Rules not working as expected

[Blaz]

Hi,
I am experimenting with Sheetcam rules and inspecting the simulation and G-code they produce. I also modified my post processor quite a bit, but unexpected feeds are already displayed in simulation, so I don't think post processor has anything to do with it. I didn't yet bought the licence, so the demo job was made on trial version. Is possible that this would cause the unexpected feeds? On the website it is written, it is only limited to the size of G-code file.

I noticed that order of the rules seems to be important. E.g. if you have multiple rules for small circles, you need to sort them from biggest to smallest in the rules list.
What is the purpose of "Path rules" drop down menu in the edit operation dialog? It doesn't seem to have any affect, when rules are already applied in the "edit tool" dialog. Also, what does "Always apply rule set" do in the edit rules dialog?
Can somebody confirm that arc size means arc radius, and circle size means its diameter.

In the attachment you can try simulation and see resulting feedrates for yourself. It has a demo part with all features (shapes) where I want to control the feedrate. However sometimes it uses particular rules only on part of a region (e.g. not whole circle) or doesn't use correct rule at all. At least it looks like so. Nominal feedrate is 1000mm/s, each rule has different % reduce feed, so it is easy to tell which rule is active when. The following are described issues, marked on the screenshot:
1. Arc has diameter 20mm, so 40% feedrate should be applied, but it is 70% (the second half of hole has correct 40% feedrate)
2. Ok
3. Last quarter of the circular hole has 70% feed instead of 60%. Meaning, arc rule is applied for the last quarter. If generated lead in is not moved, this issue seems to not appear.
4. Similar to 3. but with feeds for smaller circle / arc size (40% feed in last quarter)
5. Ok, but corner, following the arc doesn't have 50% feed reduce, but 40%
6. First arc quarter has 50% feed (corner rule is applied) instead of 70% like the rest of the arc.
7. Ok
8. Ok

Do I miss anything or are rules buggy?

Thanks a lot, Blaz

demo_rules.png (12.29 KiB) Viewed 176 times

demo_part.png (44.93 KiB) Viewed 176 times

[Blaz]

And here is dxf, if anyone wants it

[djreiswig]

Didn't look at your job yet, but did you take into account kerf width & offset in your calculations? I'm pretty sure this can have an affect on when rules are applied.

[Blaz]

Yes I checked that circle / arc size in rules is bigger than the actual hole / fillet of the shape + kerf width.

What offsets? I am pretty sure that sheetcam doesn't support additional offsets to the paths (beside the kerf offset direction)

[djreiswig]

Yes, that's what I meant. Inside or outside offset.

[Blaz]

Offsets are not the problem as circle / arc size is much bigger than the actual circle in the part.

Can anybody try to replicate the issue with the same set of rules, please?

[bLouChip]

Les commented recently on path rules in another forum, perhaps his comments apply to your situation also, or at least in part.
https://plasmaspider.com/viewtopic.php?p=235821#p235821
and here...
https://plasmaspider.com/viewtopic.php?p=235861#p235861

[Les Newell]

1. Arc has diameter 20mm, so 40% feedrate should be applied, but it is 70% (the second half of hole has correct 40% feedrate)

Hmm, that's an odd one. It does look like something strange is going on in SheetCam. It is triggering the 40% override then the 70% override, which is the wrong way around.

3. Last quarter of the circular hole has 70% feed instead of 60%. Meaning, arc rule is applied for the last quarter. If generated lead in is not moved, this issue seems to not appear.

This is because the overcut is not considered part of the circle. That last bit is the overcut which has the arc rule applied.

5. Ok, but corner, following the arc doesn't have 50% feed reduce, but 40%

When rules overlap the lowest feed rate should win. In this case you have a 40% on arcs overlapping with the on corners. Fro soem reason this overlap code is failing in example 1, which I am investigating.

6. First arc quarter has 50% feed (corner rule is applied) instead of 70% like the rest of the arc.

This is due to the way rules overlap. Internally arcs never exceed 90 degrees so for instance a circle is 4 90 degree arcs. That explains why it only affects the first quarter.

Do I miss anything or are rules buggy?

Maybe a bit of both...

[Blaz]

Thank you for the responses.

1. Hmm, that's an odd one. It does look like something strange is going on in SheetCam. It is triggering the 40% override then the 70% override, which is the wrong way around.

Yes, but only for the frist half-circle. The second half circle has correct (40%) rule applied.

3. This is because the overcut is not considered part of the circle. That last bit is the overcut which has the arc rule applied.

Yes, I expected the arc rule to apply to the overcut section, however, the last quarter of the circle (before the overcut) also has the arc rule applied instead of the circle rule. Probably what I replied down at 6. ("90deg resolution" stuff..)

5. ... for soem reason this overlap code is failing ...

Ok, so it is confirmed to be an issue on SheetCam side.

6. This is due to the way rules overlap. Internally arcs never exceed 90 degrees so for instance a circle is 4 90 degree arcs. That explains why it only affects the first quarter.

If I understood you correctly, arcs/cicrcles are divided internally into 90deg quarters and particular rule can only be applied to the whole 90deg section or not at all (and not like only to first 10deg section). This results that the the arc touching the corner has slower corner rule applied to the whole first 90deg, not only for the first 3mm (the distance that corner rule applies to)
In that case, Sheetcam works as expected however due to simplified implementation details, there is some limitations ("90deg resolution" on arcs/circles).