SG

Thank you for sharing that video , I had searched for one like that with almost the same keywords but never saw that .

That is really what I am tying to do . The video made me realize the serious amount of tooling required. Along with the extra amount of material needed in order to hold it down . I suppose I thought I could get away with just a cutter for the sticking and one for the cope (like we would do traditionally on a shaper). And maybe use backer blocks , but really had no idea how to hold them . The video you shared is perfectly clear to me .

SG can you say wether or not you found it as a time saver or worth doing this way ? The tooling is expensive , and after watching the video I have my doubts on its efficiency.

Richard , the doors I do now are done more like a carving . A wood panel is laminated and placed in the door during glue up . The whole door sits on the CNC for this process .if I chose , I can flip the door over and route out the back so the lights are divided .

My programming is simpler, just using the cope/stick heads I run on my shaper (they were made by Royce/Ayr to run on both CNC/shaper, with aluminum heads, 4" dia., custom arbors/spacers). They were fairly expensive (but a good deal for what they are - they were the most competitive by a good bit), but the versatility of being able to run them on both machines has proven to be worth it. I end up having to make the stops separately, but that's easy enough. Don't know how large/heavy your shaper tooling is, but if it isn't too huge, you might be able to put it on one of the HSK arbors Tecniks makes and use it on the router. Having a set that will do both makes it an easy process, and using something cheap, like poplar, as a backing material, doesn't add much to the cost. Doing simple radii and such on the shaper isn't so hard, though of course is much more precarious, but the beauty of the router is the unlimited nature of the possibilities, as well as the ability to run the tooling with the grain consistently (by dividing up the tool-path and running with the radius on all quadrants).

The video shows a great way to make muntins, and the coping would not be difficult in that case. Not very fast, but about as safe as can be. All the skills move to the CAD draftsman, and I would think they are as knowledgeable in cope/stick munition work as millennial. As in, not much.

Glueing to particle board with paper is promising. That is an old turner's trick to make half turnings. Once complete, insert a chisel on the edge of the paper and split the part in two.

As with anything in CNC, it is in the programming. Disregarding the running times on the CNC, how long would it take to draw and program the elliptical arch window in the above photos?

One need not be a specialist to do CAD/CAM work. I don't know how familiar you are with the process, but simple lines representing the geometry is all that's required. A good tutor could show you most of what you need to know to program 80% of what you'll ever want to run in two days (as I've done for a couple of other shops in my area). It does require astute and careful work, but once you get over the first hump of the curve, it becomes a quick and logical progression. I think what's confusing on the face of it is that the software companies want to show you everything their product can do down to organizing your sock drawer, while the reality is that almost everything you'll ever run will be represented by a few lines, easily generated by using an arc, circle, ellipse (or just straight) line commands. It's from these you'll determine cutting heights, direction, feed/speed, etc. There is a great deal to learn about running CNC over conventional equipment, to be sure, but that's part of the fun. Having a deep understanding of how things are best machined goes a long way toward getting the most out of CNC, esp. when cutting solid wood.

My 2D CAD skills are good, but nothing extraordinary - I would think 30 mins. to draw and an hour to program. As mentioned previously, I like to divide radius work into halves or quadrants and run with the grain at all times. Of course you're still fighting the rotation, but the result is better than if it's ignored and run in one direction only. If you were running something that was difficult and needed to be perfect, you could even flip the head over (in the case of bore-type tooling) and reverse the spindle. I've not done this, but it's totally feasible.

If you do want a model this does not take as long as you would think if you have the correct software. I find it helpful to see the parts, but some do not need that. I was curious how long it would take to model David's window, so I gave it a try. In about 45min I am well on my way. It is not perfect, but with a proper set of drawings showing some key dimensions it would be. I am far from being a millennial, but I do have experience 3D modeling.

I think trying to figure out how to cut David's parts would be a fun challenge. The first time would be a hair puller, but after that I bet you could copy a lot of the strategies to other projects. I suspect 4 - 6hr would make a good dent into creating the tool paths since many are mirror images, but I forsee some time making fixtures that could add some time and headache.

Interesting discussion.

View higher quality, full size image (1713 X 927)

View higher quality, full size image (1679 X 853)

I think it's important to underscore for those that don't have CAD/CAM experience that a 3D model is a nice visual aid, but is irrelevant to the process.

They are also not difficult to generate if one wants to have the reference, and I often do so if I have any doubt as to the part intersections or the like.

Mr. Sochar's example has only three radius parts (Brad has complicated it a bit my making one of the radial muntins radiused, which in addition changes the length of one the short decorative muntins) - the outer frame is one, the individual decorative muntins between radial muntins is two, and the continuous (though manufactured in parts) elliptical muntin that is a simple offset of the outer frame is three. I'm assuming the straight parts would be manufactured on the shaper. Which is not to say this is a quick and easy job, but there is a lot of symmetry which would make the programming faster than one might think. Actual manufacture, including fixturing, would take some time, but I think I could draw, program and cut the parts, excluding assembly, in a 10 hour day.