2 years dealing with 3d routing problem!!

[danielcoyle]

Hello. I have been teaching myself CAD, CNC and CAM over the past several years and usually able to solve most problems. I am using the original version of Rhinocam (version 1)
I can machine things that have been saved as Rhino CAD 3 or older.

The problem is when I am machining several versions and scalings of helmet models. I machine the inside of the helmet with a rough and then a finish pass out of a stock block and then,of course, create a 0,0,0 reference on the block. I flip the block over using the reference in my setup of course. In machining the top of the helmet off is where the problem happens. The helmet lays front to back axis along the X axis of my router. In this axis almost all (but with an exception here and there, and I can't figure out what causes the exception) of my helmets end up being machined almost as if the reference point has been moved forward 5 or 6 millimeters (and this is a problem because that is almost the entire final thickness of the helmets. But that is not exactly correct. It is almost as if the top surface of the helmet comes out shaped just a little different from the bottom. I have tried using a single surface and stock offsets to machine BOTH the top and bottom surfaces AND I have tried machining from a model that has a top and bottom surface with apparently uniform thickness throughout. No idea what is wrong here, at all, and have been having to monitor the machine and re machine "by eye" for 2 years now. Any ideas?

you can see in the image the final result of the problem where the front of the helmet is shaved off and the back is significantly thicker than the intended 6mm thickness.


Similar Threads:

• Need Help!- Major problem reinstalling Mach3 after running for years it crashed
• YOOCNC 3020Z-D52 routing STL problem
• Newbie- Routing problem
• Need Help!- Acrylic routing problem
• Speed problem when routing a lot of small lines.

[awerby]

Where is this reference point; in the middle or in the corner? How is the original part drawing centered? For doing things like this, I usually set up a spoilboard with a couple of pins sticking up along the centerline which correspond to holes in my workpiece, outside the area that's being carved. The part drawing should be centered in the middle, not somewhere on the edge. Then when you flip the part over and set it back on the pins, you won't lose registration. From looking at your part, it's difficult to see how you're supporting it when you flip it over, but if it's not held down firmly, that would account for the malformation you're talking about.

[danielcoyle]

Where is this reference point; in the middle or in the corner? How is the original part drawing centered? For doing things like this, I usually set up a spoilboard with a couple of pins sticking up along the centerline which correspond to holes in my workpiece, outside the area that's being carved. The part drawing should be centered in the middle, not somewhere on the edge. Then when you flip the part over and set it back on the pins, you won't lose registration. From looking at your part, it's difficult to see how you're supporting it when you flip it over, but if it's not held down firmly, that would account for the malformation you're talking about.

The physical reference point is in the corner. There is a rectangular .75 inch datum that is left on the piece during machining that is used to clamp onto. The piece does not move. I am not entirely sure what you mean when you are referring to how the piece is centered in the drawing (like I said, I have been teaching myself and might be missing something obvious here. Actually hopeful that I am at this point.) The original part is centered in the stock block horizontally and then "moved" via one of the corners, to the 0,0,0 to first create the machining operations to cut out the bottom surface with the entire stock block BELOW Z=0 (see picture)
.
Then the part with block is flipped 180 in the window that shows the RIGHT profile. Finally this flipped part is moved back to 0,0,0 via the corner opposite (along the Y axis where Z=0) the corner that had been at 0,0,0 in the first machining (of the bottom surface) Now (see picture of top machining)

this gets machined.
Any feedback?? Super helpful if so.

[awerby]

You're making things difficult for yourself by having the zero point in the corner, since everything changes when you flip the part over. Try moving the model so that the X, Y and Z zero points in the original drawing are exactly in the center. That way when you turn it over, the zero points won't shift, which sounds like what's happening. You'd set your X and Y zeros in the middle of your block, then touch off to the top of the block and set the Z zero to half its thickness. (Make sure your tool change position corresponds to this offset, so you don't crash your tool).

It also helps if you make a fixture like the one I described above, with pins on the Y zero line, so that you don't get any inadvertant movement of your workpiece in relation to the zero points. If you're carving everything away so that you don't have a physical datum point to touch off to in the middle, you can fix a block in a known location in relation to your zero points and touch off to that, correcting for the offset.

