First project - Hexagonal staircase newel cap

Tool(path) time :slight_smile:

Repeat warning, these feeds, speeds etc. in the toolpaths are based on what I would run, whilst using my ears, feed rate override and small amount of common sense I posess to try not to break the machine, cutter or workpiece. Don’t assume any of them are ‘good’. We should also expect that I have made at least one stupid mistake here that I have not seen in toolpath simulation but would only see when running the job in the real world so start on some cheap stock. Run the Fusion simulation for all the toolpaths before you send them to the machine so you know what they’re meant to do before pressing START

I’m also sure there’s better ways to do this and will not be at all offended when somebody else points out better ways or problems with this one.

First up, I made a new component for the base hex stock, flat this time to let the CNC do the work, you’ll get to play with how deep and fast you can run an adaptive clear on the Pro :pick:

The stock is just an extrude up and down of base hex outline, offset by Hex_Stock_Extra parameter

It then extrudes down to the base with total height Hex_Stock_Height from the parameters.

The same MDF flat base is in there (depth doesn’t matter as we set Z zero on top of the MDF) with X, Y zero set in the bottom left corner where the bitzero likes to do things. I have attached the various other components to the main hex with an as-built joint to make sure things don’t move around.

The first machining setup is “Main Hex”, this uses the MDF base, sets the Main Hex Stock component as the stock and the Main Hex as the body. If you leave the main hex on this board attached to the Shapeoko until the end the rest of the alignments should work OK. If you take it off you’ll have to be sure the rectangular board is square to the machine both times.

First job is a big adaptive clear. If your workpiece is well held down on the Pro you can probably push faster or go to a deeper DoC, your ears will tell you once you start, use the feed rate override to test going slower & faster. It’s a long job so it’s worth tweaking to see how fast you can go.

Next up it’s the 1/8" ballnose to clear out the slots for the wedges, this is done in two steps as the 1/4" didn’t clear to the end of the slots and the depth of engagement might be a bit much for the 1/8" bit. First pass is at 2mm stock to leave

Second pass is the whole length with no axial stock to leave, both have radial stock to leave to stay off the walls.

Last op in this sequence is a Pencil toolpath down the side walls of the wedge slots to get them straight and clean using the 1/8" flat ended cutter. This will not quite give a square corner due to the slope, shaving a bit off the bottom edges of the wedges should fix that and those are hidden anyway.

So now we’ve got the main hex roughed out and the wedge slots cut.

The wedges need machining too. First is the wedge stock component, this has two bodies, one is the starting stock to cut from, the second is the part-machined wedge ready for glue in to the hex. There a sketch with the wedge outer contour projected from the first wedge body and Wedge_Stock_Flat_Extra around the edges of Wedge_Stock_Thickness.

Inside that I then borrowed your trick of slicing planes which are offset Wedge_Stock_Face_Extra above the flat faces of the finished wedge to chop the stock body down to the machined item we want to glue in.

It seems over complicated, and it probably is, but it seems to work.

The Wedge From Stock CAM setup has the wedge stock flat, Z zero at the spoilboard (or whatever you mount the stock on) and X axis along the wedge. I assumed that you’d probably have the grain running along the long axis of the wedge.

There’s then an adaptive clear

Followed by a contour around the edges, which has tabs in it, you may or may not want those depending on whether you’re blue tape & glue or just clamping down a larger bit of stock.

So that leaves us with a roughed out wedge with some extra stock left on the face, tip and root for glue in, hopefully.

Once the wedges are all glued in we can go back and do the rest of the machining (as you described earlier). This setup has the same rectangular board and zero point, but this time all the wedges are included in the body selection, but not the center post.

Stock is another solid chopped down with cutting planes offset from faces by Wedge_Stock_Face_Extra to give us the extra from the glued in wedges.

First set of CAM paths on the 1/4" straight cutter. We clean up the center post hole with an adaptive clear (there’ll be some air cutting here), a contour around the inner edges and a parallel along the bottom.

Then we run around the outer edge with a contour

And then we do a ramp with a little stock to leave to rough cut the hex and wedges down close to final dimensions. This one might be fun to watch.

Then it’s the 1/4" Ballnose to do the finishing pass on the main hex and wedge faces, play with the stepover on this one to adjust surface finish, or use a larger diameter cove cutter.

Finally the 1/8" flat cutter to give us a small radius on the inside corners for the center post.

