The Mach2 config that the machine came with ref'ed y positive and was not zeroing the axes. The guy I bought it from was a wannabe who did not even understand general machining practices so he never noticed.

I am using whatever version of Mach3 ArtSoft had available for download 1 month ago. I use stable code, I do plenty of beta testing in my real job. Haven't checked whether there's a new release available, I will do that now.

Joe

Sounds like the guy may have had it set up correctly if I am understanding. If you home the Y so that the axis is furthest away from the column then when you reference the axis it should show your machine coords as Y+ Max. This is what the Home Off value does.
Hood

I don't think so. Softlimits im Mach2 were 0,9999 and the original owner never did a REF ALL. He bought the machine retrofitted and never messed with any settings. It was set to home Y+, though. I'll run it in an hour or so and report the results.

Okay I changed the homing/limits as follows
y homes negative
ymin limit 0
ymax limit 6.47
REF ALL
and the servo drives the saddle right up to a mechanical stop (past the limit switch) and stalls the servo, and I jump for the estop. Note that this is the first change that I've not tested first time by tripping the limit with my finger and been poised over the estop on the first actual run, and the first one to end up driving the machine to a mechanical stop. @%R^&*!
I check the y- limit switch and its position is fine
I crank y back several inches and REF ALL again, actuating the switch with my finger which it ignores completely, I jump on the estop in time to prevent metal to metal.
Changed the homing/limits back but added home off 6.47
Note that the y- limit WILL trigger the estop and x home. In other words, I REF ALL, let Z and Y ref, when the X is ref'ing, hit the troublsome y- limit switch with my finger and it acts as if it has hit the x limit, let go the y- limit and x stops and zeros. I thought all of the limit switches were wired in parallel? How can it detect an open limit switch for purposes of estop and ignore it for homing... but selectively?
to recap
y- limit switch is ignored when homing y axis but not x axis
y- limit will trigger the estop.

WTF?

BTW I did run my test program once after adding the home offset with no softlimit warning. Logically it does not make sense, but I can see where it could be a bug in Mach3 that only surfaces if the work envelope in y machine coords are all negative.

Joe

You say you have them in parallel,  you sure? normally you have NC switches wired in series.

So are you sorted now? Some of the earlier versions had a bug for the softlimit warning.

Hood

I did not alter the wiring, and I have no schematic for the machine.
I assumed that all the limits would be NC and wired in series, and that opening any of them while an axis is homing would be interpreted by the control as the appropriate axis, after all there's only 1 axis in motion at a time.
I am going to pull the limit switches at some point to have a look at how they are wired.
I have the most current release of Mach3 and have never had an older release. I will run a few more test passes with the program that was giving me the errors. As for the limit switch mystery, I will solve it in the future. Thanks for all your help.

Joe

You seem to have two circuits in play for the limits and home switches and they seem kind of weird in the way the are wired. Reason I know there are two circuits are there are two different pins asigned for all of your home and limit switches in Ports and Pins, but without actually seeing them its hard to envisage the connection.


Hood

Now that I am looking at that, it doesn't make any sense. X home is assigned the same pin as X+ yet X homes negative. Should I just change the Y home to the other pin and try to home negative? I don't see a problem with how I have it set up now, essentially the same effect, but I'd like to understand it better.
Joe

I wouldnt mess to much with asigning the pins until you understand how its wired, it certainly seems unconventional in its pin asignment
If its working for you now I would leave it until you have time to investigate and rewire if needed, if its not working then I would investigate/rewire before I started messing with pin asignments.

Hood

Actually, I tried it already, assigned the same pin to y home that is assigned to y+, changed the y home direction to negative, changed the y home off to 0, and triggered the y- limit switch it by hand inches away from the actual limit to test (other hand poised over the e-stop the whole time). It's perfect now.

I bet they have the pin assignments for the limits backwards, after all every axis now has assigned as the home switch pin the same pin as the limit switch at the opposite end of the axis.

I tested REF ALL, soft limits by rapiding up to the limit of each axis, and ran program that was giving softlimit warnings from the home position, no warnings. What else do I have to worry about? Any of the limit switches will trigger an e-stop, so what other potential problems could I be looking at? I can't think of any.

Can you see any benefit to having 2 separate circuits (and pinouts) for the limit switches? Seems to me that one circuit for all would suffice.

Joe