A new 3d carving with a new problem

[Chuck Keysor]

Hello Shopbot Friends:

I have cut another Victorian house "applique", which is shown (still attached to the wood blank) in the first picture.

In the second picture, you can see that by using a 3/4" core box bit, running at 10,000 rpm, 2.5IPS and 9% step-over, and a .25" "boundary vector offset" for the 3/4" core box bit 3d finishing pass, I got a pretty smooth surface (the small ridges sand away easily) which is vastly better than on my previous applique, (where I used a 1/16 tapered ball nose end mill to do the entire 3d finishing tool path). With this large bit, the first pass 3D finish cut took only 45 minutes (on my previous applique, using a huge 30 something percent step-over, the 1/16" tapered ball nose end mill took 4 hours and 4 minutes!)

I did a visual comparison of the 3D model versus the simulation of the cutting results of the 3/4" core box bit, and then manually selected those detailed areas where the core box bits clearly had not been able to get in to cut. This was very simple, and much quicker than using the "Rest Milling" technique as shown in the Aspire training video. Then I was able to hit just those areas needing more detail with the 1/16" tapered ball nose end mill, which took only 1 hour and 41 minutes since I was cutting such a small total area. And because I was running this operation at 18,000 rpm as Brady suggested, I didn't get any fuzzing (well, in part that is an assumption, as this wood is funny, and on the previous applique, most of the wood didn't fuzz at 10,000 rpm, but some of it fuzzed terribly. It really varies significantly from board to board. Maybe I got lucky.:confused:) A few of those detail areas run with the 1/16" tapered ball nose end mill marks can be seen in the second photo, but I did not circle them as I should have.

One area that was hugely improved over my previous applique, is the smoothness of the vertical "2D Profile" pass, as shown in the area marked B in the second picture. In the previous applique, this surface all the way around was really jagged, and took LOTS of time to clean up by hand and was miserable work. I discovered what was wrong, and was able to fix this for this applique. The problem was that the vector I used to guide the profile pass, was created by Aspire's "Create Vector Boundary from selected components", aka"cookie cutter" (training video term). That vector follows every jagged pixel on the edge of the model, creating a terrible cutting result with my 1/4" end mill. I tried smoothing the jagged curve by editing and various methods of smoothing, but the resulting curve looked bad to me, so I actually MANUALLY traced around half the part (and mirrored to get the other half).

The FOURTH image shows a comparison of the jagged vector created by using Aspire's "Create Vector Boundary from selected components", aka"cookie cutter" in GREEN. The smooth vector I drew by hand is shown in red. So the smooth red vector is what I used for the profile cut line. This produced a really clean profile cut with a 1/4" end mill, @ 3IPS, .25" per pass, all conventional cuts.


Now, looking again at the second attached image, my big NEW problem this time, is indicated by those areas marked "A". This extra inset groove goes around the top edge of the applique's entire perimeter. I didn't have this problem on the previous applique. This new problem groove was created in my 3d finishing pass, using the 3/4" core box bit. (On the previous applique, I did 3d finishing with the 1/16" tapered ball nose end mill.)

If you look at the 3rd image (Over-wires Model detail for forum), it kind of looks like the new problem groove I have is shown in the model. But when I carefully studied the model before making the cut, (Brady told me to carefully study the preview before cutting), I just thought it was a shadow...duh!

I am guessing that this problem is somehow connected to the fact that I used a .25" "boundary vector offset" for the 3/4" core box bit 3d finishing pass. If anyone can shine light on why I got this really ugly extra groove as shown in the 4th figure, marked as "A", I would really appreciate it! I will make another post showing the tool path for the 3/4" core box bit.

Thanks again, Chuck

[Chuck Keysor]

PS: Because my new problem/extra groove appeared during the 3D finish pass, made using the 3/4" core box bit, I have attached a detail showing the toolpath for that bit. Possibly this will provide a useful clue as to the source of my problem. Thanks again, Chuck

[Brady Watson]

Hey Chuck,
A few pointers & observations - Looking at the zoomed in pics of your simulations, it shows that your model voxel/pixel density/resolution may be spread a little too thin. The tell-tale sign being the faceted/stair-stepped curves instead of smooth and fair curves in the preview. If you are going for quality, your model space should match the size of the relief that you want to cut with just a little extra around the border. So for a 20x48" relief, I would make the model space something like 22x50" to minimize spreading those voxels too thin. You should be running at least 7x resolution, although if you hold shift when you create a new model, you can choose 25x & 50x to increase density. 25x may show an improvement, although it does take longer to process on your computer.

