3d roughing question

[Chuck Keysor]

Hello Shopbot Friends:

In designing the 3D cut files for a large Victorian house decoration, I began to wonder, how do you determine how good of a 3D rough cut you need to make? (Conversely, how crummy/quick of a 3D rough cut can I get away with?)

I ran one 3D rough cut simulation for this big part with a 65% step over, (and .25" per pass) and the simulation said it would take 2 hours and 13 minutes. The 3D finish pass took 3 hours and 22 minutes.

Then I ran the same 3D rough cut simulation except with a 30% step-over, which raised the simulation time to 4 hours and 25 minutes. I didn't change the parameters of the 3D finish pass, and it's simulation time remained unchanged.

I took screen shots of both sets of 3D finish simulations, running at highest modeling resolution (7x slower) and I carefully compared them. I couldn't see any difference in the end 3D finish cut. IE, it appears by the Aspire simulator, that the 3D rough cut does not even affect the end quality of the 3D finish cut.

So why would anyone bother to run a 3D rough cut with any detail? And how big of a step over can you go on the 3D rough cut before something does show up as being fouled up in the 3D finish cut? How can you tell by simulation that you have made too rough of a 3D rough cut? Or does the simulation even reveal problems related to the 3D rough cut, that do show up in the real cuts made on my Bot?

Sorry, lots of little questions, but they are all related to the nature of figuring out how to correctly choose a minimum 3D rough cut strategy.

Thanks, Chuck

IMPORTANT PS: When I ran the actual part, I kept the big 65% step-over, and the wood was rather splintery, and in places, I imagined that with a smaller step-over, I would have had less splintering of my stock. But none of the splintering was so bad, that it busted through "machining allowance".

[adrianm]

The important thing to realise is that the finish cut is always to the full depth required to finish the model in ONE PASS. It does not use the pass depth specified in the database.

It's just something to be aware of if you leave too much material behind on the roughing pass the usually delicate finishing tool can have too much to do and snaps.

If you're cutting in something like foam you can often get away with skipping the roughing pass entirely and dished models can often be cut with a finish raster path without using a roughing pass.

[Burkhardt]

For very deep carvings I rough only with a step depth of the finish bit flute length minus the skin. Even if the finishing bit is very fragile, the stepover is usually so small for finishing that it can handle a full flute length cut.

For shallow carvings than don't go deeper than the finish bit flute is long, I don't rough at all.

However....I make sure that the very first cut starts outside the wood in air or is above the surface because cutting that first full width/depth trench will likely break a slim bit.

DeskProto has more options than Cut3D to define layered finish cuts or to do a rough cut in a narrow strip as as starting slot.

If your model has deep holes or pits with steep walls that are smaller than the roughing steps, the roughing pass will ignore them. You are then running the risk that the finishing bit will plunge at full speed into the wood in these locations. I broke a few bits that way but don't know yet how to avoid.

[khaos]

... However....I make sure that the very first cut starts outside the wood in air or is above the surface because cutting that first full width/depth trench will likely break a slim bit. ...

Don't miss this KEY info.

[Brady Watson]

Don't miss this KEY info.

Ed Zachery.

Do a profile pass before or after roughing, before your finishing pass. Leave some meat by not going full depth so you have one continuous tab of sorts around the whole thing. This saves cutters & your sanity - since it will reduce or eliminate that horrid screeching noise at the end of each raster pass. If the part is conducive to it, I carpet tape or vacuum it down & then remove the scrap material from around the part before roughing and finishing. This also saves tools as long as your hold down is rock solid.

It isn't uncommon for me to run 4 or more toolpaths on a 3D part in order to get the quality and finish I want. An extra 2 minutes of cutting to clear out for the finishing makes a big difference. Additional toolpaths are pretty much manditory in denser materials in order to keep the tip on the tool.

-B

[shilala]

I just wanted to comment on the end of your post, Chuck.
Soft wood will get butchered pretty bad without reasonable feed/speed. So will hardwoods. It likes to chip rather than pile up.
I keep stepover under 20% all the time. But I also keep the depth of my model within boundaries that I've found work best in different woods.
I always figure that saving an hour of machine time isn't worth losing the whole project. It always takes a lot more time for a do-over.

[Chuck Keysor]

Thank you everyone for providing very useful and educational replies to my questions on 3D roughing!

Adrian, I had wondered about the 3D finish cut pass depth, as there is no setting for that in the cutting parameters. That observation actually sent me astray in my thinking as to what might be going on. So your concise statement that Aspire will calculate the 3D finish cut in one pass without regard for how much material is above the finished surface was very clarifying for me!

G., your suggestions on when to make a roughing cut, and how deep to make them was also very helpful!

Thanks Joe for emphasizing the key bit about having the bit start its cut in air.

