Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: kmacht on February 16, 2011, 02:18:06 PM
-
I posted this on the yahoo group but figured I would try here as well.
I need some help. I built a JGRO wood router and am playing around with it
using Mach3 Mill. I am having a problem where the stepper motors are having a
slight hesitation between each line of g-code. It does not matter if it is a
circle or a straight line. If I write code for two straight lines connected to
each other I get a slight pause inbetween each line when I run the code. I have
tried it both with CV on and with exact stop on. It didn't make a difference.
I have backlash set to zero so I know it isn't a compensation issue. It is
almost like the computer takes a break from sending the pulses to the parallel
port each time the screen updates to the next line of code. The final dimensons
come out ok so I know it isn't actually loosing steps, it just makes the machine
run very jerky. I don't know if it matters but here are the details of my
setup.
Probotix bi-polar drivers with the non-isolated breakout board
256 oz/in powermax II bi-polar motors
1/2-10 acme threaded rod with dumpster anti-backlash nuts and lovejoy's
motor tuning 2000, 70ipm, accel 4, pulse width 2 - no microstepping
The program was drawn in autocad and cut2d was used as the cam program
Cutting at 35ipm and .200 depth per cut in mdf
Everything seems find with the setup except for the pauses between lines of
code. The machine jogs, rapids, and moves using MDI with no problems.
Any suggestions?
Keith
-
You say you have backlash set to zero, does that mean you have it enabled? If yes then disable and see if it helps.
Hood
-
No. As far as I know, backlash is turned off. I checked backlash with a dial indicator and it was under .001 so I didn't bother to turn it on in Mach3. Even if it was turned on, it shouldn't make a difference with two straight lines in a row. If I write a program that such as:
G01 X0 Y5 Z1
G01 X0 Y10 Z1
I wll get a slight pause when it goes from the 1st line to the 2nd one.
Keith
-
Can you attach your xml so I can see if I get the same issues.
Hood
-
What is an xml? Do you mean the g-code?
Keith
-
xml is your config file, its located in the main Mach3 folder and will have the name of the profile you are using, so for example if using the standard Mill profile it will be called Mach3Mill.xml.
You will need to rename it for the forum to accept so copy it to your desktop and rename something like kmacht.xml and then use the additional options button on the reply page to browse and attach.
Hood
-
Here it is. I tried a few other settings quickly. The amount of pause is directly proportional to the accel rate. If I up the accel rate you can still hear the quick hesitation but it is alot harder to actaully feel.
Keith
-
Keith,
Try 7 for accel value for X & Y, and 3 for the Z.
RICH
-
Make sure all CV options are turned off, but CV is on. Uncheck all the CV options in General Config, and turn off CV Distance and Feedrate on the settings page.
-
Kmacht, I want to piggy back on your post if you don't mind.
I have a similar problem. It manifests itself in the following way:
When cutting a circle, Mach generates two semi circles split across the y axis. Each semicircle is it's own line of code.
When the code moves from line to line their seems to be a pause, which creates a flat spot in the circle. I have just run this with a pen, rather than a router to make sure I am not having backlash/load problems. Flat spot and line change coincide.
I have some very slight dimension problems which could be explained by this issue as well.
Any thoughts?
Thanks,
Dan
-
Dan,
Mach is computing X & Y velocities / change in Y with respect to x and as you aproach 0-90-180-360 one axis velocity is realy slowed down
to maintain position. So there seems like a pause and in CV mode it is transparent. Since there is a change in time at those points the tool
dwells there slightly longer and with tool runout you will get marking. Depending on axis resolution, accuracy of the axis, and spindle run out you won't get any flat spots.
RICH
-
Also make sure there is not a G61 in the header of your gcode file. Regardless of how you have CV set in Mach, G61 will turn it off. You should have a G64 in the file header, which turns CV on.
-
So if I'm understanding, get settings to CV mode, without additional options checked as mentioned in a previous post, and/or add G64 to the initial string/header and I should be good?
I'll try it.
Tested this with a pen to eliminate tool deflection, and I removed the motors, to eliminate the possibility of backlash, and observed the behavior of the shaft: only where the line change in code occurs do I get the dwell/pause.
-
I should also say, THANKS for the feedback!!!
Appreciate it!
-
I've updated my CV options as described and I still have this problem. I've pulled my motors and observed the pause at the change in code between the semicircles. Any help much appreciated.
Thanks,
Dan
-
Can you attach your xml please.
Hood
-
I don't have it with me but will attach it this evening. Thanks for your response!
Dan
-
Hi Hood,
I've attached my XML. Let me know if there is something I should change.
I am running a 'home brewed' machine (originally brewed by someone else...), with Gecko drives and servo motors.
I appreciate you taking a look.
I have had a strange issue with my General Configurations. Sometimes when I load G Code, my Distance Mode will flip to Incremental from Absolute, where I typically have it. Then I will get an "Abnormal Condition" flashing button and an error at the end of a semicircle line of code, saying the next location cannot be determined (don't have the exact verbage). Any thoughts on that appreciated.
Best regards and thank you,
Dan
-
I dont see any problems in simulation here with the xml and hesitation, can you attach some code that you have the issue with?
Regerds the distance mode, it is likely your code that is changing the IJ mode. Because the mode has changed and not been set back the next code that is loaded without a G90.1 or G91.1 in it may be wrong. It is good practice for your code to have the correct IJmode in it, that way no matter whether you use Abs or Inc as default your code will change it to the mode it needs.
Hood
-
Hood here is some code.
G17 G20 G40 G54 G64 G91
...
N660 G00Z1.0000
N670 X27.1768Y4.8232
N680 G01Z0.3700 F15.0
N690 G17
N700 G03X18.8232Y13.1768I-4.1768J4.1768 F90.0
N710 X27.1768Y4.8232I4.1768J-4.1768
N720 G00Z1.0000
N730 X31.2603Y8.9860
N740 G01Z0.3700 F15.0
N750 X31.2611Y8.9720 F90.0
The issue occurs between N700 and N710.
Thanks for the info on the XML. I'll add the IJ code to my header. (Isn't that the G91?)
Best,
Dan
-
Ok tried the code with your xml and I dont see any hesitation in the toolpath view or by watching the units per min DRO. Can you run the driver test and see what that is like.
The IJ mode is set with G90.1 or G91.1, the code you posted is Inc IJ mode so G91.1 would be inthe header.
Hood
-
Hi Hood,
Thanks for checking that out. What's the driver test? Seems like I'm discovering new Mach features all the time...
Thanks for the G Code tip too. I'll change my header in my post and config.
Dan