If I were machining this model, I would carefully analyze the geometry and then choose a combination of 2D and 3D strategies to get the best machining results. Looking at your previous work on this relief, it appears that some of the center portion would benefit from some 2D profile toolpaths to get those crisp edges in the design. I would also be inclined to use a tool in the 3/16-1/4" range (ball end mill) to resolve the actual relief detail itself. This model really doesn't have a lot of detail to it, and certainly doesn't command using a 1/16" ball to machine the relief.

At the end of the day it boils down to spending the time creating optimal toolpaths - or spending longer machining time with a small diameter bit. If you plan on cutting this same relief multiple times, then it pays to put the time into well thought out toolpaths. If it is a one-off, then you just blow it out as efficiently as you care to.

I would avoid the large core box bits for 3D relief machining, unless the relief has very low detail, with a large sweeping shape, or for texturing. Core box bits are not very accurate down at the center point, unlike an end mill. They take a lot longer to calculate (3D TPs) and in some cases can cause gouging in unwanted places that don't show up in the preview. This, combined with lower model resolution (how closely does a stair-stepped model come to "ideal" ?) - can compound into unwanted tool marks.

If you export your model as a .3DClip and email it to me, I will check it out and make suggestions...

-B

[Chuck Keysor]

Hello Brady. Thank you for your continued assistance! Let me reply point by point.

This applique is smaller than the previous one, being 36" x 21". My "job size" is 40" x 26". I aimed to have a 2" margin all around the edge to accommodate my mechanical fastening methods, though perhaps I am being overly cautious. So I could work to improve this figure as you pointed out.

I did run the model at 7X. (Can I switch to 25X after my model has been created, or must that be selected before I begin to create my model?)

Yes, the earlier/larger relief, after making it, and getting feedback from the forum, I had wished I had used some 2D cutting strategies, as there are some parts with geometric forms to them. On this newer/smaller applique, it seemed much more organic, and even the middle has an arching cross-section. So I didn't see quite where to apply 2D tool paths with the exception of course being the 2D profile cut.

As an important digressional note, before I get back to answering your questions....... much of this model was left simplified. That is because as I was trying to model what I wanted, I felt uncertain of the details on the leafs and the modeling of the torch. The thing that looks like an egg, was I assume, actually a torch, but the lack of detail in the original photo made it look like an egg! When I tried modeling a torch, I was most unsuccessful, and so I decided to let the Bot produce the egg, and the blobby leafs, and that I would carve these by hand to help me understand what I was trying to achieve. It may seem non-intuitive, but I can carve what I want better than I can draw what I want, and the computer modeling requires me to draw what I want. So by carving what I want for the torch and the leaf profiles, I will get a much better link in my brain, and can take pictures, that will then help me to better model the acanthus leafs that are in 4 upcoming house plaques. I will post pictures of the before and after of my hand carved details. (Double digression, I didn't do any post carving of the first bigger applique, because it had so many visual elements going on, that the lack of definition in those acanthus leafs wasn't as troublesome. And I simply wanted to get through that and complete it, since it was my first such project.) And from the picture, the larger branches did seem to have a very simple arched cross-section. So I am taking my experimental, learning liberties with the smaller details, and leaving the largest branches alone.

It may well be that the type of detail I want to create in the end, with my chiseling, will affect the type of tooling that you would recommend. Hand carving is satisfying, but time consuming. So in the future, I would like to have the Bot bring the wood as close as is reasonable to my desired end goal, without the need for significant hand carving.

These appliques are all one of a kind, and each requires a new model. Though because the first applique and this applique had almost identical main acanthus branches, I was able to copy that part of the first model and bring it into my second model. (But as I noted earlier, without any "history" attached to the 3d swept components, I effectively had to start all over again anyway in my model creation.)

The large core box bit did work nicely for most of this applique, because it is comprised of large sweeping arcs, with rounded details. But that should change in the future based upon my above stated intentions.

Your point on the accuracy of the core box bit is of course of significant concern to me. But before I had posted my questions on the FIRST applique, I had already purchased a new Amana 3/4" core box bit which since I didn't use it on the first applique, I was itching to use it on this second applique. I still have to purchase a couple of sizes of larger ball nose end mill.

The 1/16" tapered ball nose end mill was used not just because I had it, but all of the webbing between the fingers of the acanthus leafs was left out by the large box core bit, and I wanted to at least have these areas cut out to reduce the variations between my different chiseled details. Though that now seems almost illogical since I am carving a plant, and not a machine.

