Machsupport Forum
Third party software and hardware support forums. => SmoothStepper USB => Topic started by: Overloaded on March 06, 2014, 11:16:18 PM
-
Just wondering, anyone here know if the encoder spindle lathe threading thing with the SS is at all possible/probable ?
Any progress at all ?
Not much going on on the SS forum. (about a year ::) )
And what we might expect in the quick pull-out ? Unlike the CS Lab and (EMC)
Just curious. :)
Thanks,
Russ
-
Can you explain what you mean by 'quick pull out'?
sam
-
no Z axis pause at the end of the thread before the X retracts the tool from the thread, preferably completed within 1/2 to 1 revolution of the part while maintaining the pitch with the Z axis.
Mach does this quite well, some others ... not so good.
Russ
-
Russ,
Don't have an answer for you but do have dumb question.
From a "fastener use" viewpoint....... Why does it matter if the pitch is maintained or is gradual?
The stress point is moved and intensification factor would still be subjective, and in calc's, the max loading would still be based on the thread geometry.
Even the max loading is subjective.
Just wondering if you are needing to address someones spec for the bolt........
RICH
-
Hi Rich,
I'd say it doesn't really matter that the pitch is not maintained "exactly", actually with one routine here at 1000 rpm Mach pulls out in 1 revolution while the 48 tpi pitch diminishes roughly 30-50% (by eye) which is pretty good and eliminates cutting the annular groove.
Good point, thanks,
Russ
:)
-
Linuxcnc has always done this...
https://www.youtube.com/watch?v=Z_LxyosF2yc
http://www.linuxcnc.org/docs/2.5/html/gcode/gcode.html#sec:G76-Threading-Canned
You can specify an entry or exit taper - this move is synchronized with the spindle.
-
Thanks Sam, I might look into that.
Other emc vids I've seen show what looks like a groove, maybe I didn't see correctly.
Thanks again,
Russ
-
Russ,
dont know if its been done or not but suspect not. However even if it has then I would say its unlikely to work as it seems using full encoder sync with Mach3 will always produce that delay in pull out. I talked with Steve (Smurph) about this issue and he said he had exactly the same issue with the Galil so had to revert to single index. Rufi of Vital systems said to me it works fine in DSPMC/Hicon but neither he or anyone else could show me video evidence of it so......
Sam,
does LinuxCNC still do tapered threads wrong or has that been sorted?
Hood
-
Hi Hood, thanks.
Watching a thread on encoders running now at the W9 forum. Looks like it may get worked on, will be interesting to see how it transpires.
Thanks again for sharing your research,
Russ
-
Well - it isn't really that it does it wrong as it doesn't do it like any other control and no one has made a canned cycle to fix it yet.. (that I know of anyway)
G33 spindle synchronized motion in linuxcnc was made to be very flexible.. You are not limited to z axis moves. ie - you could cut a scroll if you wanted to. or a fusee. That though required that the pitch be calculated along the move. So if you move x and y the same amount (45deg angle) the pitch is along the hypotenuse. If you are cutting a taper - you have to specify the pitch along the xz move - not the z move. (clear as mud?) People have made some gcode subroutines to do the calculations.. Here is a good thread (no pun intended).
http://www.linuxcnc.org/index.php/english/forum/40-subroutines-and-ngcgui/13006-ntp?limitstart=0
Here is G33 from the manual
http://www.linuxcnc.org/docs/devel/html/gcode/gcode.html#sec:G33-Spindle-Sync
sam
Sam,
does LinuxCNC still do tapered threads wrong or has that been sorted?
Hood
[/quote]
-
Hi Hood, thanks.
Watching a thread on encoders running now at the W9 forum. Looks like it may get worked on, will be interesting to see how it transpires.
Thanks again for sharing your research,
Russ
Be sure to update please Russ, will be very interested specially if it can be proven to work correctly.
