Results 41 to 53 of 53
Thread: G02 and G03 Radius Calculations

11082019, 12:35 PM #41The problem is this is wrong.
As others have pointed out you have to compensate for the tool nose radius.
So unless you use G41/42 Tool nose raduius compensation, you have a whole load more triangles to construct.

Think Snow Eh!
Ox


11082019, 12:39 PM #42
Doing it algebraically, all you need to do it adjust your offset values to compensate for the tool radius.

11082019, 01:18 PM #43

11082019, 01:20 PM #44
???
When you solve for the .375 rad on the part, just solve for a .344 rad in stead.
When you solve for the .750, just solve for a .781 in stead.

Think Snow Eh!
Ox

11092019, 03:23 AM #45
Hello Ox,
I assume that by using the feature radii fudged by the TNR, you're saying that a program results where TNR Comp by the control is not necessary (G41/G42). I don't see that you're right on this occasion.
Following are two program snippets for tool paths based on a shape similar to that of the OP's (diameters unknown for the OP's example). In the profile I used for my calculations, I used a 2" diameter for the Small Journal, 6" for the Large Journal and an angle feature of 45deg between the two radii. The first code example used radii of 0.375 and 0.750 respectively in the part profile with the 0.031 TNR compensated for in my calculations. Accordingly, TNR Comp at the control is not required.
In the second code example, radii of 0.344 and 0.781 were used in the part profile.
Given that the numbers in the first example are correct for a 0.031 TNR and therefore, TNR comp at the control is not required, the numbers of the second code example, using the fudged radii values, should be the same; clearly they are not.
Its not just a case of substituting the actual part radius with the fudged, or compensated radius. You have to first calculate the centre of the actual part radius using its value, the compensated (fudged if you like) radius is then used to calculate the centre of the TNR. Depending on how the tool is set, the TNR (2 x TNR for the X axis) is either added or subtracted to get the true coordinate for the tool in the tool path.
Regards,
Bill
Example using 0.375 and 0.750 feature radii and compensated for the 0.031 TNR
G00 X3.4715 Z0.9261
G01 X2.0000
G01 Z0.8759
G02 X2.2014 Z1.1190 I0.3438 K0.0000
G01 X5.5424 Z2.7895
G03 X6.0000 Z3.3419 I0.5524 K0.5524
G01 Z5.0000
Example using 0.344 and 0.781 fudged feature radii
G00 X3.4715 Z0.9261
G01 X2.0000
G01 X2.0000 Z0.8575
G02 X2.2015 Z1.1008 I0.3440 K0.0000
G01 X5.5425 Z2.7713
G03 X6.0000 Z3.3235 I0.5523 K0.5523
G01 Z5.0000Last edited by angelw; 11092019 at 05:41 AM.

barratt liked this post


11092019, 05:02 AM #46
This is the only way I know how to do it.
After calculating the intersection points you have to compensate for the TNR in order to account for the difference between the virtual cutting point & the actual cutting point.
Compensation for the first intersection is just the TNR value in Z. There is no compensation needed in X.
Then you have to take the tool so that it lays tangential to the second intersection point. That requires compensation in both X & Z. The R value in the G02 line would be RTNR. I don't know how to do this without constructing more triangles.
A CAM system does all this for you.Last edited by barratt; 11092019 at 05:14 AM. Reason: better drawing

