Steps/unit calibration
I used VCarve Pro to drawn a 2" square (50.8 mm) and profiled it to a depth of 1/4" (6.35mm). I then measure with digital calipers the size and depth of the square. What I ended up with is a width (X) of 2.761" (70.2 mm)and a height (Y) of 2.711" (68.9 mm)and a depth (Z) of 0.2655" (6.67 mm).
I used the Mach2/3 mm postprocessor.
Question is: should I adjust the steps/unit (not sure if unit is inches or mm) proportionately to get the square to come out exactly as designed or is there some other problem?
If I do adjust the steps/unit, the new values calculate out to be:
X - 96.58, Y - 98.36, Z - 125.55
S2/600, Industrial HF spindle (StoneyCNC)
DL445, 4th axis, touch probe, 3D probe
UC100 + UCCNC (Win10)
Update: I just looked at the gcode file generated by VCarve Pro using the Mach2/3 mm postprocessor. The bit used was 1/4" (6.35mm). If I look at the XYZ values for the 4 corners of the square and subtract for the width of the bit, I get a width of 63.65mm and a height of 65.501mm whereas I programmed VCarve to be a square of 50.8mm.
It now looks like the G-code being generated is not correct regardless of the sets/unit value.
S2/600, Industrial HF spindle (StoneyCNC)
DL445, 4th axis, touch probe, 3D probe
UC100 + UCCNC (Win10)
Assuming you are using a Stepcraft 2, the steps per mm (unit) should be 133.33 on all three axes. You don't say if you are using UCCNC or MACH3/4, if UCCNC look at the "Configuration" page X, Y, Z axis steps per unit settings. From your calculated values they would seem to be correctly set already, so don't change them.
Examine the GCODE to find the co-ordinates being demanded of the router. If you are milling as a pocket, the Vcarve will have compensated for cutter diameter and be milling inside your design outline, and this will be reflected in the co-ordinates of the toolpath.
Selecting a "Profile" tool path rather than "Pocket" in VCarve settings could easily give the errors you have described.
Obviously you did "Home " the machine first and then set the Z axis to zero with the cutter at the at the top of your material when you set the Machine Co-ordinates?
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
Yes, I have a Stepcraft-2/600 and running UCCNC under Win10 on a MacBookPro along with a UC100 USB adapter. The values for steps/unit in the profile were all set to 133.3333 as you indicated.
I examined the gcode for another 100mm square generated by VCarve Pro 8.5 and, while it does compensate for the width of the tool, it is generating a square that is too large. I design in mm and use the Mach2/3 Arcs (mm) postprocessor.
I have a Replicator 2X 3D printer and we change the parameters for steps/inch to get the correct extrude length. I new to this aspect of CNC and assumed that the steps/unit could be modified also, but evidently not. Any ideas on why the postprocessor isn't generating the correct toolpaths?
S2/600, Industrial HF spindle (StoneyCNC)
DL445, 4th axis, touch probe, 3D probe
UC100 + UCCNC (Win10)
So you are confirming that the Stepcraft executes the GCODE correctly, and that the design in VCarve is correct?
That being so it would appear that the post-processing is at the root of the problem. I can't offer any sensible comment on that except to say that I use the Mach2/3(mm) in Cut2d and all my parts are correctly dimensioned to better than 0.002" as far as I can tell on my digital calipers.
Maybe there is some subtle parameter in VCarve that is messing things up for you?
SC 420/2, Industrial VFD spindle from StoneyCNC
UC100 + UCCNC
Cut2D, Autosketch10, Draftsight, Eagle 9.5.1
There is no problem, however simple, that cannot be made more complicated by thinking about it.
The pitch of the screws is the pitch of the screws. 133.333 is correct.
you can't adjust the steps per unit based on measured dimensions like this - especially a square as a square induces the backlash values in X and Y. This a way to calibrate a CNC - but not in this case - for calibration you need to be going in a direction + stop... then continue in the direction + calibrate. this means you don't expose the calibrated value to drive backlash. the dimensions are most likely caused by backlash. this would be the best place to start.
- 44 Foren
- 7,389 Themen
- 63.3 K Beiträge
- 1 Online
- 26.5 K Mitglieder