Messages

51

General Mach Discussion / Re: Bugs in Mach 3.42.015

« on: October 31, 2008, 02:25:00 PM »

I'm in millimeters. I have 800 steps/mm. This is the resolution of my lead screw/motor combination. I use 0.9° steppers with 10microsteps with a 5tpi ballscrew.
My parts come out to the correct size and I can step the machine in about 2micron increments so all works.

What does  that have to do with me clicking the "zero" button  and the DRO not going to zero?

No comments on bugs # 2 and 3?

Brian

52

General Mach Discussion / Bugs in Mach 3.42.015

« on: October 31, 2008, 11:42:39 AM »

1. DRO's don't always zero they will have 0.0013 left in them sometimes.

2. Step and Dir pulse lengths on some installs will reduce upon opening and closing the motor tuning window.

       ie.  Step=2  Dir=1    after opening, setting, closing , opening motor tuning will change each by -1   

3. Program extrema is not being calculated correctly.

53

General Mach Discussion / Re: CV backwards (G61/64)

« on: October 30, 2008, 05:17:04 PM »

So with further testing and thinking I had this licked I am seeing the real problem.

When raster cutting along XZ-axis while indexing the part with a rotary table A-axis I am having CV issues. The issue is that the velocity is not continuous as if you set it into exact stop mode or turned on CV dist tol and made the tolerance really really small.

If I open up a new session of mach and load the code and run it it works great. If I make changes to CV dist tol value, turn CV dist tol on/off, feedhold, stop, start, restart or some combination of those commands is when it starts having issues. It will act as if its in exact stop mode making it take forever to machine my contours.

The trouble is I don't know what the combination of commands is to repeat the bug. I just know that it does happen enough to cause me headaches in ver 3.41 and 3.42.015. I will be trying to figure out exactly how to reproduce the error repeatedly. Does anyone have any inclination to what the issue may be or have they seen this before. Attached is code that it happens on.

Brian

54

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 12, 2008, 02:30:02 PM »

So far I cut 3 jobs and it seems to be performing without putting any large radius on my parts sharp corners that I'm contouring on and my feeds are
in the same 400mm/min-1000mm/min range. So at a quick glance yes, I really appreciate your responsiveness on this matter. I will let you know if I have any more troubles.
 

55

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 12, 2008, 02:15:06 PM »

Basically there have been some changes made to CV and my settings no longer ran my machine like they used to.

I used to be able to use really small CV dist tolerance values and the machine ran fast. My values now have to be much larger
to keep the speed up.

56

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 12, 2008, 11:29:46 AM »

Here is a video of my problem.

http://www.jbdstudio.com/cv_vid.wmv

Here is my xml

http://www.jbdstudio.com/xml.zip

The problem is CV dist tolerance. Wha I have learned is something has changed in your trajectory code where targeting a small radius the accel decel planning is not nearly as fast as it used to be. I used to always run with the dist tol on set to a small number and get really nice crisp parts. I ran parts last night with it off and they seem to fine and its running smooth. I bumped my radius up to .5mm. I Used to get the same accel decel at .1 that I now only get at >.5.

So the real issue is something changed with CV dist tol, but I seem to be doing ok without it now. Now I'm not sure if there is a bug or maybe now CV distance tolerance is working properly and maybe before it never did.

Brian

57

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 11, 2008, 08:17:04 PM »

Somebody needs to run this on their machine to believe me. If you can't duplicate this I will make a video and post it to prove this. I understand how CV works.  I have two of my units and one other of my friend's units doing this and they use completely different hardware.

Basically if you try and 3d contour with exact stop the machine is going to sound jerky like a machine gun stopping at every little point along the way. In CV mode it runs smooth through the points cause it calculates the trajectory accel and decel based on the lookahead. Then we can control how continuous through a target radius. I get it. I've been using it and watched it get better since mach2. Then on one of my upgrades, I'm not sure which one its backwards. It's working backwards so when its off its on, when its on its off.

Maybe I have a bad install or there is some other cause of this, corrupt xml or something but I don't believe so. CV Feedrate is off, CV limit angle is on and set to 179, CV dist tolerance is on and set to anywhere between 1mm and 0.05mm depending on the scale and geometry of my part.

We'll see what Hood has to say in the morning. Thanks for your attention on this issue.

Brian

58

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 11, 2008, 06:44:06 PM »

Great. What I supplied you with was code for a 4 axis mill. You need at least 2 axis simultaneous motion to perceive the problem. I can generate code for that if u like. I don;t know about the lathe but it should be the same. Its really noticeable on 3d contours.

Brian

59

General Mach Discussion / Program extrema failing to read file properly

« on: September 11, 2008, 05:39:39 PM »

in ver 008 the program extrema are reading wrong. This codes extrema are read fine in version 3.041 but in 42.008 they are way off. Please correct this.

60

General Mach Discussion / Re: CV backwards (G61/64)

« on: September 11, 2008, 05:35:45 PM »

No. The machine states the correct modes but I can tell by the way the machine is pulsing that its in exact stop mode. Take the Gcode attached and run it with CV on then run again with CV off. You will see that mach is stuttering in CV mode (its stopping at each point) then turn CV off and you will see the pulsing is smooth and fast ummm... continuous.
So its like some boolean value in the mach code is backwards.

Turn light switch on and the light goes off turn it off and the light goes on kinda deal. But the leds and displayed modes are correct. Its the performance thats backwards.