Hello Guest it is July 14, 2020, 08:06:11 PM

Author Topic: Mach did again!  (Read 21446 times)

0 Members and 1 Guest are viewing this topic.

Re: Mach did again!
« Reply #60 on: October 06, 2010, 08:58:13 AM »
Dennis
What you experienced sounds different then what I have been seeing, However I don't trust the run from here feature because in my experience it does not always load all the position information so I only use it at a tool change so I know all values for position are called and that seems to work fine
Re: Mach did again!
« Reply #61 on: October 06, 2010, 11:46:15 AM »
Ergerber
Well i agree Mach has some quirks but it works well what version of Mach are you running ? I am always skeptical of updating.

Dennis
Re: Mach did again!
« Reply #62 on: October 06, 2010, 01:15:58 PM »
Dennis
I am running R3.042.040 of Mach and Version 3 DSPMC plug-in. I have been making headway with some of the issues and the appear to be G code related so currently we are breaking down some of the complex lines to see if this helps, I am testing a program currently the drills and taps three different size holes at 96 different locations using 7 different tools, so we will see.
Re: Mach did again!
« Reply #63 on: October 07, 2010, 12:16:58 PM »
Egerber
So you have found the problem it was all code related have you found anything weird in the R3.043.40 version seams that when i update my mill version thing's go ok but when i update the lathe version it always screws up.

Dennis
Re: Mach did again!
« Reply #64 on: October 07, 2010, 01:44:06 PM »
Dennis

I found that the problem was directly related to the tool change I changed the config to Ignore Tool Change from Stop Spindal wait for cycle start and added  befor the next tool M0 and every thing is fine now
I also found that if you don't make tool changes ther is no problem
There is another thread under plugins DSPMC that has moor info
Re: Mach did again!
« Reply #65 on: October 07, 2010, 10:56:47 PM »
Hi Guys, I have changed to a different machine program and none of the issues that I was having with the origin changing or whatever we want to call it are happening. I would like to thank everyone who has helped, especially Hood, Steve at PMDX and David Cole of Cole Controls, but I still believe that there is a bug somewhere that is creeping up.
« Last Edit: October 07, 2010, 10:58:54 PM by ranchak »
Re: Mach did again!
« Reply #66 on: October 08, 2010, 12:38:11 PM »
Ergerber
 Good to hear you have the problem solved i use a tool change but not as many as you have noted so far anyway and found no problem to date in the mill application lathe is another issue.

Dennis
Re: Mach did again!
« Reply #67 on: October 10, 2010, 06:03:44 PM »
Good afternoon,   ;D

as promised, I spent the weekend in the lab.

1.
As I stated, I am quite sure that the M6End Script is sometimes not started, or is interrupted by another "Ghost" sending commands to the trajectory planner. Maybe, Mach3 does not see wether the M6End Script is terminated before continueing with other (paused) operations.
To be sure that my tool-change script is able to finish properly, I attached the content of M6End Script to M6Start Script, generated a Message window to have a pause for the manual tool change. As a result, the M6End is no longer needed.
--> After this change, I could do tests for hours, without encountering any surprising events.

2.
Secondly, I noticed a clearly reproduceable position error after completing the tool change: Z-Axis was wrong.
So to say: Machine DRO and Work DRO are showing correct coordinates, but, the "Z target position" was incorrect.
"z = GetToolChangeStart(2)" delivers the WORK coordinates before start of the toolchange. Unfortunately, this value does NOT include the tool lenght offset. So, when commanding the return movement with a "straight feed" or "straight traverse" command, the Z position is wrong, because those commands need the "controlled point" as target coordinate, which includes the tool lenght offset.
Of course, it has to be taken into account that you need the tool lenght offset BEFORE performing the tool change and new tool measurement... So, you have to do it all in one script, or you have to use a global variable to store the "old" tool offset.


So, I am fine now, waiting for new adventures... I will inform you about the next crashed drill bit.  >:D


Regards, Konrad

Offline rcaffin

*
  •  1,029 1,029
    • View Profile
Re: Mach did again!
« Reply #68 on: October 10, 2010, 06:13:50 PM »
--> After this change, I could do tests for hours, without encountering any surprising events.
Valuable information.

Quote
"z = GetToolChangeStart(2)" delivers the WORK coordinates before start of the toolchange. Unfortunately, this value does NOT include the tool lenght offset.
One could argue about whether that means the 'Z coord' is wrong or just misunderstood, but clearly one has to be very careful about what ASSUMPTIONS about definitions one makes in writing the code.
Once again, very valuable information.

Quote
So, you have to do it all in one script, or you have to use a global variable to store the "old" tool offset.
Good point. Fortunately there are a lot of spare global variables available.

All very helpful: thank you.

Cheers
Re: Mach did again!
« Reply #69 on: October 10, 2010, 06:22:31 PM »
Quote
"z = GetToolChangeStart(2)" delivers the WORK coordinates before start of the toolchange. Unfortunately, this value does NOT include the tool lenght offset.
One could argue about whether that means the 'Z coord' is wrong or just misunderstood, but clearly one has to be very careful about what ASSUMPTIONS about definitions one makes in writing the code.
Once again, very valuable information.
[...]
All very helpful: thank you.

Cheers


Thanks for your feedback :-)
I better should have said: "wrong for what I wanted to do with it"  ;-)

My reference was the default script which comes together with Mach3. I just took apart and reassembled it. So, my conclusion was that I just could return to the coordinates delivered by "GetToolChangeStart" function...  At least my conclusion was wrong...  ;-)

Konrad