I am kind of frustrated today...

[Burkhardt]

Well yes, my wife is 6000 miles away and I am on my own for Christmas. But what really ticked me off today comes from the following:

So I glued up a huge kind of butcher block from redwood and pine 2x4 studs:

https://lh5.googleusercontent.com/-cSsWce-xZf8/UroTYLsPvMI/AAAAAAAAEao/99Ej1niJ_kk/w1024-h768-no/DSC02069.JPG

Surfaced and trimmed it on the CNC:
https://lh4.googleusercontent.com/-iQ6FuZ78NRE/UroTb7GRSrI/AAAAAAAAEaw/XjZHEAf4sCw/w1024-h768-no/DSC02070.JPG

Made a ton of shavings roughing out the shape (reducing from 30 pounds to about 8 pounds:
https://lh5.googleusercontent.com/-O8e1D3xLjls/UroTfIov6lI/AAAAAAAAEbA/FcmnSlW_tsk/w1024-h768-no/DSC02072.JPG

So far so good. But at 20% of the finishing run, the bit (1/2" long ballnose) worked itself loose in the collet, dug in, started wobbling, shut down the spindle controller and stalled the machine (not easy to see in the picture). I could have sworn I tightened it properly. Probably a loss but I will try another finish run 1/4" lower and hopefully sand out the remaining scars. Or start over again.....
https://lh4.googleusercontent.com/-H_bcriToXbA/UroTgWz9tgI/AAAAAAAAEbI/siRfatYx-a8/w910-h862-no/DSC02073.JPG

[Kyle Stapleton]

Hope it is all dry or in a year in could be very bad.

[gene]

Why dont you use a flat bottom surfacing bit if you are smoothing it out?

[Burkhardt]

Why dont you use a flat bottom surfacing bit if you are smoothing it out?

I did use a surfacing bit to plane the block from the top. But that was not the problem.

Maybe I did not explain it right...I used a ballnose for the 3-d carving and that got loose while finish cutting. In the meantime I ran the finish cut again and I guess I can salvage the piece with some sanding. You can see the unitentional notch on the left side (after the 4th strip). I may anyway sand blast the surface to better hide the imperfections.

https://lh4.googleusercontent.com/-pLAQYNCOz9g/UrpSXnpQwpI/AAAAAAAAEbg/Lhnm7jOZTGg/w1024-h768-no/DSC02077.JPG

[myxpykalix]

G,
Well yes, my wife is 6000 miles away and I am on my own for Christmas

Where is she...off helping Santa at the north pole?:eek:

That looks really neat, and what you could do is is map out those couple areas and do a 3d carving in that area that would emulate a ripple within a ripple like when you throw 2 rocks in a pond and the waves cross each other?

I think you would have to do a seperate 3d model as opposed to your ripple program. Either way it looks great.:D

btw learn to not be frustrated when stuff like this happens, it happens to many of us...it's how you figure out how to fix those mistakes that makes you a good botter;)

[Burkhardt]

G,
Where is she...off helping Santa at the north pole?:eek:

That would not be so bad, North Pole is only 3900 miles from here.
She is actually in Germany. Her mom is old and leaving her alone for Christmas is kind of unconscionable. Normally I would go, too but I can't this year for business reasons. But I can handle it. Couple of totally free play days in the shop :)

Normally I am quite frustration-proof and just do it over again if something goes wrong. But in this case I spent some 2 days cutting, glueing, roughing and on the PC and thought for a while I was @%^&$%. I would not have started over, given that I just shredded 20 board foot of lumber. Cheap lumber, but anyway.

[scottp55]

Gert, What stepover finish pass? Closer pic custom "rumblepad?". I realize you are totally frustrated with it and don't ever want to see it again---so please advise me on shipping costs-got a perfect spot for it in my house:)

[mikeacg]

It turned out beautiful regardless! I would just get out my gouges/chisels and handcarve it to blend it in with the rest of the design. Due to the nature of the design, I don't think anyone would notice it if you just smoothed out the dings a little...

Mike :)

[rb99]

Maybe you could find some large glass jewels and inset them into the spots.

This would give the piece some focal points for the viewer.

[cowboy1296]

Interesting to say the least.

Recently I was having issues where I thought that I had not tightened a collet and my bit slipped. Then I realized that it was my porter cable collet that was defective. I upgraded on the collet and have had no problems since.

When in doubt: sand and then sand some more.

Now other than some kind of wall art what is the intended use for your project?

[scottp55]

G., Crushed Malachite or turquoise or MOP to "Celebrate" the mistake and make it look like you meant to do that? Don't totally infill but follow the contours of your mistake (world within world?). Crushed mica mixed with shellac and "Paint" divots? :) Almost looks like fantasy topography.