I looked in the Aspire menus, and didn't find any way to export a.3DClip. I could only do an STL. I am guessing I would need to export the STL file to Rhino, and then create the .3DClip file there. I haven't gone to look at Rhino to see if it will do this. However, maybe I should post pictures of the completed chiseled applique first, since that is what I want to model in the future. Please let me know if that seems appropriate, or if I should use Rhino to create a .3DClip file for your examination.

Thank you again, very sincerely!!!! Chuck

[Brady Watson]

It may well be that the type of detail I want to create in the end, with my chiseling, will affect the type of tooling that you would recommend. Hand carving is satisfying, but time consuming. So in the future, I would like to have the Bot bring the wood as close as is reasonable to my desired end goal, without the need for significant hand carving.

Hey Chuck,
This is where hand carving one model & having it laser scanned makes a lot of sense. Even though each 'custom' sized part might be a different size, you can usually reuse a good bit of the original & or make derivative works of the original. I do this practically every day...from relief carvings to guitar parts and everything in between!

You can easily export a .3DClip in v3.5 or newer by right clicking on the 3D component in the list on the left & select 'Export as 3D Clipart'. It will default to .3DClip. Feel free to email it to me.

-B

[Chuck Keysor]

Thanks Brady! I have created the 3d Clipart, but I don't know how to send it to you. I clicked on your avatar, but didn't find any way to attach a file to a private message. I went to your company's website, and found an "email us" button, but that brought me off to some strange sign-up form that I backed out of. In any event, can you direct me to a means by which I can transmit this 3d Clipart file to you? Thanks, Chuck

[Brady Watson]

PM sent...

-B

[Chuck Keysor]

Hello Brady. Here is what the entire Victorian "applique" looks like after I finished the hand carving. For comparison, I also included shots of before the hand carving.

I just got done coating this with penetrating epoxy for water proofing before I took the finished pictures. That has of course changed the color. (I will later prime this. Ultimately, it will be painted to harmonize with the house.)

I could have kept on carving, but I realized that I had gone far enough. I got a good handle on what I want to try and produce on coming acanthus leafs. Clearly I will need to change the cross sections that I use in my single rail sweeps. Does this help you in offering suggestions for my tool-pathing strategies?

This finished carving is much more to my liking. But clearly I need to get the Bot to do the carving!

Thanks, Chuck

PS: The little squiggle at the very tip of the branch is intentional. I saw that on an acanthus leaf carving I found in my research and thought it looked good.

[myxpykalix]

Chuck,
That's beautiful detail work and i'm wondering if all the nice detail work will be appreciated if this is something that is going high up on a house? I hope you are getting paid for all that work. Excellent job!:D

[POPS 64]

That is fine work ,something to be proud of not everyone has the patience for that , good job Jeff

[dan_nelson]

Chuck, I'm just asking. If your doing all that hand work after your piece gets off the bot after 8 hrs plus, why not just cut 2d and save about 7 hours of cut time? Unless I'm missing something that the first picture shows. Dan

[Chuck Keysor]

Thanks Jack and Jeff for your positive feedback. And thanks Dan for your inquiry, which is very logical. I will ramble for a bit, to answer your question, and Jack's.

a) Yes, this applique is mounted up really high, and the detail will probably not be noticed. But this project was undertaken as a learning experience where I am getting paid almost nothing. I have a great interest in Victorian houses, and when I was planning to buy my Bot, it was before the economy crashed.

One thing I wanted to make were carved parts like for Victorian houses. In Elgin, there had been a thriving business in Victorian house restoration, and the only way the local restorers had found to replace things such as the applique in this thread, was to use something salvaged from another old house. I thought I could make it a key part of my business (I don't have a business, still just a concept....) to make such house parts, one of a kind reproductions. But after the economy tanked, the old house restoration business in Elgin dried up and blew away. It is still dead all these years later even though we are told that the economy is picking up a little.

Any way, a retired school teacher friend of mine who fixes old houses out of personal passion offered me $500 to make all 4 appliques for this house. (He is providing the wood.) Because I wanted to learn how to do this (3D work) and it involved old houses, I accepted.

Now, getting to Jack's and Dan's points.... Based upon the original photo (which I posted a month or two ago), the original appliques were not very detailed, as best as can be told. That makes some sense, since they are up so high in the air. The first applique I made (also posted, last month), I slavishly copied the picture, and did not include any details that I felt weren't present in the original. The first applique, the biggest one, turned out like the computer model, and took far less total time than this second applique that is the subject of this thread. However, when I took the second applique off the machine, I at first went, "Cool, this looks like my model, which looks like the original applique!" But then I looked at it, and said, "Gee, this actually looks pretty clunky and uninspired!" I already had the applique off the Bot, and decided that I would try and make it look better by hand. I figured I had to spend time to clean up the applique by hand anyway (maybe 6 hours worth, based upon the first applique clean up time). So, I began to sketch on the part, and to chisel!

