Foam XYUV feed rate issue on UV axis
07 Aug 2016 15:34 #78442
by delinq
Foam XYUV feed rate issue on UV axis was created by delinq
Please help
I've buildt a 4 axis hotwire cnc.
Trying to cut wings that I generated in profili2, I've got a problem with the UV axis when cutting the right wing(root of wing on UV axis)
What happens is that the UV axis looks to ignore feedrate and speed off.
When cutting the left wing where the root of wing is on XY axis, everything is normal.
When cuttering letters etc, without tapering, everything seems normal.
Running linuxcnc 2.8.0~pre1.2497.geec48e1
I've buildt a 4 axis hotwire cnc.
Trying to cut wings that I generated in profili2, I've got a problem with the UV axis when cutting the right wing(root of wing on UV axis)
What happens is that the UV axis looks to ignore feedrate and speed off.
When cutting the left wing where the root of wing is on XY axis, everything is normal.
When cuttering letters etc, without tapering, everything seems normal.
Running linuxcnc 2.8.0~pre1.2497.geec48e1
Please Log in or Create an account to join the conversation.
- Todd Zuercher
-
- Offline
- Platinum Member
-
Less
More
- Posts: 5008
- Thank you received: 1441
07 Aug 2016 18:55 #78447
by Todd Zuercher
Replied by Todd Zuercher on topic Foam XYUV feed rate issue on UV axis
Is the problem that the UV is going slower or faster than you think it should?
(I don't think that naive cam detection or multi line look a head of the new tool planner work for axis other than XYZ, so lines with movement in ABCUVW will often be slower because of single line look a head and the old path-blending system.)
(I don't think that naive cam detection or multi line look a head of the new tool planner work for axis other than XYZ, so lines with movement in ABCUVW will often be slower because of single line look a head and the old path-blending system.)
Please Log in or Create an account to join the conversation.
07 Aug 2016 19:07 #78449
by delinq
Replied by delinq on topic Foam XYUV feed rate issue on UV axis
It goes a lot faster, I have to slow it down with the "Feed Override" to about 35%.
Please Log in or Create an account to join the conversation.
- Todd Zuercher
-
- Offline
- Platinum Member
-
Less
More
- Posts: 5008
- Thank you received: 1441
07 Aug 2016 19:32 - 07 Aug 2016 19:35 #78452
by Todd Zuercher
Replied by Todd Zuercher on topic Foam XYUV feed rate issue on UV axis
Is the XY movement slower than the UV when the problem occurs?
It might be that the movement of XY are obeying your feed rate and UV is going at what ever speed it has to to keep up? Still sounds like a problem with ABCUVW not being included in the trajectory planning properly.)
If this sounds like the problem you are having, There was a test branch made earlier that might help, but it was an older one from before the JA merger.
github.com/robEllenberg/linuxcnc-mirror/...ure/uvw-blending-dev
I only played around with that test branch a little while, before moving on with another work around for what I was trying to do.
It might be that the movement of XY are obeying your feed rate and UV is going at what ever speed it has to to keep up? Still sounds like a problem with ABCUVW not being included in the trajectory planning properly.)
If this sounds like the problem you are having, There was a test branch made earlier that might help, but it was an older one from before the JA merger.
github.com/robEllenberg/linuxcnc-mirror/...ure/uvw-blending-dev
I only played around with that test branch a little while, before moving on with another work around for what I was trying to do.
Last edit: 07 Aug 2016 19:35 by Todd Zuercher.
Please Log in or Create an account to join the conversation.
07 Aug 2016 19:36 #78453
by delinq
Replied by delinq on topic Foam XYUV feed rate issue on UV axis
Hi
Found this over at the profili2 forum:
devcad.forumfree.it/?t=69476862
Apparently, the same issue is also present in Mach3.
The solutions sounds plausible, going to try to create a mirror config so that I can get my wings cut.
Found this over at the profili2 forum:
devcad.forumfree.it/?t=69476862
Apparently, the same issue is also present in Mach3.
The solutions sounds plausible, going to try to create a mirror config so that I can get my wings cut.
Please Log in or Create an account to join the conversation.
09 Aug 2016 11:55 #78519
by andypugh
Replied by andypugh on topic Foam XYUV feed rate issue on UV axis
Depending on how difficult it is to modify the G-ode source, you could consider using inverse-time-mode feed-rates. This tells the machine how long a move should take, not how fast it is.
linuxcnc.org/docs/2.7/html/gcode/g-code.html#gcode:g93-g94-g95
Fundamentally LinuxCNC does not know how long a path is between a pair of UV coordinates. Their meaning is much less clear-cut than XYZ, which can be assumed to be a cartesian axis set.
In your specific geometry the system is probably only considering the XY path length when calculating the velocity.
If inverse time feed (G93) solves the problem then it ought to be possible to create a simple input filter for your LinuxCNC installation that converts from feed-per-minute to inverse-time. The reason this can work is that the filter knows what U and V mean on _your_ machine, whereas LinuxCNC can only work with what is true (or assumed to be true) on every machine.
linuxcnc.org/docs/2.7/html/gcode/g-code.html#gcode:g93-g94-g95
Fundamentally LinuxCNC does not know how long a path is between a pair of UV coordinates. Their meaning is much less clear-cut than XYZ, which can be assumed to be a cartesian axis set.
In your specific geometry the system is probably only considering the XY path length when calculating the velocity.
If inverse time feed (G93) solves the problem then it ought to be possible to create a simple input filter for your LinuxCNC installation that converts from feed-per-minute to inverse-time. The reason this can work is that the filter knows what U and V mean on _your_ machine, whereas LinuxCNC can only work with what is true (or assumed to be true) on every machine.
Please Log in or Create an account to join the conversation.
12 Aug 2016 18:41 #78684
by delinq
Replied by delinq on topic Foam XYUV feed rate issue on UV axis
Hi
I've tried setting G93 in my gcode, but there where no difference in how it reacted.
Had some sucsess running the mirrored settings, but it still runs too rapid when told to move long distances on the UV axis, and totally ignores Feedrate.
I've yet to try the other build that Todd mentions, but find it odd that no one else is running into this issue.
I've tried setting G93 in my gcode, but there where no difference in how it reacted.
Had some sucsess running the mirrored settings, but it still runs too rapid when told to move long distances on the UV axis, and totally ignores Feedrate.
I've yet to try the other build that Todd mentions, but find it odd that no one else is running into this issue.
Please Log in or Create an account to join the conversation.
- Todd Zuercher
-
- Offline
- Platinum Member
-
Less
More
- Posts: 5008
- Thank you received: 1441
12 Aug 2016 19:46 #78688
by Todd Zuercher
Replied by Todd Zuercher on topic Foam XYUV feed rate issue on UV axis
You can't just add a G93 to your old code. In G93 mode every line of code (G1,G2, or G3) must have an F-word that is the time it takes to complete the movement specified by that line.
That' all I've got. (never used it myself)
That' all I've got. (never used it myself)
Please Log in or Create an account to join the conversation.
12 Aug 2016 21:08 #78691
by delinq
Replied by delinq on topic Foam XYUV feed rate issue on UV axis
Every line has a F set in my test, no change.
Please Log in or Create an account to join the conversation.
15 Aug 2016 12:06 #78794
by andypugh
Can you show us a sample of the code?
Replied by andypugh on topic Foam XYUV feed rate issue on UV axis
Every line has a F set in my test, no change.
Can you show us a sample of the code?
Please Log in or Create an account to join the conversation.
Time to create page: 0.085 seconds