It’s hand sanding now.

Last setup is the Center Post, there’s a stock object of Center_Post_Stock_Size and Center_Post_Stock_Height.

Simple setup of the stock and body with X, Y, Z zero in the bottom corner

1/4" flat cutter to do an adaptive and contour

Then the 1/4" ballnose to ramp round the top to reduce the sanding for the top pyramid part.

Here’s the F3D file, have a poke around and let me know if you’ve got questions.

Large Newell - Another Way v11.f3d.zip (1.8 MB)

2 Likes

Wedge Tip Depth and Wedge Root Depth define the thickness of the wedge at both ends. I wasn’t sure exactly what decisions you’d made in projecting the wedge but it looked thicker at the root than the tip so I copied that.

Mostly I was just happy to get some kind of plane in place. Ideally I would have made the wedge a uniform thickness, because that’s what traditional woodworkers would have done, but as long as the tip thickness is enough to cover the “Flat Base Thickness” with some to spare it’s fine.

The wedges are slightly oversized for glue in to be levelled in the final toolpath. I think that’s what you meant?

Exactly, thanks.

The main issue to me is the loss of rapids in the free version.

OK, something to be explored later.

Next up it’s the 1/8" ballnose to clear out the slots for the wedges

OK, I was going to ask why a ballnose instead of another end mill and then I realized that the toolpath can use the tip of the ballnose to get slightly deeper into the narrow part of that pocket (which is 2.5mm at the thinnest (aka WedgeTipWidth).

This will not quite give a square corner due to the slope, shaving a bit off the bottom edges of the wedges should fix that and those are hidden anyway.

Yeah, this is pretty typical SWP (Standard Woodworking Procedure).

There a sketch with the wedge outer contour projected from the first wedge body and Wedge_Stock_Flat_Extra around the edges of Wedge_Stock_Thickness.

I’m having trouble visualizing this, maybe I should play around with the numbers in Fusion.

Followed by a contour around the edges, which has tabs in it, you may or may not want those depending on whether you’re blue tape & glue or just clamping down a larger bit of stock.

Yeah, now that I see them, having tabs around the wedge should be OK since they don’t project all the way up to the top surface, so I can be aggressive cutting them off without affecting visuals.

Once the wedges are all glued in we can go back and do the rest of the machining (as you described earlier).

I’ll want to run the center hex pocket clearing when half the wedges are in to clear out extra material from those wedges, which would interfere with the other half of wedges being put into place.

Then it’s the 1/4" Ballnose to do the finishing pass on the main hex and wedge faces, play with the stepover on this one to adjust surface finish, or use a larger diameter cove cutter.

Interesting. I’m now paying the penalty of previously buying 1/2" shank router bits when I could, as all my ball noses are 1/2" shank, even for smaller sizes. There is this cheap 1" diameter bit: 1/4 Inch Shank Cove Core Box Router Bit 1 Inch Cutter Diameter Woodworking Cutting Tool (Red): Amazon.com: Tools & Home Improvement and Rockler has a ¾" bit here: https://www.rockler.com/core-box-router-bits-1-4-shank Since it’s just for the finishing pass, do you think I should spend $10 and try the cheap one out? EDIT: I see Amana has a $50 version, too.

Finally the 1/8" flat cutter to give us a small radius on the inside corners for the center post.

I’ll want to run that 3 times: First before I glue in wedges, second after half the wedges are in, and third in this final pass as you indicate.

On the Center Post, I was thinking that the zero point should be the top center of my stock. This way it doesn’t matter if the stock is a bit thicker or thinner and it looks easier to find the center of the stock than to find a certain offset from eventual corners, but I guess it does mean I will have to trim to length manually afterwards. But, I’ll have the extent of the side milling to show me where to cut, and that cut is hidden in the final piece anyway.

Again, huge thanks - I hope this is fun, maybe even a bit of a challenge for you.

2 Likes

No problem, yes, it’s a fun problem to work out how to machine and folks here help each other. Be sure to post pics of what you make too :wink:

OK, good, the shapes should update when you change the parameters, just be sure to regenerate the toolpaths afterwards.

Yep, that too, I hadn’t realised that part. I chose the ballnose because it’s a sloped face inside the wedge cutout and ballnose bits tend to work better for that.

I have fallen into a pattern of roughing out with the largest cutter I can run, I frequently use a 1/2" no helix worktop cutter in my spindle to rough out with, it’s an animal. Then switch to a smaller, shaped cutter to do the detail. That’s really important on Aluminium.

