Floating Z Configuration

[Blugreen]

Hi. I have a CNC Plasma that has a AXBB-E and UCCNC. I have proximity sensors (Pin 1 Port 2) on 1 end of each axis for homing (working perfectly). I also have a micro switch (Pin 2 Port 2) on a floating Z.

I use SheetCam to generate the TAP file.

I am a little lost in the next step to get the Z to move down to then touch off on the stock prior to each cut.

ANy help would be amazing. TIA

[spumco]

[EDIT - if you don't know what a 'post processor' is, it's essentially a text document with instructions for Sheetcam. The 'instructions' cause Sheetcam to output a TAP file with certain commands, in a certain format, and certain order.]

This will depend on your machine configuration, but generally goes like this:

1. Set floating Z port/pin as "Probe 1" on general I/O settings tab. This makes that floating Z sensor a 'probe' that responds to G31 probing moves.
2. Get/write a post processor for Sheetcam intended for use with a floating head switch.
3. Measure the switch distance (tip touch to switch trigger distance) and adjust the post processor "switchoffset" (or whatever the variable is called in the PP file).
4. Direct Sheetcam to use that PP when outputting a cut program.

Basically, the PP I'm using causes Sheetcam to output a program file that includes the following functions before starting the actual arc-on/pierce/cut moves:
- G31 probing move in Z-minus (probing distance set very long so it never stops short of the stock. G31 causes the Z to stop when the probe triggers)
- (probe triggers)
- G92 (set Z0 at probe trigger point)
- Retract 'switchoffset' distance
- G92 (set Z0 again. This should now be the top of the stock if you've got your switchoffset value right)
- Retract to pierce height
- Start the main functions (Arc on, wait for arc-ok, pierce delay, down to cut height, THC delay, etc.)

Couple thoughts:
1. G31 stops the probing with decel, so there's a bit of over-run depending on probing (Z) speed. Read the UCCNC manual for G31, and you'll see that variable #5063 will be the actual position when the probe triggered, not including the over-run distance. If you have a fast Z-probe speed this overtravel distance could be considerable and your "switchoffset" value won't be super-accurate.

A better post processor than the simple one I described above would either move Z to the #5063 value before setting G92 Z0, or compare the difference between current position and #5063 and subtract the difference before applying the G92 Z0.

2. Get a spring for your floating Z so that thin material isn't distorted (as much) by the weight of the torch and slide. You can't have it too light, because then fast Z moves during the cut (triggered by the THC) can cause the floating head to float up from inertia lengthening the arc. The THC responds by driving down the Z-axis, and you wind up with the floating head acting like a bobble-head and your Z-axis doesn't control arc length very well.

Not as much of an issue if your Z-axis is slow, but a servo-Z capable of 800ipm and high accel will certainly un-seat the floating head if a preload spring is too strong.

There are other folks significantly more knowledgeable about plasma operations than me who will hopefully be along to give some help, but I think this is enough to get you started.

-Ralph

[Blugreen]

Hi Ralph.

Thanks for your reply and the detail.

I was able to get it working before I saw your comments. I tweaked a post processor and have it working perfectly so am very happy.

I saw you mentioned a spring for the floating Z. I was headed in that direction then stumbled on 12mm Linear Bearings on eBay. I have one of those as the floating Z with a hard stop on the bottom. The plate that slides on it is 6mm steel so with the plasma torch it has a reasonable amount of weight with almost zero friction allowing it to always seat on the mechanical stop after probing.

[edit] Just re read about the spring. I will check that and make sure that the floating Z is not causing the stock to distort on probe. Thanks

The only other thing I am keen to have working is the reset triggering at the end of the program to take the load off the steppers.

Thanks.

Mark