Trying to solve an odd toolpath issue

I have recently designed and built a largish 2.44m x 1.22m CNC router running LCNC.

It is going pretty well overall, although as a mechanical engineer there is a steep learning curve for much of it.

At the moment I am trying to machine some 32.2mm wide Y axis slots into the spoilboard so I can screw in some 32mm wide T slot extrusion. I am using a 10mm 2 straight flute endmill.

However when I cut the slots they come out only 30mm wide and offset by 1mm positive X.

The toolpaths are generated in Solidworks/HSMworks.

I have checked some things;

- The tool is definitely specified in HSMworks as a 10mm end mill
- I am definitely using that cutter in the CAM setup
- I have no tool compensation applied in CAM
- I have updated and checked the Linux CNC tool table for this tool, the diameter is correct, there are no offsets specified
- I have checked the axis scales, these are correct
- The graphical interface on LCNC looks like the tool is following the path pretty much bang on (apart from the corner movements but that is a seperate issue i have a solution for now)
- I have looked at the Gcode, the finishing cut X axis movements work out correctly, ie this is the finishing cut movement;

N320 G1 Y5.
N325 X181.1
N330 Y1227.
N335 X158.9
N340 Y1219.896
N345 G0 Z15.
N350 G61
N355 G53 Z0.
N360 G53 X0. Y0.
N365 M30

181.1 - 158.9 + 10 = 32.2mm

What else should I be looking for?

Is your slot realy 30mm precise. Or 30.2?
2x 1mm would give an error of 2cm
CAM compensates probably for your 10mm mill. = 1mm
Plus U use 10mm in tool table. I guess including compensation = next 1 mm

Have you eventuell mixed diameter and radius?

Mike

You mention that it is a home built router. Are you sure you don't have a total of 2mm of slop in the X axis direction. It can come from a number of sources besides just backlash in X axis. It can also come from side to side slop in the Y and Z axis and also router spindle bearings, not to mention any flex in the frame. (I have seen 3.5hp Porter Cable routers with nearly a whole mm of slop) All of these sources of play can add together to ruin your day. You also mention issues with your corners, problems there can also be indicative of slop and backlash issues.

Todd wrote:

You mention that it is a home built router. Are you sure you don't have a total of 2mm of slop in the X axis direction. It can come from a number of sources besides just backlash in X axis. It can also come from side to side slop in the Y and Z axis and also router spindle bearings, not to mention any flex in the frame. (I have seen 3.5hp Porter Cable routers with nearly a whole mm of slop) All of these sources of play can add together to ruin your day. You also mention issues with your corners, problems there can also be indicative of slop and backlash issues.

Hi Todd

I think I have figured out the issue and it is basically what you are suggesting. I have measured the actual router movements vs the programmed and it is moving the correct distance in reality. I think it is a combo of tool, gantry and belt deflection. I am running an ER20 collet which is what came with the spindle, and I think the shaft size is also allowing a bit much deflection for the speeds and feeds I am trying to run. If I do a second finishing pass things improve somewhat, the error is not as great as I thought it was but is definitely still there.

I am finding it genuinely fascinating to learn how the machine reacts, I have spent a lot of time trying to get the speeds and feeds right for the capabilities of the machine while still maintaining reasonable accuracy. I have been slowly improving the machine as well, so I will see what i can do about this...

Cheers