2.8pre gantry racking

Try setting HOME_IS_SHARED

It is possible that the homing moves are starting before the other axes are off of the (shared) switch.

Yes the halmeter does show the pins changing state. All are FALSE and turn to TRUE when I trip them one by one.
I've also tried setting HOME_IS_SHARED both ways, 0 and 1 to no avail.

Z completes, the Y completes, both are at HOME which is not touching the micro switch. The gantry (X) will then head towards the switches, X0 will trip and back off, X1 will then head in, trip the switch and keep going as though it never found the switch.

My next move is to move the shared switches as they are on pins 11/12 of parallel port 0 to 11/12 of parallel port 1. I want to proof to myself that the wiring and ini/hal files are ok, maybe the G540 in the way somehow.

Rick

Can I suggest using Halscope to watch the joint.N.home-sw-in pins, just to make completely sure that they are tripping during the actual sequence?

Yep, will do.

Rick

Well, the saga is over.

I moved the home switches to a second parallel port card so I could have enough pins for separate home switches instead of paralleling them. Same issue, the X axis would set one side and rack the other. Grrr. Halscope does show the joint.N.home-sw-in tripping as expected.

Since it made a difference before, I swapped the pins for the X joints in the hal file. Success !!! It now homes ZYXX as I want, and works repeatedly.

Why the pin swap makes a difference I don't know. Different parallel port, cable, BOB, individual wiring to the micro switches, all gave the same problem. After all this and finally getting a working system, I got an update notice for 2.8pre. May swap the pins back to see if the update does anything when I have some time.

Off to make parts again.

Thanks all,
Rick

llamatrails wrote: Why the pin swap makes a difference I don't know. Different parallel port, cable, BOB, individual wiring to the micro switches, all gave the same problem. After all this and finally getting a working system, I got an update notice for 2.8pre. May swap the pins back to see if the update does anything when I have some time.

I wouldn't expect the update to have any effect. I am pretty sure nothing much has changed there.

homing.c hasn't changed since february last year: github.com/LinuxCNC/linuxcnc/commits/mas.../emc/motion/homing.c


Can you show us the before and after versions of the code, perhaps it is in some way explicable?