Trouble with dual motor gantry setup

peterdownunder wrote: I have currently switched to gmoccapy mainly for the ability to override limits. If you go past limits on this machine it is hard to get it back (the gantry does not easily freewheel and weighs a ton)

That statement leads me to believe something is not right there:
-Axis has override limits, so do most GUI's of LinuxCNC,
-A machine, big or small should never hit physical limits if it is set up correctly.
Since i do have plenty of industrial retrofitted machines, i am pretty sure they never do, one of them is used daily by a not so "careful" user and it is a big Esab, gantry must weigh a ton.
A better description with some pictures and config files might help, also add a switch, crippling the machine for a limit switch just does not compute.

Thanks for the help. I will upload the current config and get a better desc of what is currently happening.

Currently the homing just errors out when hitting the limit switch.

I will post a desc of the machine in a few hours.

here is a link to some photos photos.app.goo.gl/HMw2qjZ6G6dreh567
You can see that there is an inductive sensor on the gantry that is activated when the gantry is in the normal range of the axis. Holes drilled at each end mean that the sensor is deactivated when it approaches. This is typical for all axis.

In a few photos you can see the physical endstops welded into the rail mountings. Things get a bit serious if you run into those.

The Silverdragon UI does not have the limit override, which is why I used the gmoccapy. This machine is part of a shared makerspace and part of this is to improved the usability for users. I will be looking at other interfaces as well.

Part of this challenge is that the UI gives a good indication of the state of the machine. Previously had mach3 which could be perplexing to say the least. Also have the chinese huan yang vfd which does not seem to work with mach4. Am also happy to modify any interface to help improve it for our specific case. We have quite a few programmers, myself included.

Quick feedback from the forums here indicates the reason we are looking to change, I already have linuxCnc installed on my desktop taig mill at home. I have setup dual microswitches on each axis on mine and the homing appears to do as expected. Will be doing some tests tonight.

So its good you are having more luck now.
I would be a little bit concerned about your sensor triggers. (or un-triggers). Its easy to invert a signal in hal but Linuxcnc expects your home sensor to remain triggered all the way to limit of travel. It requires this to know which way to travel to clear the home sensor if the home switch is enabled at the start of homing. If you are near the max limit and you try homing, the result won't be pretty as it will find you hard stops very quickly ( I know about this fist hand!) Those holes really need to be slots.

You might like to have a look at the sensor triggers I devised for my spaceship plasma table as they would be easy to add to your table. You just need to raise the prox sensor to clear them which might be easier than cutting slots.
See the very first post for details here. forum.linuxcnc.org/show-your-stuff/32029...cutter-build?start=0

A machine that size really is crying out for more I/O than a parallel port offers. You could add a second par port, but the Mesa 7i76e would be ideal. If you live downunder in .AU like me and you want one in a hurry I might be able to help out. You'll find my number on my web site www.vmn.com.au/

And as Tommy said, a machine should never hit a hard limit because the limit switches are inside those limits and the home switch (if not shared) should always be inside the limit switches and remain triggered all the way to the limit switch. I know a lot of people only use a single home switch and rely on soft limits which is OK if you have limited IO. I chose to do it properly because I figured if they were supported, it was for a reason. I will say the hardest axis I set up was my Z axis which used a shared home/limit switch. For such an axis, you must use a home offset to move the home switch off the trigger otherwise your limit switch will fire once homing is completed. I just resented giving up travel to achieve this so would set my next machine up the same way as my Spaceship.

I did check the files last night, or this morning as it was 5:45AM, but was to tired to reply.
In the ini file X axis has 0 velocity for search and latch so that will not work. The files are set for a normal 3 axis machine, not for a tandem axis, so there should not be a need for two limit switchesH.
As for holes, they seem big so should do the job if the search velocity is low, i would set it to 5 and test, then 10 and i would leave it at that.

Thanks for the advice, and will probably go that way. The construction of your system is similar and you are correct, we could easily implement something similar.

One advantage of the current sensor setup is that if something screws up, ether one motor stopping causing severe racking, or something is on the rails causing the X-Axis to lift, it triggers a limit, which is nice.

We have a lot going on at the moment, but I will probably take you up on that Mesa offer. We would like to install on a small Lenovo thinkcentre that will fit in the control cabinet, but has no slots, so an ethernet connection would solve that (or allow us to install elsewhere with easy cabling).

Also have some safety concerns about the system and would like to put as many safety interlocks as possible, given the use by not so experienced users.

We currently limit the X speed as that is an exposed area and is driven by dual Nema 34 at 5.6 amps, so can move at a frightening speed considering the mass of the gantry.

As part of this SW upgrade we are also replacing the spindle, which has suffered too many crashes and main bearing are a bit sad. The new one has an ATC so we need to drive all the pneumatics for that as well.

Currently running off an old PC that has 2 parallel ports, but that is for testing only.

Add to that we are moving premesis in a couple of weeks time. Not looking forward to that.

Fun local fact: this is an Australian CNC that the previous owner used to make surfboard blanks (how very Australian )