Brady: I assume Ed Zachery is Japanese English for exactly,,, I scratched my head the first time I read that. The second time I said , Ah so!
Your suggestion of doing a profile pass makes sense to avoid the above issues. And I had noted the screeching of my bit at various moves which made me a bit nervous at first.

To your point of multiple finish passes, I first did a raster finish pass, moving with the grain, over the entire piece at 1.5IPS (figuring that since that is the fastest I can move without problems in my V-carving, why should 3D finishing not have the same limit?) with my 1/16th inch tapered ball nose end mill, at 10,000RPM, with a step over of 32%. That took 4 hours! That is why I chose such a big step over. Then I went back over key areas where hand cleaning up would be most difficult, and did an offset finish pass. That was fascinating to watch, and amazing to see how precisely the bit tip lightly touched the already cut surfaces, barely leaving faint marks. Effectively my step-over was much finer with the two usually non-parallel tool paths. I spent 7hours and 45 minutes doing finish passes.

Scott, my wood was pretty butchered! The wood is from South America, and is called heura crepitans (I have no idea what the spelling is). I had thought of shellacking the entire piece after the first 3D finishing pass, to help with the fuzz, but decided not to, as this was a first time type of project.

Coming:
There are 4 separate carved Victorian "appliques" in this project. I will post pictures of this piece, along with one of the other 4 appliques that I cut out of blue Styrofoam with a request for material suggestions. The blue foam cut great, the wood was really crummy. I'll keep and use the wooden one I am now finishing by hand. But that leaves 3 more to go, that I could make out of other material. I will save that though for the next post.

Again, thank you all for your greatly appreciated replies to my questions! Chuck

[khaos]

When I 3D finish my stepover is 4-8%. Granted it does take longer.

[shilala]

Chuck, that stuff sounds super soft. Like down around 600.
I keep a pretty comprehensive list of wood hardnesses. I choose my wood for projects based on those hardnesses, or "pressure to mar".
I find that I like to work with woods that are around 900-1100. They tool really nice for 3d applications. Once I get into red oak and things that are hard and open-grained, the chipping is awful, unless the 3d features are nice and smooth and flat and rounded.

My point is, if you can select your crazy south american lumber with hardness and shapes in mind, you'll save tons of grief. The cool thing is that lots of South American woods are in the mahoghany family, and the range of hardness goes from moosh to nice, and lots of them look the same.
So you can easily match a look, even if you don't use the exact type of wood a customer requests. I find that folks will easily defer to a suggestion, especially when I let them know "it'll look much the same in the end, but this wood will give you a much nicer piece".

I find that if I match the right wood with the right feeds and speeds, I can get done a lot quicker, and I don't need a bunch of sanding and lacquer to fix things up.
I'll do damn near anything to avoid sanding.

[Brady Watson]

To your point of multiple finish passes, I first did a raster finish pass, moving with the grain, over the entire piece at 1.5IPS (figuring that since that is the fastest I can move without problems in my V-carving, why should 3D finishing not have the same limit?) with my 1/16th inch tapered ball nose end mill, at 10,000RPM, with a step over of 32%. That took 4 hours! That is why I chose such a big step over. Then I went back over key areas where hand cleaning up would be most difficult, and did an offset finish pass. That was fascinating to watch, and amazing to see how precisely the bit tip lightly touched the already cut surfaces, barely leaving faint marks. Effectively my step-over was much finer with the two usually non-parallel tool paths. I spent 7hours and 45 minutes doing finish passes.

Chuck,
Every relief is different in terms of what stepover and diameter tool I choose for finishing. I absolutely hate sanding & would rather let a robot do as much work as possible than for me to aggravate my arthritis. Typically I run 8-12% on a 3D finishing pass, with 6% being the lowest & 20% being the highest - unless I was doing the 3D finishing pass as part of a roughing strategy. I may run 35% SO with an 1/8" ball before finishing with a 1/16" @ 8-10% SO. It is highly dependent on the material and the relief.

Simulating the 3D toolpath is your friend. In most cases it will tell you what bit you want to use by running a few different toolpaths selecting incrementally larger or smaller tools as you go. There is certainly a diminishing return when running smaller tools on larger reliefs & you have to weigh whether or not it is worth it to you.

I'm not sure what relief you were cutting, but it may have made more sense to finish the entire relief with an 1/8" or 3/16" ball & then go back over areas you isolated with a vector to machine those areas with the 1/16" ball. In either case, you'd probably want to run 10% to get a nice finish. Just remember to Zzero at the EXACT same spot you zeroed out your first tool on the material. I just write the location on the material or spoilboard.

Got any pics of the piece in question you can post up?

-B


PS - those small cutters like high RPM. Crank that spindle up to 18,000, especially in soft materials and watch your quality go up. I routinely run 18-20k with tools 1/16" and less.