Messages

41

Feature Requests / Z grid to prevent work from separating from the stock

« on: July 30, 2010, 02:57:25 PM »

I would like to see a built in grid that can be set with a Z depth and grid spacing (as well as on and off).
Purpose: When cutting flat stock there is the problem of the part being cut free from the stock and it bouncing into the tool, both damaging the part and breaking the tool. This problem could be minimized by setting up a grid that the tool will not cut into as it is cutting out the part. Currently, to prevent the part from coming loose from the stock, one must stop cutting before cutting all the way through the material, and cleanup can be very difficult if the flash is too thick. One could manually program in several links between the work and the stock so that the work will not come loose and could be snapped free after the job is done. However, it is very difficult to program in such links unless very specialized software is used. However, if an internal grid existed in the CNC software, the operator would only need to set the link grid to stock thickness minus the link size as desired, and would proceed to cut the part out to full depth. The CNC software would then automatically step over the grid while cutting the part to the desired depth.

I think that the link should be some shape (perhaps half circle or ramp) that would allow for smooth tool travel while cutting over the link. By setting a grid spacing, one can control how many links will hold the work to the stock. The grid might be best centered on 0,0.     

If the link were a 45-deg ramp, the Z depth setting would determine the link size in the finished cut. If the link grid settings were controlled by G-CODE, even better.

Is something like this reasonably possible?

Thanks, Bill

42

Tangent Corner / Re: eliminating taper on deep cuts

« on: June 24, 2010, 04:36:04 PM »

Rich,
I rarely cut aluminum deeper than 0.010 or I tend to break tools. I judge the path speed (usually around 35ips) & motor speed by the chip size, not too small or too big. I usually have the run the spindle speed as slow as I can get it to run smoothly. I use a 1.5" long 1/8" 2-flute end mill (looks to have the outer edges longer than the center).  If I try to use a 3" long tool (drill-point and not carbide), I can see it wander all over the place. 
If the tool consistently cuts too large in the CW direction, do I need to allow for a finish cut in the CCW direction? In theory, the finish cut will not try to pull away sideways from the ideal tool path as much as it does in a bulk cut. Do I want to finish cut only at near full depth or on every pass?
Thanks, Bill

43

Tangent Corner / eliminating taper on deep cuts

« on: June 23, 2010, 05:25:40 PM »

I have a CNC machine that is made out of mostly wood and will cut thin aluminum (0.1" or less) fairly accurately. I want to cut a piece of aluminum (T61) that is 1/2 inch thick. The cut is 2D and I need to use a 1/8" end mill. The problem is that if I cut clockwise (CW) around the piece, it is larger than I want, and if cut counter-clockwise (CCW), it is smaller than I want. To make things worse, the part is tapered toward the direction of the error, that is, the bottom of the part has more error than the top.

I am aware that this due to the lack of stiffness in the CNC machine design, and in part, to the flexibility of the small sized tool.  

I was wondering if I can resolve much of this error by first cutting each pass CW and then following with a CCW finish pass (same tool path same depth). It occurred to me that the CCW pass might not take out all of the error, and it may be beneficial to lower the tool a small amount into the material on the CCW pass to help hold the tool on course while it is finish cutting the part.  

Is there anyone who knows the best way to do this kind of cut with this kind of problem?

If I need to use a different tool path for the finish cut, is there an optimal finish tool path offset (% of tool diameter) for the CCW cut?

My primary goal is to get a part with a vertical side and around 0.010 tolerance on overall dimension.

Thanks, Bill

44

G-Code, CAD, and CAM discussions / G-code equivalent to goto or jump

« on: September 25, 2008, 11:27:04 PM »

Is there a G-code command that will allow me to skip over a block of G-code? I want to be able to skip over subroutines embedded in the code (to keep things in a specific order) or code that will be used only once per job, like a setup routine. 
Thanks,
Bill

45

G-Code, CAD, and CAM discussions / M3, M4, M5 have permanent pause and require a Cycle Start to continue

« on: October 01, 2007, 01:40:37 PM »

Hi,
I find that whenever I use M3, M5, S0, or S100 that The controller (Mill mode) pauses indefinitely or until I press the Cycle Start button again.
I found this when I tried to stop the spindle motor after moving to Z0.5000 and before moving to X0.0000 Y0.0000. I did this to use the going home time to replace the spindle spin-down delay. As a result, the program never starts or finishes without manual intervention. I tried a variety of G-Code combinations, none of which seem to get past the permanent pause. I have the spindle delays set to 1 sec (spin-up) and 0.1sec (spin-down).  This code snipet was my latest attempt:

N  6  G0 Z0.5000
N  7  G0 X0.0000  Y0.0000 M03 S100
N  8  G83 X0.7000 Y1.0000 Z-0.035 R0.01 Q0.025 F15 G99
...
N 49  (End of Program)
N 50  G40 (Cancel Compensation)
N 51  G43 (Cancel Tool Offset)
N 52  G80 (Cancel Canned Cycles) 
N 53  G00 Z0.5000 (UP)
N 54  G00 X0.0000 Y0.0000 M05 S0 (Go Home)
N 55  M30 (Rewind & Stop Spindle)

Any Ideas?