So Dan, it was not my intention to go and do all this work, and Jack, I am not expecting that the detail will be appreciated three stories up in the air! And no, I was not being paid for the extra work.

So, why did I do this? Because I wanted to see if I could do it, AND to get a good, intuitive sense (all those hours carving will help there ;)) of what I like in an acanthus leaf. With the improved sense of what I think an acanthus leaf should look like, I will be able to more confidently make good computer models, that WON'T require all this terrible amount of handwork. (Even I reached my limit, as I could have kept on and added detail to the entire piece, and maybe spent another 20 hours. But my interest was in the acanthus leafs.)

So, what are the chances of ever making another acanthus leaf???? Well, funny you should ask..... Years before I bought my Bot, and before the economy crashed, I promised friends of mine that if they provided the materials, I would make for them at no cost, THREE big triangular panels that go in the pedimented gables over three sets of stairs on their Victorian house's big wrap around porch. Those panels are typically 10 or 12 feet up in the air, and are very visible to anyone entering the house. When their house was "remodeled" in the 1960's these three heavily carved relief panels were removed when they wrapped the house up in white aluminum siding. From what little can be seen in a few crummy old photos, these panels were full of acanthus leafs/vines/scrolls. Almost all of the rest of the house is now restored, except for the three carved pedimented gable panels that are still missing. So that is what I will make next.

If someone wants to stay busy, making free, or low cost acanthus leaf decorations, there are a bunch of people in Elgin who will take your work!

So I hope that explains a bit of what is going on with these appliques. Now if it were so easy to explain why I do these things...... :confused:

Thanks for your positive comments and questions, Chuck

[dan_nelson]

Chuck sent you a email via forum Dan Nelson

[Chuck Keysor]

Hello Dan. I checked my private Message folder on the SB forum, and didn't find any message from you. Have I looked in the wrong place? Please let me know, thanks, Chuck

[myxpykalix]

Well Chuck that certainly explains your motivation!:D It certainly explains why i have spent dozens of hours making a bar complete with carvings on the front of the bar face and carvings on the cabinet doors....for my son...for free.

I would however try to do something to be able to scan your finished applique into a 3d model so that you could keep the finished product as a file so the next time you make one you don't have to hand finish it.

Also why limit yourself to only doing this work locally? Make yourself a simple website specializing in these details, let people take a picture and you reproduce the part and ship it to them.
People who want this type of stuff have the money to pay for it. I'm currently working on some fireplace details for a restorer along with some period balusters and she's gladly paying for my service because you can't go down to lowes and buy this stuff.

[dan_nelson]

Chuck it got sent to your email ( the third option on the private message center )
It might have went into spam.
Anyway shoot me a email at Nelsonwoodcrafters@yahoo.com

[Chuck Keysor]

Thanks Jack. As to making a website, etc, in my perhaps overly analytical way, I need to know IF I can do this work in a satisfactory manner. And I must know how long it takes me to do this work. And I have to find out if I like doing this work.

In terms of relying upon the local economy as a base, there is the real concern that making a website will not attract customers in and of itself. Working/networking with the local craftsmen/women who do old house work in my area seems more logical for ME to establish my skills, portfolio and reputation. I would feel after getting local traction, and finding it could not support me, then it would be appropriate to try and expand. The Internet would then play into that effort.

Dan, I have sent you an email.

OK, now I have to revise my Aspire model for the last two appliques. I will do my best to create a more chiseled look than the models that I have already created.

Thanks, Chuck

[Chuck Keysor]

I had one pair of Victorian house appliques left to make in this set. These are a mirrored pair to flank an attic window. The attached image shows the left side applique, as I originally modeled it to match the photo (as best I could see it), and then the revised model which shows a more chiseled approach.

I tried to keep some of the rounded character of the original design, so that it would relate to the other two appliques in the set. However, as may be expected, the revised design does have some improvement concerning the chiseled aspect, but now the design lacks a little balance. The chiseled top and left fans now look to be too light relative to the leg and central spiral. I suppose I could go back and spend another hour or so to fiddle with this. But at this point, I will declare it "good enough"....... I think.............