OK, this is one of those things that’s not really intuitive, until you know how to do it. It might just be a dumb way to do it and there may be a much faster way I don’t know too.

What we’ve got is a sketch in the wedge stock component whose sketch plane is set on the base of the first wedge, which is the Wedge Infill component.

I moved into the wedge stock component and made that active (dot to the right on the left hand browser menu thing). I had the wedge infill component visible too and right clicked on the flat base of the wedge infill and said “create sketch” so I had a sketch whose plane matched the wedge.

I then hit “P” to project and clicked around the edge of the wedge finding the bits that stuck out the furthest to get those purple projection lines into my sketch.

Then I just drew around the purple lines from the wedge, told Fusion that the lines were parallel to the lines from the wedge and inserted a “D” dimension which is Wedge_Stock_Flat_Extra, basically how much wider and longer the stock ought to be than the final wedge shape.

I then extruded it upwards by Wedge_Stock_Thickness to get a piece of stock that we could tell Fusion to cut up with toolpaths.

Yes, play around with the numbers and see what changes, if you select the component on it’s own then it’s easier to go back and forth in the history and see how it happened, you can right click on each feature in the timeline to get to the config dialog for that feature too.

Ah, yes, forgot about that bit :frowning: You probably just want the adaptive clear toolpath from that group as that has 0.5mm stock to leave.

I have used the Trend cove cutters which are quite cheap (Rockler sort of price over here) and perfectly good. I generally need little excuse to add to my cutter collection…

Once you buy a new cutter you also get to learn how to create the new cutter in Fusion too, so win-win :wink:

Yep, there’s some interference inevitable but nothing a hand saw and chisel can’t sort out, I couldn’t think of a clean CAM way to do it as we don’t know exactly how far down the wedge slot the infill will slide.

Yep, that works, the big thing about choosing zero point is to know why you chose that zero point and remember where it is when you set up the job on the machine. I follow a suggestion I got here which is to put “Front Left Spoilboard Zero” or similar in my .nc file name as a reminder.

As an alternative to trimming to length you could still zero Z at the spoilboard and just change the parameter Center_Post_Stock_Height to whatever you measure and just re-post the toolpaths.

Good luck machining :wink:

2 Likes

Real quick - what would you think about using the #112 .0625" Flat Cutter to get a pass that gets into the small part of wedge slot pocket?

Didn’t realise you had a smaller cutter, give it a go in place of the 1/8th ?

Just be careful about how deep and how fast you send a small cutter or you’ll have two parts of a small cutter…

I don’t have the smaller bit, yet, but I’m going to order one or three (I hear they break easily).

@LiamN, I’m going through you file step by step to understand how you constructed everything. My first question is you lofted the main hex up to half the center hex opening? Right after lofting you extrude the 25mm center hex from the bottom up as a cut, so why start with a 12.5mm center? It ends up that the height of the resulting piece is under the 38mm initial location of the top hex. You can see the top hex floating here:

It’s not a big deal, but I am wondering why you did it this way? I do realize that if you lofted to the actual opening height then due to the way the two hexes are rotated to each other the extruded cut wouldn’t quite work since some of the cut is slightly below the top hex. That means we can’t have the top hex be the actual size of the extruded cut hex, but why not choose something like 80% instead of 50%? Is there something mathematical going on?

1 Like

Nope, there was no clever maths going on. I thought about doing the trig to just set an angle for the pyramid wall and have the height adjust itself, but then realised that I didn’t know whether you had height or angle as your requirement measurement.

I realised as I was copying your shape that the requirement was for the center hex to go up far enough to cut through with the 25mm hex and give the sawtooth outer contour you have. In order to keep things parameterised I took the center post hex size and used half of that. When you set dimensions with parameters you hope to be able to change it’s best to stick to the parameters instead of fixed dimensions. There’s no good reason for 50% other than it was easy at the time.

It might have been neater to loft to the center point and not have that upper hex at all looking at it now and that’s how I did the center post as that needed to come to the full point.

If you wanted to change the top hex then you might need to do something a little better thought out in the sketch that sets the wedge cross section as that projects the top hex and uses it as anchor points to make the wedge move with the other dimensions.

OK, well, I changed the hard-coded 50% to 80% and am still stepping through the file. It would be great if Fusion had the ability to add comments to each step. I’m adding comments to the parameters as I figure them out.