[bob_reda]

In some circles, they would call that 'character'

Bob

[waynelocke]

Best woodworking advice I ever got. If you can't fix it feature it. Intent is everything.

[Burkhardt]

Thanks to All for your motivational and practical advice!

I kept sanding and in the end the gash may indeed become a feature, although it is not that prominent anymore. I have a suspicion, that the collet got loose because I failed to clean and lubricate it for a long time. The dust buildup inside may have prevented it from clamping tight. At least there was a lot of gunk when I blew it out.

Couple of things that I learned (I have never done such a massive 3-d carving before):

- roughing is overrated. It takes a lot of time without improving the surface finish. In the future I will rough only deep enough that the flute length can get the rest while finishing. If it is shallow enough then not rough at all.

- 80/1000" horizontal step finish cut leaves a lot for sanding with a 1/2" ballnose. It took about 3 hours at 2.5ips but I should have invested more time in a finer step for finishing. Given my trouble with the gash and remediation that is a moot point, though.

- Redwood (I picked mostly heart pieces of Con Com studs) is fabulous for such large carvings that don't need intricate detail. It machines cleanly, has fine grain and sometimes wild pattern. It is soft and sands very smooth but will splinter at fine details. At 80 cents a bdft it is rather inexpensive. It should also be great for v-carved designs.

- The pine studs are not so good. Much coarser and fibrous grain that tears out.

[Burkhardt]

Now other than some kind of wall art what is the intended use for your project?

That is exactly the idea: Wall decoration for our new bathroom. Couple of months ago I wrote the wave program that generates the g-code and still don't know exactly what to do with it. But I could put it to good use here. I probably played several hours with the parameters until I had a pattern that I liked.

[coryatjohn]

>> I have a suspicion, that the collet got loose because I failed to clean and lubricate it for a long time.

Could it be the bit was starting to get hot? I've had that happen to me (1/2 end mill). Did you change any cutting parameters right before the problem?

[Burkhardt]

Could it be the bit was starting to get hot? I've had that happen to me (1/2 end mill). Did you change any cutting parameters right before the problem?

Could be, not sure. But I did not change anything in the program. One thing I did notice is that the spindle sounded much noisier and I almost thought the bearings are on the way out. It might be as simple as a tiny chip getting in between the bit and the collet leading to some imbalance/wobble and vibration which in turn led to loosening over time. I will never know. But after cleaning and re-tightening the collet nut the spindle is it's own old self.

[Bob Eustace]

I have a very expensive piece of timber hanging on the wall - a large 3D clock. It has a huge gouge right down the middle! This was done by two STUPID things. Leaving the Bot running during lunch out of the workshop and not replacing high use collets at least every 6 months. Brady put me onto this and I also took his advice and never buy cheap collets. Anyway its on the wall to remind me if I forget to take the Alzheimers pills!

[Burkhardt]

Phew...with a lot of sanding, sandblasting and abrasive bristle brushing I think I got it into a decent shape so that the mistake is not obvious. Actually, I had planned to sand it smooth but I like the more textured look of the blasted/brushed surface. Makes the grain stand out on the softwood. I will see tomorrow by daylight if it is good enough to hang it. For now I will take a shower to get the blasting sand out of my hair.

https://lh6.googleusercontent.com/-Xwo8vd2a6Po/Ur5484jbNvI/AAAAAAAAEb4/pvSST20ZioM/w1024-h768-no/DSC02083.JPG

https://lh6.googleusercontent.com/-60MI_Fgt4YM/Ur549xwfaOI/AAAAAAAAEcA/ZZQv-Qon5ys/w1024-h768-no/DSC02085.JPG

[mikeacg]

I think the added texture makes it look even better! Good job!

Can you tell me more about this ripple application you wrote?

Mike

[scottp55]

Sand dunes spring to mind, Beautiful save G., Better than it was! Please advise me on shipping costs :)

[gundog]

Here is a tid bit you may already know but never seat your bit all the way in a collet ever. As the collet tightens it draws the bit in and if the bit is bottomed out it will never be as tight due to the bit stopping that small upward movement.

Mike

[coryatjohn]

I'm kind of curious as to how you developed the "G" code for that work of art. I have topographical data I'd like to convert to 3D maps by creating "G" code from a program.

Anything you can add would be of interest. Thanks.

[Burkhardt]

I'm kind of curious as to how you developed the "G" code for that work of art. I have topographical data I'd like to convert to 3D maps by creating "G" code from a program.

Anything you can add would be of interest. Thanks.

Oh my. Now I will get the "nerd of the week" award. But here it goes...

Like most of my experimental programs this one has no user interface yet. I just change the parameters in the code and compile it again. I can define up to five wave generator points inside or outside the machined area. Each point can be parametrized with position, amplitude, wavelength, decay and delay. The one I described in this post has 2 generators. More than 3 is kind of useless and will just show a jumbled mess. It took me a while to find a simple enough algorithm that resembles water ripple. The physics of water waves are quite complicated as the propagation speed is proportional to the amplitude.

