Hello Guest it is March 28, 2024, 09:48:36 AM

Author Topic: Rounded corners still problematic after switch to exact stop over 89 degrees  (Read 3747 times)

0 Members and 1 Guest are viewing this topic.

Hi guys I have a bunch of orders to fulfill and I thought I had this issue ironed out by telling mach3 to switch to exact stop mode on angles above 89 degrees but on random corners it still seems to try to run them in constant velocity mode.

It seemed fine the last program I ran but since then I did reduce my acceleration on my x axis but I also don't feel like that's relevant, otherwise feeds and speeds are the same.

It's not a gcode issue, i'm using vectric aspire and the rounded corners aren't consistent from vector start points or anything it really just seems random.

I can't afford to make anymore scraps, if anyone can tell me some settings to look for or change it would be highly appreciated.

Since these programs are just squares and rectangles you think I should just switch to exact stop mode for these? My only concern then is when I do run a program with other shapes or engraving and I know that stuff runs terrible in exact stop mode so I'd like to figure it out while leaving it in constant velocity if possible.

random rounded corner


good corner




Hi,

Quote
It seemed fine the last program I ran but since then I did reduce my acceleration on my x axis but I also don't feel like that's relevant, otherwise feeds and speeds are the same.

The ability of a CNC machine to follow a trajectory is critically dependent on it axes acceleration. The higher the acceleration the better a machine can follow a path.
You need to have as high an acceleration in your settings as you can consistent with not losing steps and having sufficient reserve to accelerate the mass of the table
and workpiece.

Craig
'I enjoy sex at 73.....I live at 71 so its not too far to walk.'
Hi,

Quote
It seemed fine the last program I ran but since then I did reduce my acceleration on my x axis but I also don't feel like that's relevant, otherwise feeds and speeds are the same.

The ability of a CNC machine to follow a trajectory is critically dependent on it axes acceleration. The higher the acceleration the better a machine can follow a path.
You need to have as high an acceleration in your settings as you can consistent with not losing steps and having sufficient reserve to accelerate the mass of the table
and workpiece.

Craig

I understand, but if it was running in exact stop mode like it should be given the setting then the acceleration should be irrelevant in this case

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Dont reduce acceleration, increase it.

Try setting stop CV to a lower angle.

Offline rcaffin

*
  •  1,052 1,052
    • View Profile
You need to have as high an acceleration in your settings as you can consistent with not losing steps and having sufficient reserve to accelerate the mass of the table and workpiece.
I don't think that is exactly right.

If you tell Mach3 that your X axis can accelerate faster than the physical hardware can manage, you WILL have all sorts of problems.
If you tell Mach3 that your X axis can only accelerate slowly, Mach3 WILL handle that just fine. The trajectory calculations take the stated acceleration value into account. The only virtue in telling Mach3 about a high acceleration value is to get faster movement.

What you have not stated is whether the rounded corner is on the entry or the exit to the corner. That is super-critical for diagnostics.

Cheers
Roger

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Not so i think, It is trying to do constant velocity - if the machine settings say it can only accelerate slowly then in order to maintain constant velocity something MUST give, what gives is accuracy in path following.

This is extremely evident on plasma cutting where CV is critical - i had all sorts of rounding errors until i tweaked acceleration UP to the hilt.

IMV, acceleration is king of the hill when in CV, in exact stop all it will do is save some time.

Offline ger21

*
  • *
  •  6,295 6,295
    • View Profile
    • The CNC Woodworker
The trajectory calculations take the stated acceleration value into account. The only virtue in telling Mach3 about a high acceleration value is to get faster movement.

That depends.
In CV mode, Mach3 tries to maintain a constant velocity. Physics dictates that you can't turn around a corner without decelerating, and then accelerating back up to speed. In CV mode, you are telling Mach3 not to slow down. So when Mach3 gets to a corner, in order to maintain that velocity, it starts making the turn before it gets to the corner, rounding the corner off. Acceleration dictates when Mach3 starts to make the turn. The faster the acceleration, the further into the corner it can go before it turns, resulting in less rounding.


Try setting stop CV to a lower angle.

I already resolved this issue at CNC Zone a few days back.

Stop CV on angles only works when you have two moves at an angle, say a square with 4 linear moves.
Most CAM programs will create toolpaths that roll around the corner, so they have 4 linear moves with 4 arcs at the corner.
So your toolpaths are straight segments with tangent arcs at the corners, so there are no toolpaths at angles to each other.

Thus, Stop CV on angles will have no effect. If the accelertion is so low that Mach3 can't follow those tangent arcs at the corners, it will start turning the corner earlier, rounding it off. This is the issue that the OP had.
Gerry

2010 Screenset
http://www.thecncwoodworker.com/2010.html

JointCAM Dovetail and Box Joint software
http://www.g-forcecnc.com/jointcam.html

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
I was right about the acceleration then ;)

Interesting about the StopCv angles though, explains a lot.
Killer;

As you can see, this can get complicated, and language can get in the way...

Killer;
All are trying to help here. 

Not to throw a spanner in the works, but is there more than 1 tool involved in cutting each corner?  Looks a little like there is some tool deflection.  Could be just the dust/ camera angle.  Tried to attach pic, but not not sure it made the trip. I'll try again if needed.

My summary of what I read:
1) How fast can the hardware physicaly accelerate/ decelerate (Craig's & Roger's posts)
2) How fast are you commanding the hardware to accelerate/ decelerate (Roger's & Davek's posts)
3) How is the code generating software set up to make a corners (Line segment turn 90 degrees, next line segment, or line segment, arc, line segment) (Gerry's post)
4) How is MACH set up to cut corners? (CV) (Davek's & Gerry's posts)

... and then my question, 5)Is there tool deflection as it cuts the corner?


Offline rcaffin

*
  •  1,052 1,052
    • View Profile
It may be just my imagination, but ...
Coming vertically down from the 'rounded' corner, it almost looks as though the 'rim' tapers outwards to about 20% oversize, before shrinking back to its nominal size.  Or is this just an artifact of the photography?

Cheers
Roger