Calypso CNC User Forums

I am running MSM version 2.0.10 Pro and Mach3 Version R3.043.066.

What is the "correct" way to continue operation following a Feed Hold? When I did it the tool crashed into the part!

This has happened twice now. The first time I did it but definitely had the spindle running when I tried to restart. The second time when my my son and I were debugging the 3D printer attachment we made for the mill and, not knowing the problem, he hit Cycle Start which promptly snapped off the extruder. I mention this to show the spindle state was not the issue.

Hi,
I'm not clear as to what you are seeing happen.

To resume from a feed hold state I just click cycle start.

Have you done something else during feed hold (jogged maybe)?

Dave

I'm not sure at this point but there may have been a jog of the A- axis which, in this case, is driving the stepper of the plastic extruder. Is jogging a problem during feed hold? The "manual" says it is OK to do this.

Also not sure but I think it was a rapid down movement in Z that occurred when Cycle Start was pressed. For what it is worth, this is a different profile for the 3D printer configuration and Master Tool mode, and all auto TLO functions are off. After a Ref All, X and Y are set to the approximate center of the printing bed and Z is manually zeroed by the old fashioned method of a piece of paper as a feeler gauge with the extruder moved down to the bed. I checked and there is no G43 code in the file and TLO is not active when the program starts (G49 listed in the mode line near the bottom of the screen). However, I *don't* know if TLO was "auto-enabled" when Cycle Start was pressed after the Feed Hold but seem to recall reading in your documentation about bugs in Mach3 in this area. Have you ever heard of this behavior before?

I said this happened a few months ago while machining metal. I just remembered I made a note of how to resume after a Feed Hold that I determined by trial and error. I found the paper and it says:
To restart from Feed Hold:
- Press Run from Here
- Press Cycle Start
- Cancel Prep Move
- Press Cycle Start

It has been a long time since I did the trail and error but I don't think I did anything like jogging, etc.

Any ideas?

Hi,
A couple of ideas:

1) I've heard of problems with a jog of A during a feed hold causing a problem with the positioning of the X,Y,X axes.
I don't have details, but other MSM users have said that tales of this can be found in looking thru the mach forums.

2) If you are not using tool offsets, make sure (as insurance) that the tool table has a zero TLO for every tool you may mount.
Mach has some bad habits: It tries to be "helpful"... of you mount a tool that has a non-zero TLO, then mach can sometimes auto-magicaly turn on TLO compensation - even if the gcode did not ask for that to happen. Msot often this shows up when when TLO=on, moutn tool wiht TLO=0, and G49 happens (without a G49 command!). But it has been known to happen the other way too.
See documented and unfixed mach3 bugs 4.23 and 4.27 in the MSM release notes for some examples.

3) I see that you used to use "run from here". That's a "feature" of mach that is known to be fragile.
I'd also note that "run from here" is totally different from feed hold. A feed hold simply stops motion at the next convenient moment (exactly where is dependent on mach's internal buffering operations). Resuming (via cycle start) should just pick up from where mach was held.
ON the other hand, using "run from here" causes mach to rewind the gocde file and "internally execute the file from the start to the "from here" point - and then start motion again. This is VERY different - as all the state of the machine is changed as each gcode line is executed between the start of the program and the "from here" line.

Dave

Thanks, Dave. This is excellent information and I appreciate you taking the time to write it and for ideas on what is the cause (in Mach) of this problem.

Some clarifying information:

1) The "procedure" I listed for restarting from a Feed Hold is only what I found by trial and error after the crash I had a while back. It's not what was used in either of my crashes. As you have said this is NOT a good way to do it and I agree but, at the time, it did seem to work in that one situation and gcode file. Note that I cancelled the Prep Move which I assume tells Mach to not rewind and run thru the gcode file so it "should be" equivalent to simply pressing Cycle Start assuming, of course, no jogging or other changes are made following the Feed Hold.

2) The profile I am using for the 3D Printer configuration has no tools defined and nothing is enabled in the TC Options & Master Tool Mode section of the Tooling screen.

I am going to do some air cutting to try to understand how to use Feed Hold/Cycle Start and what not to do in between. I'll do this for both cases where the A-axis is enabled or not and report back my findings. My concern is that Mach clearly appears to be non-deterministic so what I conclude may not always be the case.