I’m on this step, which “fills in” the center hex opening:

It’s an “Extrude2” which I believe takes the center hex sketch and extrudes it in two directions. I get the extrude down to object, which is the bottom of the main hex, but I don’t understand the upper object to which the center hex is being extruded. It looks like it’s 26mm away. I also don’t understand why 3 objects are being extruded - isn’t it just the one center hex drawn on the base plane? I totally get the result here, but don’t understand how it gets that result. And if I wanted the top hex pocket to be more shallow, there doesn’t appear to be any parameters I can change, even though I thought I had “Center_Plug_Depth” to control that (there’s no parameters being used in the dialog, right?)

1 Like

Yep, a double sided extrude just saves one extra operation.

The extrude target object is the construction plane set up in the previous step, which is “Center_Plug_Depth” down from the “Hex Top” construction plane which is the plane that the Hex Top sketch is drawn on, “Hex_Height” above the zero Z plane.

I put the construction planes in to make it a little easier to see where things were going to end up, rather than hiding those params in extrude distances or similar.

That’s the downside of putting several things in the same sketch to keep them easily aligned. Anywhere that a solid sketch line (not a construction line) crosses through a shape you get a separate selectable area. The lines which are used to project the wedge cross-section run through the center hex here.

The hex is broken into 3 parts which need selecting together.

Center_Plug_Depth is exactly the parameter you want here, but it moves the construction plane in the previous step. I just checked and yes, it does seem to work :wink:

OK, I walked through the whole file step by step and believe I understand each step, although some of the details surprise me (like how the wedge stock operation expands on multiple faces), but I can learn about that later.

I decided after all that I should do some main hex raw stock trimming on the tablesaw after all. One concern I had was that with the 3.175mm edge thickness, I’d have a hard time getting exact cuts with that edge riding on the tablesaw, especially with a throat plate that wasn’t extremely tight to the blade. The answer there was to cut the raw stock so that the resulting raw edge thickness was twice that.

I therefore changed the first raw stock creation operation to extrude not just down from the base plane by Flat_Base_Thickness, but to be symmetrical, so it also extrudes up that same amount, resulting in a 6.35mm thick edge in the raw stock.

Then I changed the extrusion up to the top to have a taper, and made that taper angle a User Parameter. I ended up choosing 65 degrees which cuts away a lot of material yet leaves more than enough material that if my tablesaw angle is off I’m not near the final surface:

BTW, I didn’t see any ill effects from changing the top hex target for the Lofting operation to be 80% of the actual center hex. On the good side it meant the actual height after cutting the center hex pocket was much closer to the Hex_Height variable.

Another question: how much tolerance is there in the raw stock sizing? In other words, if my raw stock is, say, one or two mm larger than I’ve defined it in Fusion, will that cause a problem or are the toolpaths tolerant of such deviations? If they’re not, then I’ll be sure to cut my raw stock slightly undersize, but if they are I’ll just go for it.

I seemed to have messed something up with the Center Hex milling:


Which is weird, but I check Liam’s original file and it’s correct:

I’ll start poking around to try to find it, but here’s my current file:
Large Newell - Another Way v17.f3d.zip (1.7 MB)

Great, sounds like your Fusion learning curve is shallowing out then. My model is by no means exemplary and much of CAD CAM is down to the preferences and biases of the person doing the modelling, there’s hardly ever a ‘right’ way to do anything.

As for the stock, sounds sensible and it looks like you’ve updated the stock sensibly to tell Fusion what it needs to cut away. I prefer to err on the side of telling Fusion the stock is larger than it really is if the stock isn’t precise. A bit of ‘cutting air’ is nicer to deal with than an over-deep toolpath which breaks a cutter or pulls chunks out of the stock. Those are going to be the first loop or two in the adaptive clearing and then everything will be the size it’s been roughed down to and the other toolpaths will be fine.

On the wedge stock, I think the bit you’re asking about is how the first sketch profile for the extrude was done, I’ve gone back through the steps to show how it was done (again, not an exemplar, just one way to make it work).

First, a new sketch on the base of the wedge, select into the wedge stock component, right click on the base of the wedge and select “create sketch”

Then hit “P” for project and click round the extreme edges of the wedge to project them into this sketch plane (project is one of the most useful tools in CAD IMO)

Now we have the purple projected lines, these are linked to the wedge object and will move if it changes dimension or position, that’s why I like project.
Draw four lines around the purple projected outlines to give the rough shape of the stock we want to machine away.