- The software first creates a z-height matrix with a point-to-point distance of the horizontal x and y steps for the bit.

- Then it scans all points and calculated the z-contribution from each wave generator. I can also let the amplitude go to zero at the perimeter of the working area to have a smooth edge.

- Then it creates a matching matrix of all the normal vectors for each vertex of the z-matrix. That is needed later to find the ball end position relative to the target surface.

- It displays a color coded 2-d height image for quick control of the result.

- It generates an .stl file of the wave surface that I can use to get a realistic picture in an .stl viewer. It also generates an .stl file of the tool position trace, allows to see if some curves are too tight for the ball end radius.

- Then it scans the z-height matrix y-first and then x to generate roughing g-code for an endmill. After that finish g-code for the ball end bit.

-Optionally it can read another g-code file (e.g. v-carved text) and warps it to conform with the wave surface.

That's about it. The output can be done in one piece or tiled for larger objects.

If you have your topo data in text format as mm or inch values in rows and columns it would be very simple to do a software making g-code from it. Please PM if you need more detail.

[myxpykalix]

It took me a while to find a simple enough algorithm that resembles water ripple. The physics of water waves are quite complicated as the propagation speed is proportional to the amplitude.

HUH?:confused: I think i understand.Here is a thought and whether it is relevant or usable i don't know....Is there a way to translate/use/convert oscilliscope generated wave patterns into something you could use?

I i have seen some wave pattern generators out there before, just not sure where since it's been a long time.

[Burkhardt]

A simple sine wave (like you see on the scope) is a starting point. It may give you a nice pattern but is not realistic for a water ripple:

https://lh6.googleusercontent.com/-nZ2TefebE88/Ur9p5nxczlI/AAAAAAAAEc0/B2aFCzvI1sE/w1289-h696-no/ScreenHunter_03+Dec.+28+16.01.jpg

If you add the decay and wavelength reduction it looks more like the real thing but not exactly:
https://lh3.googleusercontent.com/-JLf8Y7bDjog/Ur9p6xshzjI/AAAAAAAAEc8/0Ep6d1KKiPE/w1362-h862-no/ScreenHunter_04+Dec.+28+16.02.jpg

Then you can add a time delay for the wave to propagate from the center and that is as far as I got. I think it is close enough toe a stone drop in a pond that I can use it.
https://lh5.googleusercontent.com/-MXKmzhB-5ZM/Ur9qAuTBoqI/AAAAAAAAEdE/6uB70ZbJz1s/w1353-h827-no/ScreenHunter_05+Dec.+28+16.11.jpg

[David Iannone]

That's a great job. The texture is cool

[scottp55]

G., Like Jack said "Huh?". Pictures explain so I could better understand, thanks. Training a computer "Savvy" person this weekend and understand "Nerd" has negative connotations while "Geek" is now positive(sometimes). So you did qualify for "Geek Speak of the Week" :)

[coryatjohn]

Oh my. Now I will get the "nerd of the week" award. But here it goes...

Like most of my experimental programs this one has no user interface yet. I just change the parameters in the code and compile it again. I can define up to five wave generator points inside or outside the machined area. Each point can be parametrized with position, amplitude, wavelength, decay and delay. The one I described in this post has 2 generators. More than 3 is kind of useless and will just show a jumbled mess. It took me a while to find a simple enough algorithm that resembles water ripple. The physics of water waves are quite complicated as the propagation speed is proportional to the amplitude.

- The software first creates a z-height matrix with a point-to-point distance of the horizontal x and y steps for the bit.

- Then it scans all points and calculated the z-contribution from each wave generator. I can also let the amplitude go to zero at the perimeter of the working area to have a smooth edge.

- Then it creates a matching matrix of all the normal vectors for each vertex of the z-matrix. That is needed later to find the ball end position relative to the target surface.

- It displays a color coded 2-d height image for quick control of the result.

- It generates an .stl file of the wave surface that I can use to get a realistic picture in an .stl viewer. It also generates an .stl file of the tool position trace, allows to see if some curves are too tight for the ball end radius.

- Then it scans the z-height matrix y-first and then x to generate roughing g-code for an endmill. After that finish g-code for the ball end bit.

-Optionally it can read another g-code file (e.g. v-carved text) and warps it to conform with the wave surface.

That's about it. The output can be done in one piece or tiled for larger objects.

If you have your topo data in text format as mm or inch values in rows and columns it would be very simple to do a software making g-code from it. Please PM if you need more detail.

Thanks! Great detail. I can go directly to the stl with the data I have. The rest will be easy.