I was in hopes that the claims that the DSPMC/Hicon could be proved as it would have provided me with a large boot to kick some arses ;)
Hood
-
Well - it isn't really that it does it wrong as it doesn't do it like any other control and no one has made a canned cycle to fix it yet.. (that I know of anyway)
G33 spindle synchronized motion in linuxcnc was made to be very flexible.. You are not limited to z axis moves. ie - you could cut a scroll if you wanted to. or a fusee. That though required that the pitch be calculated along the move. So if you move x and y the same amount (45deg angle) the pitch is along the hypotenuse. If you are cutting a taper - you have to specify the pitch along the xz move - not the z move. (clear as mud?) People have made some gcode subroutines to do the calculations.. Here is a good thread (no pun intended).
http://www.linuxcnc.org/index.php/english/forum/40-subroutines-and-ngcgui/13006-ntp?limitstart=0
Here is G33 from the manual
http://www.linuxcnc.org/docs/devel/html/gcode/gcode.html#sec:G33-Spindle-Sync
sam
Ha ha Sam, yes I cut fusee and scrolls every day of my life, really handy, who the hell needs pipe threads anyway ;D
Better not ask you about all the other issues Steve Blackmore has mentioned that really pi$$ him off about LinuxCNC as you will have to get your spin doctoring skills turbocharged.
Just having a bit of fun Sam :)
Hood
-
Oh that is ok :) .. - steve has been more than vocal about problems he has with linuxcnc. the ones I know about are..
1. tapered threads (you have to calculate the pitch along the hypotenuse) - Not a show stopper as you can cut tapered threads either with a little right angle trig or gcode subroutine.
2. Jog while paused - work has been done on this maybe in the next release (not been keeping up on it)
3. 1 segment look ahead limitation in the trajectory planner - This is what I have been helping test and am really excited about. A developer had taken up the challenge and has made a N lookahead trajectory planner that is in the testing stages right now. we have used steves gcode to test that the new tp.
It is what it is.
sam
Well - it isn't really that it does it wrong as it doesn't do it like any other control and no one has made a canned cycle to fix it yet.. (that I know of anyway)
G33 spindle synchronized motion in linuxcnc was made to be very flexible.. You are not limited to z axis moves. ie - you could cut a scroll if you wanted to. or a fusee. That though required that the pitch be calculated along the move. So if you move x and y the same amount (45deg angle) the pitch is along the hypotenuse. If you are cutting a taper - you have to specify the pitch along the xz move - not the z move. (clear as mud?) People have made some gcode subroutines to do the calculations.. Here is a good thread (no pun intended).
http://www.linuxcnc.org/index.php/english/forum/40-subroutines-and-ngcgui/13006-ntp?limitstart=0
Here is G33 from the manual
http://www.linuxcnc.org/docs/devel/html/gcode/gcode.html#sec:G33-Spindle-Sync
sam
Ha ha Sam, yes I cut fusee and scrolls every day of my life, really handy, who the hell needs pipe threads anyway ;D
Better not ask you about all the other issues Steve Blackmore has mentioned that really pi$$ him off about LinuxCNC as you will have to get your spin doctoring skills turbocharged.
Just having a bit of fun Sam :)
Hood
-
It is what it is.
Sure is Sam and it is the reason I have the utmost respect for you, yes you tout LinuxCNC as being the dogs danglers but unlike all the others you are willing to admit that its not perfect, the rest dont like to admit that and I think that is why it takes many years for even fundamental things to change.
Mach3 is not perfect either, far from it in some areas, and some things just cant be done with it hence the fabled Mach4, lets just hope we both live long enough to see things improve in both camps ;D
Hood
-
Russ, got a mail from Wotjek today saying they will at some point be doing a M76 for Mach3, so suppose now its just a wait to see which comes first, SS managing to use encoders without the delay, M76 macro from CS-Lab, Mach4 Turn or proof the DSPMC/Hicon can. I know which one I would bet on being last ;D
Hood
-
;D
Thanks Hood,
Russ
:)