[3] SOLVED - Milling direction reversed

[jm_araujo]

Hi!

When waterline roughing the bottom shape, the exterior pass reverses direction and is done as conventional milling (clockwise in my case), even tough every other pass is done correctly.

Couldn't find any report on this bug. Is there a solution?

See attached picture and cb file inside the zip.
I'm using CamBam  0.9.8P (beta-2) - 14 October 2013 build date.

[lloydsp]

It does the same in "N", so I guess it's a bug.  I tried changing everything in the MOp I could think of, and nothing altered the path of that last circuit.

Unless I misunderstand what CB is doing there, I'd call that a bug.

If so, please report it in the 'bugs' thread.  But wait for some other 3D guys to report.  I do almost no 3D work, just 2.499999 D.

Lloyd

[Dragonfly]

When pocketing too.
This was acknowledged by Andy some time ago.

[killerguppy101]

When pocketing too.
This was acknowledged by Andy some time ago.

Has there been any update or workarounds on this?  I just got CamBam and this is proving to be a big issue for me.  Any pockets I do, the last pass reverses direction.

[pixelmaker]

hello
with the waterline roughing this is something special. The error occurs only at the third, the bottom shape.
It looks that this bug has something to do with the size of the object. The third shape is a bit larger then the other.
If you increase the roughing clearance to 1.9 the toolpath is in right direction.
I call it a minor bug because with roughing the toolpath direction is not important. I work with mixed toolpath to minimize rapids. A waterline finishing toolpath don´t have this wrong toolpath.

With pockets there is different thing. The size of the pocket is not important. Since the version N I know the bug only with polyrectangle. There are different "workarounds".
• convert the polyrectangle in to a polyline (CTRL+P)
• produce a new node at the middle of one side. For this you draw a short line that crossing the middle of one side. Select the rectangle and the short line and use -> edit -> break at intersections (CTRL+B). The rectangle gets a new node at the intersection point and the wrong toolpath disappears.

ralf

[killerguppy101]

• convert the polyrectangle in to a polyline (CTRL+P)

Thanks, this seems to have worked for me!  I'll keep playing around and see if this issue pops up anywhere else.  I mostly do 2.5D machining, so I probably won't get to test out the waterline and scanline MOPs.

[dh42]

Hello,

A drawing that show when reversed toolpath appears in pockets.

EDIT: not only on pocket, in this case the pb is the same with a internal profile, and in addition the lead in/out are reverted

++
David

[lloydsp]

Yep, in "N", too.  THAT is a dangerous bug.  One should be warned to examine toolpaths carefully before submitting to g-code. 

That could destroy an expensive part!



Lloyd

[dh42]

Hello Lloyd,

THAT is a dangerous bug.

Yes sure ! ...

I started to review bugs reports to create an excel sheet. I also adds numbers to the topics (here [3]) ... same number = same bug (I think ... only Andy can be sure). I also moves the posts that have nothing to do with bugs ..

the same problem appears also with reversed lead in/out, related to the drawing position this time ... not sure it's the same bug, but the result is the same.

http://www.cambam.co.uk/forum/index.php?topic=3723.0
http://www.cambam.co.uk/forum/index.php?topic=3584.msg23448#msg23448

With the WL rough, I remember that there is also another heavy bug ; the tool plunge directly to the the last machining level in the more external toolpath ..

The P version seems to solve problem with inconsistent fast plunge height ... but not each time ; it solves the problem on a file from a French user (I can't retrieve it) ... but not on the file given here.

http://www.cambam.co.uk/forum/index.php?topic=3724.0

++
David

[Dragonfly]

Hello,

A drawing that show when reversed toolpath appears in pockets.

EDIT: not only on pocket, in this case the pb is the same with a internal profile, and in addition the lead in/out are reverted

++
David

From my observations it seems to be connected to the step over and the actual value of the step over of the last pass.  Sometimes I play with step over value until I visibly see more or less equal steps, because CamBam seems to do strange things with pocketing  
- change of cutting direction - does not happen always and at least I could not find a logical link and therefore it's difficult to reproduce an example.
- cutting direction reversal is the logical approach when pocketing a region (the area enclosed between two closed curves) and 'Region fill style' is set to 'Inside + outside offsets'. The first (full tool engagement) cut is done in the middle of the area, then working toward the inner and outer boundary respecting the desired cutting direction. In this case the inner and outer boundary passes are done in opposite (CW or CCW) directions.
- it seems that the underlying algorithm is aimed at doing the final pass(es) at exactly the set step over and often, depending on the value, CB introduces parasitic passes with almost no material cutting. IMHO the order should follow the opposite logic - 1. do a first single cut with full load along a mid-line 2. work towards the edges at full step over 3. do the final pass to what material has remained (which should be less or equal to step over value).
For this reason instead of pockets I've started to use profiles with defined width because they are more predictable.
I attach a few screenshots of tool paths generated with different step over values. Note that the final passes' width is always the widest cut.

[lloydsp]

My approach, which may (actually) be approximately the same as yours is to do pockets with a substantial roughing clearance, then finish with profile cuts -- because in profile, I can specify the cut direction with some confidence.

I keep saying it, because it's true -- MOps don't cost you anything in g-code.  Sometimes, more MOps are better, in order to more-closely define what needs doing.

Lloyd

[dh42]

From my observations it seems to be connected to the step over and the actual value of the step over of the last pass.

Yes, but I think it's a different bug ... that cause the same result, because on my example, there is no step over, it's a profile. For me the problem appears often with the "mouse ears" in the edge, if the toolpath cross itself in the edge, the direction is reversed.

It seems there is at least 3 totally different problems that cause the bug to appears.

1 - the position on the drawing ; moving the part for some 0.00001... mm or rounding the coords solves the pb.(sometimes )

2 - the step over in pocket as you show

3 - the arcs in the edges

And, the problem related on this topic with WL, that maybe is another things...

++
David

[Dragonfly]

My approach, which may (actually) be approximately the same as yours is to do pockets with a substantial roughing clearance, then finish with profile cuts -- because in profile, I can specify the cut direction with some confidence.
Lloyd

I have adopted this approach long ago, can't remember exactly but it was one of you more experienced members who brought me to it. No argue that finishing with a profile with low tool load, deflection, etc. is the best way.

[EddyCurrent]

Regardless of the issues mentioned, where it matters I always finish a pocket with a profile cut.
The reason being that I cut wood, and taking a light finishing cut with a down spiral cutter prevents breakout on the top surface.
Of course this does not answer the bug issue but it means I'm unlikely to be affected by it.

[coolant slinger]

dh42,

In regards to "Sans titre-1.jpg". What color are you using for your cut width color? I haven't been able to find one like this so I can still see tool path line and arc color.