Chuck

[myxpykalix]

does the phrase..."leave well enough alone"...mean anything to you:eek::D

I have the same argument with myself on some things and it looks fine as is and your client will be happy....good job:)

[Chuck Keysor]

Hello Shopbot Friends:

In my most recent post, I showed the revised model I had made for the last 2 carvings in this project. In this post, I am showing a problem I had in actually carving that model, then the manually fixed up actual carvings of that same model, and finally the full set of the carvings for this one house that I cut on my Bot.

a) I call these last two carvings, which are mirror images of each other, "wings". I carved the "left wing" on my Bot first, and was very happy with how it turned out right off the machine. It looked very much like my model, and had its first finish path with a 1/2" core box bit, then for the second finish path, I hit the detailed areas with a 1/16" tapered ball nose end mill, then cut the part out with a 1/4" end mill. A detail of the good "left wing" is shown in the left side of the first picture.

Then I created the "right wing" model by mirroring the "left wing" model, and recomputed all the tool paths. When I machined that part on the Bot, the results were not anywhere nearly as good as I got for the "left wing", even though they should have turned out with the same quality. The 1/16" tapered ball nose end mill cleaned up some areas, but not others.

A detail of the bad "right wing" is shown in the right half of the first picture. These poor results, of what should have been a mirror image of the first part added several extra hours to the manual clean up of this part.

When I did the profile/cut-out operation on the "right wing", the registration was perfect, so the poor results of the 1/16" tbnem finish pass are not due to losing my zero reference point.

Any obvious suggestions for what may have gone wrong with the "right wing"? (No, this is not a political discussion :rolleyes:) Thanks for any suggestions.

b) The second attachment shows the completed left and right wings. (The left wing had been sealed, and the right wing had not been sealed, accounting for the different colors.)

c) The last attachment shows all of the "appliques" for this one house. The parts are primed, and holes have been pre-drilled. My recorded work times for each part are shown for general reference. This is not fast work, (the one that took 25 hours of hand work is because I went and manually recarved much of the applique. I did that when I realized that I didn't like what I had designed in the computer, even though it matched the picture I was trying to copy!).

Thanks, Chuck

[steve_g]

Chuck…
Two possible causes
Not being “there” limits my ability to see the issues but, any chance the “bad” wing wasn’t held as tightly as the good one? It almost seems like you have some vibration or scalloping from movement. Another possibility, I wonder if the grain of this piece of wood was running different than the first… Obviously not a LR vs: top/bottom but possibly rising or falling along the grain, giving you a greater “end grain” component.
My Question…
When these houses were built, were these ornaments a catalog item, or were they carved on site? Just curious…
SG

[Chuck Keysor]

Hello Steve. Thanks for your reply.

I went back to look at my pictures to answer your questions, and in so doing, I saw a new, important clue, (which is why I realize it is so hard for anyone to answer questions of this nature, as you implied, you just about have to be there to see what is going on.....:()

The first attached picture shows the new clue,,,, chatter in the first 3D finish pass, which was made with the 1/2" core box bit. The chatter seems to have resulted in an "undercut", so that would explain why the next pass with the 1/16" tapered ball nose end mill didn't clean up the previous cut correctly.

Now, back to your questions. The hold-down method is shown in the second picture, and it is exactly what I used for cutting the previous carving that turned out great. But the point of note is that this chatter of the 1/2" core box bit, was as bad at the bottom side, where the means of securement was beyond reproach. The long thin board holding the middle of the top, as ratty as it looks, is a piece of white oak, with a reverse bow on it, so it was really cranked down. Though again, this is clearly the weakest part of my hold down. (You can see I don't like to screw into my table top.) Had the chatter been worse at the top, I would have been quite likely to ascribe this problem to the means of holding the blank in place.

But this chatter of the 1/2" core box bit did not take place at all on the "left wing", were the result of that cut was really very good on all counts! And the speeds and feeds were the same for both the right and left wings, 2.5IPS, 8% step-over, 10,000 rpm.

The grain orientation for both carvings was exactly the same.

As to your question about these appliques being carved on site..... I don't officially know. I have read some histories of Victorian architecture, where they blithely state that with the invention of scroll saws, and the like, that the carpenters were able to whip up all these great details on site, and then as if to prove this, they show a picture of a treadle operated scroll saw. But if you have cut with a scroll saw, it simply seems absurd to me to think that a carpenter would sit by the house, sawing out such intricate parts. It is hard enough to do today. I have looked at maybe 5 or no more than 10 reprinted Victorian millwork catalogs, and they do show lots of factory made fancy cut shingles, grilles, brackets, newel posts, and even gable ornaments (which are not nailed to the gable wall, but are attached only to the bottom edge of the roof, and are thus suspended) but I haven't seen this type of thing I have made appearing as a catalog item. But again, these seem so labor intensive to carve, I can't imagine them being carved on site. It is my suspicion that I just haven't found the right old catalog.

Thanks again, Chuck