11092019, 09:21 AM #47

11092019, 04:34 PM #48
Hello Ox,
The result is correct when a 90deg corner is involved and its for the same reason that TNR Comp is not required at the control for surfaces that are parallel to the X and Z axis respectively.
90 Deg.
0.375 and 0.750  With 0.031 TNR Comp used in Calculations
G00 X3.000 Z0.2000
G01 X2.0000
G01 X2.0000 Z0.8757
G02 X2.6880 Z1.2197 I0.3440 K0.0000
G01 X4.4380 Z1.2197
G03 X6.0000 Z2.0007 I0.0000 K0.7810
G01 Z5.0000
0.344 and 0.781  With 0.0 TNR Comp used in Calculations
G00 X3.000 Z0.2000
G01 X2.0000
G01 X2.0000 Z0.8757
G02 X2.6880 Z1.2197 I0.3440 K0.0000
G01 X4.4380 Z1.2197
G03 X6.0000 Z2.0007 I0.0000 K0.7810
G01 Z5.0000
The numbers in my previous Post are not quite a true comparison, for I used the exact TNR of 1/32" (0.03125) in my TNR Comp calculations with 0.375 and 0.750 radii. In the example where no TNR Comp was calculated, I used feature radii of 0.344 and 0.781, which used 0.031 in the subtraction and addition respectively. Following is a correct comparison where 0.031 is used in the calculations.
45deg.
0.375 and 0.750  With 0.031 TNR Comp used in Calculate
G00 X3.0000 Z0.2000
G01 X2.0000
G01 X2.0000 Z0.8757
G02 X2.2015 Z1.1189 I0.3440 K0.0000
G01 X5.5425 Z2.7894
G03 X6.0000 Z3.3417 I0.5523 K0.5523
G01 Z5.0000
0.344 and 0.781  With 0.0 TNR Comp used in Calculate
G00 X3.0000 Z0.2000
G01 X2.0000
G01 X2.0000 Z0.8575
G02 X2.2015 Z1.1008 I0.3440 K0.0000
G01 X5.5425 Z2.7713
G03 X6.0000 Z3.3235 I0.5523 K0.5523
G01 Z5.0000
Regards,
BillLast edited by angelw; 11102019 at 05:40 AM.

11102019, 09:08 AM #49


11102019, 12:04 PM #50
I'm not sure the context of EG's remark, but corner rounding/chamfering is not standard on many lathe controls. It is an optional feature on most lathe controls. In some cases the machine builder will include it as standard, but usually it will be an added cost option for the customer to purchase.
I don't recall the OP stating what control he is working with so no way to tell if his machine is capable of rounding or not. Without know that I'd not want to muck up his question by advising use of a feature he may not have.

11112019, 04:34 AM #51
Even if the corner is at 90 degrees you still have to calculate tnr comp.
For the 1st intersection you need to advance the tool by the tnr in Z. no comp in X.
For the 2nd intersection you need to subtract the tnr (x 2 for diameter) in X. No comp needed in Z
The R value needs to be made R  tnr
So if you have a part that is 100mm dia for 50mm then a square corner to 200mm dia with a 20mm rad in the corner and the tnr is 10mm
1st intersection point for the profile is X100.0 Z30.0 (Start of radius)
2nd intersection point is X140.0 Z50.0 (End of radius)
G00 X100.0 Z3.0
G01 Z40.0 F.25
G02 X120.0 Z50.0 R10.0
G01X200.0
Our Fanuc 16T can use the G01 rounding function however it only works at 90 degrees and then only if those angles are parallel to the X & Z axis. Pretty useless really.

11112019, 05:23 AM #52
Hello barratt,
What Ox is suggesting, a radius fudged by the value of the TNR, will work for radii between adjacent lines that are perpendicular to each other and parallel with X/Z axes respectively, but in essence, is doing exactly what you're explaining.
Substituting the actual feature radius with one that's been modified by the value of the TNR, where the radius is tangent to one, or two surfaces that are not parallel with an axis movement, will not result in a correct program. This is easy to understand when you consider that two radii of different radius, tangent to the same adjacent lines can't share the same centre. Notwithstanding that this is also the case where the adjacent surfaces are perpendicular and respectively parallel to an axis, the resulting program is correct because the surfaces are parallel to the axes.
Regards,
Bill

11112019, 06:48 AM #53
Hi Bill,
You're right, in the square corner case modifying the profile rad amounts to the same thing.
Bookmarks