Constraints next, set Horizontal/Vertical on the two vertical lines, that locks them vertical, then make the two angled lines parallel with the purple wedge projected edges.

The lines are still blue, which is bad, it means they can move without us knowing, so now we apply the final constraint necessary to lock them down and make them go black. Go round with “D” and set Wedge_Stock_Flat_Extra as the distance outside the wedge for our new stock outline

When they’re all done the full outline goes black and we have a fully constrained reliable sketch to extrude a body from which has 4mm of extra stock around all four sides to machine away in the CAM.

HTH

1 Like

Aha,

This is a really easy trap to fall into, happens to me several times on every project, I really think Fusion could do some UX improvements to warn you about this.

What’s happened is that you’re not at the end of the history and the Component “Center Post Stock” does not exist yet, which messes up the CAM setup which uses that as stock to machine from.

Go back to the Design tab, select the root component and hit the ‘end’ button on the history controls at the bottom to get to the end of the design history, then go back to Manufacture and the setup should magically work again.

1 Like

Well, maybe.

I just got through mounting my MDF test blank and realized that it’s smaller than it should be because I made it when I wasn’t trimming everything on the Shapeoko. So I went to change Hex_Radius from 128.5mm to 128.0mm and got this warning:

I went back to your file and of course the same change did not generate a warning.

The error message is somewhat confusing because it doesn’t use the user naming. I think “Sketch 1” is the “Base Plan” sketch shown above. When I rotate it I think that “Face 1” is the bottom face, which is in red, right?

How should I go about debugging it? I think the file I uploaded last will have the same error, but if not let me know and I’ll upload my latest.

Ah joy, the Fusion lost references hole, your wax on / wax off experience is almost complete now :wink:

I opened up the v17 you posted and the error was already there, the dreaded yellow blob on the history line. This means that Fusion has lost some of the information which defines that sketch and is just remembering where things were, meaning they’ll get out of sync with any changes you make afterwards.

You’ve already worked out how to see the error list, sketch plane is first, the references lost may be just due to the sketch plane being lost or they may be separate.

First, look at the sketch and try to figure out which surface, origin or construction plane it was previously on. I believe this was set on the extruded base of the main hexagon body.

Right click on the yellow blob and “Redefine Sketch Plane”

The go select the plane (in this case the base of the main hex) that the sketch should be drawn on

Say OK and then see whether you still have a yellow warning. I didn’t, fixing the sketch plane also fixed the borked references.

I decided to go back to Liam’s original file and re-make my changes slowly and deliberately. So far, I haven’t encountered any missing references…

OK, the redo of the changes took without issue.
Regenerated the toolpaths. Feed speeds for the #201 cutter look OK for hardwood, as recommended by Winston.

But, the first adaptive clearing pass is going to take 53 minutes. And that’s with cutting off major parts of the raw stock on the tablesaw first!

Here’s the Generate Params:

Spindle Speed (which doesn’t matter in the file since it’s set manually at the router)
Cutting Feedrate (1500mm/min for Hardwood; Winston says 1525 = close enough)
DOC (Depth of Cut) is, I believe set on another tab:

But, I’m not sure which parameter (s) are appropriate, or match the 1.015mm max DOC.

I was going to use the Hardwood settings for the MDF practice run, but now I’ll probably just hit the “increase feedrate 10%” button 5 times for that practice run.

Thoughts?

1 Like

Winston’s feeds and speeds are a safe starting point, you can probably go quite a bit faster in MDF, you will probably go slower in hardwood than MDF.

As you say, use the feed override and your ears to figure out how fast the machine is happy to feed in the stock you’re cutting, wider, faster cuts are generally better than deep cuts on these machines.

I frequently listen to the first few passes on a long job like this, try the feed override and then update the CAM settings, that way the 2nd, 3rd etc. can go at the better speed.

Quite a bit of the time taken in this job is walking up the angled face at the Fine Stepdown increment to leave approximately Stock to Leave (axial and radial) for the next pass, if you increase that to, say 1mm you should shave off some time, we have another full face roughing cut later anyway.

What are the reasons, beside tolerating the noise, for not running my Carbide spindle at something higher than 20K? It goes up to 30K-ish, right? Only running ¼" and ⅛" bits…