if you really want to bevel the edges in small steps a basic 2D router could do that by adjustment to the cutting head, or the cutter, between passes. And if you were to use a ball ended cutter then I would have thought the steps could almost blend out.

 

[wet feet]

I typed what I did because it is relevant to the topic of bulkheads in the post.If you cut the smallest face,there will be gaps.If you cut the large face and bevel by hand you have no gaps and a better surface to fasten or bond to.We have no idea what size boat is under consideration and if it is anything bigger than can be easily lifted the cost of epoxy will soon exceed the cost of the ply.For something like a thirty five foot hard chine planing hull the bevels in the bow can be 25 degrees or so and you can be left with a huge void to bridge with your expensive epoxy.

Your suggestion of using a ball end cutter is in fact how complex shaped surfaces are machined.The negative aspects are that it can take a long time to machine the surface with a suitably fine stepdown and then when one arrives at the bottom of the edge there will be an almost semi-circular groove in the spoilboard of the machine.This will sooner or later compel you to replace the spoilboard and skim it flat and will cost you for both the time and the new spoilboard.My suggested solution of cutting steps with a flat bottom cutter is a compromise that should leave a cleanl spoilboard and panels with a minimal amount of finishing without an excessive amount of machine time to pay for.

 

[Mike Inman]

I actually learned CNC machining on a 5 axis machine...

So much depends on what's available at the shop you're at... my first spur of the moment machined part was a quick belt drive pulley - could have been done with a 2 axis machine in a pulley making configuration, but what they had available was a 5 axis machine, so that's what it was made on.