Bill

46

G-Code, CAD, and CAM discussions / Re: Possibly bad G-Code turning arc into circles

« on: January 19, 2007, 08:19:35 PM »

Bloy,
I use that function. It works pretty fair at solving chaining problems. Trouble is that it is too easy to screw it up again. Once a chain is formed, the direction of all of the segments should be locked together, changing one should change them all by default.
 
I tried to make a gear (using the gear function) and then convert it to a surface to extrude it into a solid and it wouldn't do it. I'm not sure what I did wrong or if it can be done. I had a devil of time trying to close the chain, because the segments are all very small, and I could not easily organize them. "Reorganize" helped with it, but every time I tried to select the gear, I would wind up reversing a segment or two, which broke the chain again.

I will get used to the program someday.  I have the 10 CD training series and that helps, providing I can find what I am looking for.

I now have a patch for the G83 function, and that seems to work for me for the problem below.

Thanks.

Bill

47

G-Code, CAD, and CAM discussions / Re: Possibly bad G-Code turning arc into circles

« on: January 19, 2007, 03:11:52 PM »

Bloy,
I now know what the problem is. The code generator is forgetting to add a G00 to line 11 and Mach does not know what to do with data that follows the G80.

Now to fix the code generator.

One problem that Bob Cad has is that it is click intensive. There are a few flexibilities that should never exist that are constantly in the way. The big one I notice is being able to reverse a single element in a chain. Another is that it forgets settings on some functions that tend to be repeat items.   

Thanks.

Bill

48

G-Code, CAD, and CAM discussions / Re: Possibly bad G-Code turning arc into circles

« on: January 19, 2007, 01:42:31 PM »

Bloy,
Item 3) has been resolved by by changing a parameter that I though was unrelated. There is a parameter that is used to link chains together. It is apparently is also used when generating tool offsets. I had it set very large and it was closing up the arcs with lines and then cramming in the arc

I am using BOBCAD, probably a bad investment, since it is not going to be much faster at generating code.

I now have Z values where they belong but now the Mach 3 does not like the codes it's generating for a G83 peck drill. If I peck a single hole before cutting the part outline, the first arc of the part outline chokes with a "Radius to end of arc differs from radius to start on line. .(N14).  (No peck no error)
N  1 G90
N  2 G80
N  3 G40
N  4 G49
N  5 G91.1
N  6 G0 X0.0000  Y0.0000
N  7 M6 T1
N  8G83 Z-.5 R0 F15 I.01 P.1
N  9 X2.8230 Y0.5195
N 10G80
N 11 Z0.1000
N 12 X0.6000 Y1.1625
N 13 G01 Z-0.0100F15
N 14 G03 X0.0375 Y0.6000 I0.0000 J-0.5625F45
N 15 X0.6000 Y0.0375 I0.5625 J0.0000
N 16 G01 X4.1000
N 17 G03 X4.1442 Y0.0558 I0.0000 J0.0625

Thanks.

Bill

49

G-Code, CAD, and CAM discussions / Re: Possibly bad G-Code turning arc into circles

« on: January 19, 2007, 11:21:57 AM »

Sorry wrong picture.

50

G-Code, CAD, and CAM discussions / Possibly bad G-Code turning arc into circles

« on: January 19, 2007, 11:19:58 AM »

Just had a program generate the following G-code segment. I can't tell if the code is bad or if Mach3 (latest) is misinterpreting it.
1) The code is producing two small circles where arcs are supposed to be. I have the circle code set to incremental. It is getting all of the other arcs right.
2) This particular segment is supposed to spiral down five loops. Without actually cutting a piece, I can't tell if it is doing it. I don't see it on the tool path window when I change the perspective
3) The G03 code lines 331,345,359,etc. only reset the speed and it looks as though it should also set the new depth. Is this proper usable code? Every 5 passes the code does set a new depth.

What is the solution to this?

Thanks.
Bill