As a Mecsoft customer, you are entitled to tech support from the company; you don't have to struggle all alone in the dark. Give them a call and ask them how to do a two-sided part; I'm sure they'll be helpful.

[danielcoyle]

Ok. I THINK I understand, Andrew. Will work on testing this over the next couple days and let you know what disaster I come up with I might have to ask you a question or two as I do. And, I agree that mecsoft should help me but I have tried and they have told me they no longer support the Rhino Cam version I have and that I should upgrade. Have had that conversation twice in the past couple years.

[victorofga]

Daniel

you need to leave reference surfaceon the side you start...

additionally, I believe, you should rough it out from both side, then finish ..

wood is never stable and you removing very much..

leave example a ""rim"" all around the helmet.. of course model the rim too.. then you can flip safely and find origins on both side..

================================================== ==================

in my opinion, the corner reference makes sense..
but again, in this case you removing very much material...

and again, the model should contain the ""rim" and you referencing always this ""rim""

it is not upgrading problem.. even a 30 year old camprogram could generate the toolpath..

[Melkiades]

When i need precise both side machining i used special cut template with tooth shifted on tool radius.
Then turn in at 180° and fit back in stock.
MSC in center by XY and top on Z
See pic:

This one flipped and machined:

Template cutted by 6mm end tool
Stock made from acrylic 20mm thickness.
BTW, i using work zero rotate option or setup orientation to flip machined side on my RC 2015
I not sure it's avaliable on 1.0 ver

[danielcoyle]

Alright guys. I have tried several of the suggestions, thanks. Unfortunately, I am still having the problem. I have some better pictures now. It seems like what is happening is that the top surface is actually scaled up somehow. Even though, in my models (of which there are several which come from lots of different places, the web, my own design, other companies etc.) the wall thickness is even and uniform I get what appeared at first to be a "shifted" effect where the top surface in every single helmet was shifted to X- by about 5-6 mm. After looking closer it seems like the top surface is more or less made for a larger model. So, what I think you will see in the pics is that there is an even and continuous increase in the thickness of the wall as one looks from the X+ side (front) of the helmet to the X- side (back). If it was simply a shift it wouldn't make sense, especially in the widening of the sides in the first two pictures. I marked the wall thickness, in mm, on the outside of the helmet in the second pic. Any ideas? Is this still a shift/reference problem but I am not seeing it? It happens with every single helmet model I do, regardless of the type.

[Melkiades]

I have no idea.
They only reason is if this happens on any helmet model, all models are correct, and you using same settings when creating TP, may be problem in cnc machine?
For example, wrong gear ratio at X axis?

[victorofga]

wheres the wall thicker. the roughing marks still are there...

if you made wall thickness with progam, then the model has even wall thickness

can you show picture, how youre clamping it?
I mean both position, when you mill bottom also when you mill top



a closely zoom into picture.... on the left side shows a very bumpy surface..

it is positioning error..

example if you used the corner for zeroing, the difference between the midlle part of helmet and the side part lengths could explain...
in case if you leave the ""zero point""" on the table..
when I say ""leave""" I mean same X-Y cordinatas, and you flip as the helmet front and back swapping, then that explain the difference between front and rear the result...


my conclusion is, make everything same, except flip helmet on axis around
where you could draw a line from front to rear
means left and right side shifting within flip

then it will work..

[ger21]

When you do the first side, drill a hole off to the side of the helmet.

When you flip it, drill the hole in the same place. If the holes don't line up, then your location is wrong.

[danielcoyle]

I have no idea.
They only reason is if this happens on any helmet model, all models are correct, and you using same settings when creating TP, may be problem in cnc machine?
For example, wrong gear ratio at X axis?

Well, that is something I hadn't thought of but, yes, I obviously also have no idea. Will keep this in mind.

