Hello Guest it is May 22, 2019, 05:30:57 PM

Author Topic: "go to zero" doesn't go to zero and other issues.  (Read 4045 times)

0 Members and 1 Guest are viewing this topic.

"go to zero" doesn't go to zero and other issues.
« on: July 06, 2012, 07:30:25 PM »
Hello all,
   I'm using mach 3 version 3.043.022, and am getting some abnormal results.  I just ran a program, then hit the "go to zero" button, y and x went to zero, but z went from 0.125" to .0017".  This is what mach 3 displayed on it's DRO.  This is a constantly recuring problem on the machine, sometimes if the axis is within .015" of zero, it won't move at all when i hit the "go to zero" buttom.  This happens for all the axis on the machine.

Another issue i've noticed which may be related:
load & run "vice home",zero all axis, then .001 incremental jog 10 times(all axis do this):
    1) .0000
    2) .0000
    3) .0024
    4) .0024
    5) .0049
    6) .0049
    7) .0049
    8) .0073
    9) .0073
   10) .0097

load & run "cribbage board drill 1",zero z (all axis do this), then .001 incremental jog 10 times:
    1) .0010
    2) .0020
    3) .0030
    4) .0040
    5) .0049
    6) .0060
    7) .0070
    8) .0080
    9) .0090
   10) .0099
this is repeatable, i can load either program, run it, and then incremental jog with the same results
the numbers above are the numbers shown in mach.  Actual table movement is ~ the same as measured with a dial indicator.

The theoretical motor/table resolution is .0000492".

Any ideas on what might be the issue?

Any and all suggestions are much appreciated.

Brian

Offline Hood

*
  •  25,811 25,811
  • Carnoustie, Scotland
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #1 on: July 07, 2012, 04:09:46 AM »
If your steps per unit are sufficient to allow such movements then I am not sure what the problem could be. If you attach your xml I will have a look and see if I can figure it out.
Hood

Offline RICH

*
  • *
  •  7,331 7,331
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #2 on: July 07, 2012, 08:21:47 AM »
Just because your calculated resolution is .0000492" that dosen't mean your sysytem is capable of moving  .00005" in a single jogged step. You can have some backlash or non movement because of components or the some of the components can be non linear. A screw can be non linear along it's length. A stepper rotor may move more or less for a single small step.
Pin pointing the cause below .001"   movements can be difficult since a number of components can contibute  and easily add up to a non movement of .0001 to .0005". Even the indicator can be off.

So the above could account for .0001" in the dro in your second example. First example could be due to settings.
FWIW,
RICH
 
 
Re: "go to zero" doesn't go to zero and other issues.
« Reply #3 on: July 07, 2012, 12:01:37 PM »
Hood, attached is my XML file for the machine.  I certainly do appreciatte you taking a look at it.  The steps per are 20320.  I'm using directly coupled servo motors w/ 1000 line encoders driving 5.08TPI screws.  I've also attached the G code files mentioned in my first post.

Rich, i should've clarified that my jogging results from the "cribbage board drill" program are perfectly acceptable to me.  I just posted them to show the contrast between the "vice home" program.  A lot of the things you mentioned pertain to the mechanical nature of the machine, which is all very true, but i'd like to focus more on why mach 3 isn't telling the machine to go to the desired position.  Where i get truly baffled is that i can switch back and forth between each of these programs and get the same results as mentioned in my previous post.  This leads me to believe it's not a settings problem, but i'm all ears for any possible solutions.

Thanks again for the replies!

Brian

Offline Hood

*
  •  25,811 25,811
  • Carnoustie, Scotland
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #4 on: July 07, 2012, 05:11:19 PM »
You have the motors tuned to fast for the kernel set so I am presuming you are using an external controller? If not then set the steps per unit or vel lower.
Hood
Re: "go to zero" doesn't go to zero and other issues.
« Reply #5 on: July 07, 2012, 06:02:48 PM »
That's correct, Hood.  I'm using a G100.  The computer has a 1.5Mhz cpu, motors have 1000 line encoders, so i figure ~190Khz per axis, which should be within the computers range, unless i'm over simplifying things here.  Anything else seem off to you?

let me add to the list of issues.  This is actually the most problematic of all the issues i've been experiencing.  The machine constantly faults out on the z axis during "G0" moves, or when going to the retract plane in a drill cycle.  If i replace "G0 Z.5" with "G1 Z.5 F300" (300ipm is the max speed in my settings), then the program will work.  I can't do this in drill cycles unfortunately.

I've adjusted the max speeds and accelerations numurous times going as low as 1/8 of their current values.  This doesn't seem to have any effect on a G0 Z.5 command, it'll still fault.

I used to have a stepper motor on the z axis, and had the same results.  when the stepper faulted out, i could hear that the pulse rate was considerably higher than when it would work.

I'm 99.99% sure this is not a mechanical issue.  Some attempts to fix this include:
1. installed 2 gas springs on the z axis - the z axis would go up when faulting because the springs were stronger than the weight of the head.  This saved bits, but I still couldn't get a program to run without faulting
2. replaced stepper motor with servo motor 1- no change in z axis faulting out "G0 z.xx" moves
3. replaced servo motor 1 with servo motor 2 - no change in z axis faulting out on "G0 z.xx" moves
4. I've tried a total of 4 different drives - same results
5. uninstalled mach3 and reinstalled - no change
6. spent numurous hours playing with settings in mach 3 and g100 config - still can't get rid of the issue.

I'm truly stumped.  Any and all suggestions are much appreciated!

Brian

Offline Hood

*
  •  25,811 25,811
  • Carnoustie, Scotland
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #6 on: July 07, 2012, 06:13:22 PM »
Not sure how the computers CPU would have a bearing on the KHz of the axis but then again I have never used a G100.
The G100's never really worked that well with Mach from what I hear, lots of problems with them and your issues are likely to do with that. You may get some help if you post on the G100 part of the forum but I wouldnt hold out too much hope as I only know one other person who still uses one and he will be swapping it out for something else when he gets a chance.

Sorry I cant be much help :(
Hood

Re: "go to zero" doesn't go to zero and other issues.
« Reply #7 on: July 07, 2012, 06:28:15 PM »
Thanks for looking into this anyway, Hood.  Maybe it's time to upgrade the G100 to something else :( guess it's about all i have left to change out at the moment.

thanks again,

Brian

Offline Hood

*
  •  25,811 25,811
  • Carnoustie, Scotland
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #8 on: July 07, 2012, 06:34:59 PM »
The G100 held a lot of promise and I almost bought one for the lathe but luckily for me the Smoothtepper was in the pipeline at the time so I held off.
There are a lot of controllers out now that work well with Mach, I have only tried the Smoothstepper and Ethernet Smoothstepper and they both work well for the most part, few niggles with the USB SS and one for the ESS for me. Sad thing is they are just 5v I/O and really I like 24v I/O so I need to convert the voltages. There are however some Polish controllers fairly new to the market and they look promising. 24V i/o Analogue I/O Differential Step/Dir etc. I will be buying one when they get the encoder module done so hopefully they are as good as they look on paper.
Hood

Offline RICH

*
  • *
  •  7,331 7,331
    • View Profile
Re: "go to zero" doesn't go to zero and other issues.
« Reply #9 on: July 08, 2012, 06:03:06 AM »
Brian,
No experience with the G100 so can't  be of much help to you.
Hope you get it sorted out.

RICH