Limit Switch Noise and Homing Issue

Hello All,

I have been struggling with what I think is primarily a limit switch noise and configuration issue and I am hoping to get a second opinion to help me get to a working machine a little faster.  I will preface this by saying I think I have read every limit switch, debounce, and homing config thread on the board and they have helped tremendously but I think I am still missing something.

Basic Machine Config:
This will be a plasma machine when I am done.  It uses steppers for drives and a paraport interface.  I am using this break out board:

There isn't a ton of info floating around on it and I think it might be because it is not a choice board to use.  Here has the most info I have recently found though I thought I found info on it in the forums previously.  I am open to suggestions on better parallel port board that have decent documentation and performance though I would love it if I didn't have to rewire stuff.

Travel are roughly 48 x 48 x -8 XYZ.  I would like to home the machine at 0,0,0.  X and Y home and at min axis travel and move positive, Z homes at max axis travel and moves negative.

There is one combined limit and home switch per axis and they are all on axis hitting stops at the extents of travel so the switches will be home, max and min travel limits.

The issue:

After some fiddling with the config and wiring I had the NC switches operating as they should according to the HAL monitor but the machine would not home.  I was getting joint limit errors when I shouldn't have.  I could hit limit override but the instant drives were turned on the inputs would start switching randomly and cause the home to fail. I have done my best to use shielded cable and only ground at the control end to avoid ground loops but the limit circuits are heavily influenced by the drives.  To combat this I modified the HAL to use debounce on the limit switches. I started with debounce delay of 20 cycles on the servo thread and the machine almost homed itself.  X and Y would home normally but Z would home then return a joint limit error pretty reliably.  I messed with the homing sequence to observe x and y homing ok before z caused an error. I suspected this was still a noise issue and increased debounce delay to 40 but that cause homing to fail on x and y. I haven't had time to try the other direction for debounce delay.  The z homes successfully btw the error seem to come right after the home sequence even though the z axis has come off the limit switch

Does anyone have any more advice for fixing the noise issue in the HAL or do I need to go to hardware fixes?  I was hoping the BOB had opto isolators but according to the page linked above it doesn't.  I was hoping it did so I could increase the voltage on the limit switch circuit to get more separation between on and off.


 
 

What is GFR/FID?

I thought I took all the precautions you listed when wiring this thing but I think I will need to take another look at that.

Look over the circuits for the inputs on your BOB. Is there any buffer or isolation chips between it and the parallel port? I'm going to assume that the input terminals float at some voltage and are pulled down to ground by the limit switch via the ground terminal beside the input terminal. What voltage does the input terminal float at relative to the ground when nothing is attached?

Ok, I am familiar with GFI / GFCI. All of the standard 15 / 20 amp receptacles in my shop are GFI protected. I am also pretty confident the shop is grounded correctly.

If it is simply the pull up is weak or at a marginal voltage level, adding an external pull up could help a lot. (This would be if your false triggers are when your input circuit is open.)

If the false triggers are occurring when the input circuit is closed, then the fix may be more difficult. In this case you are getting voltage noise that is raising your input voltage above the detection level (what ever voltage that is) in spite of it's connection to ground. Here the best solution would be to fix the noise, and or some how raise the voltage threshold for the input signal detection higher than the voltage induced by your noise (by adding a buffer or optical isolation)..