[danielcoyle]

wheres the wall thicker. the roughing marks still are there...

if you made wall thickness with progam, then the model has even wall thickness

can you show picture, how youre clamping it?
I mean both position, when you mill bottom also when you mill top



a closely zoom into picture.... on the left side shows a very bumpy surface..

it is positioning error..

example if you used the corner for zeroing, the difference between the midlle part of helmet and the side part lengths could explain...
in case if you leave the ""zero point""" on the table..
when I say ""leave""" I mean same X-Y cordinatas, and you flip as the helmet front and back swapping, then that explain the difference between front and rear the result...


my conclusion is, make everything same, except flip helmet on axis around
where you could draw a line from front to rear
means left and right side shifting within flip

then it will work..

Ok. So looking back over everything I see Andrew's suggestion was similar and I may have misunderstood him. I don't have an automatic tool changer (manual) so it is not easy to use the physical center of the blocks as a reference point (or is it and I am just lacking the trick to do it.?) I show the way I currently fix the blocks in the below pictures and drawing. In the drawing I also draw what it is that I think you are suggesting Victor. Using the mid point/center on the y axis as the reference point. Everything in the models is already "aligned" so that the model, which is symetrical along the x axis sits dead center in the block. I guess I can and will try what you and Andrew are suggesting but still don't really understand why it seems the top part is Scaled bigger or, even (if its not scaled) shifted. Also, in response to what you saw in the photos. It is not a rough path you see but just the ragged edges of a plastic like compound I coat the inside of the helmet with. The "shift" or scaling problem is consistent between the top rough and the top finish passes. Those passes match each other but obviously do not line up correctly with the bottom surface (here again it seems like it is a scale and not a shift problem? ) Anyway, here are the pics of my fixturing and the drawing I made. In the first pic you can see the top being roughed and the fixture for carving the bottom is in the foreground. The second picture shows the bottom being roughed but the fixture itself is covered in sawdust.

[victorofga]

what I try to explaining, rotate or call it flip, along the axis, where object is symmetric... that already filter out one error...
also what Gerry saying should work...
a known point on the object transfrrred to the opposite side..


we really like to see you going with your great project.. it looking really unique..

[victorofga]

checking drawings you made... I think you got the point I tried to show...

for roughing...
check with cambam... and I tell you why... cambam has feature the trochoidal milling.. that is the so called high speed machining..
first in 2-3 step you get down to 3-3.5 in deep
then you make smaller steps... to make graduated-stepped the sides... ready for finish...

the point is of the trochoidal milling, that your federate can keep 3-600 ipm TRUE ipm.. because you going along an oval...
ot even faster if your machine can keep up...

[awerby]

You wrote:
"Ok. So looking back over everything I see Andrew's suggestion was similar and I may have misunderstood him. I don't have an automatic tool changer (manual) so it is not easy to use the physical center of the blocks as a reference point (or is it and I am just lacking the trick to do it.?)"

This isn't hard to do; you don't need an automatic tool changer. Touch off the tool against the side of the block. Set that (temporarily) to zero. Now go to the other side of the block and touch off to that. Note the distance in the DRO. Divide that number (call it X) by 2, pull the tool up, and go to that position (X/2) and set that as zero. Repeat the process for the other axis, and you've located the center of your block. As long as it doesn't shift its position when you flip it over, it should be the same on both sides. But it helps to make some kind of fixture to assure that. If you've oriented your original CAD drawing on the center point for both sides of the helmet, both sides should line up when carved.

[danielcoyle]

Victor, Andrew, thanks. I will try what you guys suggest now that I understand better. Unfortunately the drive belt on my Z axis motor just failed so it won't be until the 28th that I get it replaced and am able to try the new ideas. In the meantime I will look into Cambam.

[awerby]

While you're at it, look at DeskProto. It has a "2-side wizard" that simplifies the task of flipping parts over to cut the other side without losing registration. You attach blocks to the spoilboard to define the front left corner, and take a cut on the back edge of the stock so that when you flip it over, the corner which formerly was on the top left now fits accurately into that spot and everything works out. The demo works for a month for free, so wait until your machine's ready to work before starting the trial, in order to maximize your time to work with it.

[routalot]

What machine are you using?It may have its own reasons for moving a toolpath around.