First Try From Image

[5artist5]

So this is my first attempt to mill something from an image.
It actually came out better than I thought it would. But I would
like to be able to do it much better.
The pattern is an inch by an inch and the cutter was a 1/16 inch ball.
I wish I could remember what the step over was but I don't remember.

So to make it better I am thinking that if I decrease the step over distance
I should get the circles to appear more round and probably improve the
shape of the text?

Then I got to thinking that if I went to say a 1/32 ball I should be able to get the text to cut deeper and improve the look, but that means the step over will have to be like 4 times as short right?

[DeusExMachina]

Looks pretty good really,

Try a 4-8% step over (translates to about 0.0025"-0.005" ridges) and try a square end end mill. You actually don't need a ball end unless you have features that require it.

This to me looks like a height map generated by an image with a lot of contrast where the edges between the circles and the background a very crisp. This type of image doesn't really require a ball-end mill and will cut down on ridges on the back of the plate.

Now if you decide to do something complicated like a portrait then the ball end mill would be necessary.

What software are you using to generate the gcode?

[5artist5]

Interesting idea about the flat mill. I can see how that would be a good idea for this part.

I actually want to do much more complex ones though. I just did a real simple shallow one cause I didn't have much time to wait around while it cut last night.

I'll try it again tonight with something a little more complex and see how it goes with 4-8%.
Thanks!

[Mini Beast]

That is really cool. I'm also interested in how you did that and with what software? Does the software do most of the work, or is there a bit of a learning curve to get 2D images to 3D shapes?

I'm planning to cut a gargoyle shape from vectorart3D when I get my machine built, but I would like to learn how to do what you're doing and be able to create these shapes myself.

[5artist5]

It was pretty easy actually, I just made a gray scale jpeg and opened it in Meshcam.

There are a lot of settings to play around with, and I am sure that I can get it to come out way better by messing around with the settings some more. Of course I expect it to take a lot longer to cut it out in order to have it come out nicer...

[5artist5]

I did it with photoshop and meshcam.
It was pretty easy. Especially compared to writing all the gcode for it out by hand!
It came out to about 48000 lines of code!

I'm going to try another one tonight, i'll post the results.

[Mini Beast]

Very cool! I'll look forward to seeing your next part.

When you are playing with step over distance and cutter sizes, do you have to scrap the part each time you try something different? Or, can you see that you want a little sharper detail and reprogram and run on the same part with smaller step over to smooth it out?

[5artist5]

you could do a test area on the part, somewhere that you were going to cut away anyway. You could put in a circle and see how it comes out before doing the whole pattern. Once you get a feel for what the settings should be you probably don't need
to though. I was thinking that if the one I did last night didn't come out that I would just mill it flat and start again. But I will probably keep it so I can have a reminder of my progress. This milling stuff has no end to things to learn!

[escott76]

You've obviously got the software you've got, so you can take this with a grain of salt. There are a couple of things that will influence the results you get. The square end (or a smaller ball mill to some degree) will give you steeper edges per a given cut depth. That will improve things a bit.
The other thing that would help is a machining strategy that does a better job at following arcs and such instead of just generating a straight raster cut. Here is where you are limited by meshcam I'm guessing. Given that a smaller end mill and smaller stepovers will certainly help. I do some mold cutting, and I use a constant 3 dimensional stepover cut of .001 or .002 and get a great finish, it does take a little while though.
I'm not sure exactly how meshcam maps pixels of an image to features size in the final part, but there may also be some stair stepping introduced as part of the attempt to map one to the other. Might be worth a look, although it's just a theory, as I said I don't know the software.

[Mini Beast]

This makes sense. You could cut the top and the floor with a square end mill. Then, go around the contours with the ball nose instead of sweeping straight lines back and forth. Is that even possible? I'm just now starting to realize how much I have to learn, haha.

[MechanoMan]

The smaller mill does not make smoother features.
The smaller mill allows you to make smaller features than a mill which is wider than the feature.

With fixed parallel stepovers (raster output), the smaller mill leaves a bigger cusp as it climbs vertically. This can be reduced by using smaller steps and extending the milling time. However, even if you reduced the stepover by half, the cusps on nonhorizontal surfaces will always be present just reduced. And they will always be smaller if you use a larger mill.

In short choose the largest mill that can still render the features. In some cases it makes sense to do a tool change to render just the fine features with a smaller mill.

[5artist5]

I see what you are saying about using the biggest mill that will work.
I am timid about doing a change on a program like this though. How would I ensure that the Z is the same between the tools? I would like to try it though, maybe I will give it a try tonight.

I tried a much more complex image last night.
The details were too small for my 1/16 mill so most of the detail didn't come out.
I guess there was also a problem with the way I told meshcam about the size and or location of the material because even though the Z0 level on the mill was actually
set below the top of the material it still didn't cut the highest details.

I wanted the Z 0 that I set to be the actual top of the piece after it was done.

Anyway, here is a picture of the result and also the image file that I put into meshcam.

I think I might try doing the models in 3D instead of trying to do these gray scale images. I don't feel like I have as much control as I would like with the images.


Mini Beast- Yeah there really is a lot to learn. There are a lot of places that problems can come from that lead to unexpected results. It is sort of overwhelming because you really need to learn how to machine things, how mach3 will handle the gcode, how to generate the gcode and how to make the drawing or model to give the cam software. I think that most people who get into this sport already know how to do one or two of these operations and that is a help. But getting all of these operations to merge smoothly into the joining ones is a juggling act for sure! Still it is AMAZING to watch the cutter shape the material. I literally sat and watched it last night for an hour and a half as it revealed the new shapes each pass. Memorized like a kid at a magic show!