Radial Digitizing only one Z depth?

[adatesman]

Got a probing question (pun intended):

Playing with Radial Digitizing and a 2.2190” ring gauge as a way to get a feel for things in CNC12, and not seeing a way to simply do a single pass at a single Z level.

If I set either Z Patch Depth or Z Step to 0, it tells me Invalid Value. Same for 0.0, +0.0, and -0.0. Best I’ve managed is to set them equal (eg 0.010”) and it’ll do two passes, one at Z0 and one at Z-0.010.

I assume I’m missing something obvious, but hours in and I’m just not seeing it. My need is to map the profile of a straight sided part, so only one Z depth needed. Spent the morning going over the CNC12 manual and just upgraded to v4.20.1.

Thx.

[cncsnw]

I think you are correct that it is not allowed, even though there is really no reason to prohibit it. Perhaps that restriction can be lifted in a future software version.

In the mean time, try using the Contour or Wall Following patterns instead. Contour will definitely do just one pass, but it may or may not work on the inside of a part (it was designed for outside contours). Wall Following will work inside or outside, and will probably allow just one pass.

[cnckeith]

wall following or contour might do what you want, i would have to see the part. you can use radial and just let it complete the 1st pass and stop it when it starts the second pass..even if you hit estop the data is still collected.

[adatesman]

Keith just responded on the FB Group, and seems Wall Following is the intended method. Had not tried it, as Radial seemed the less complex method and more likely to work.

Have not yet played with Contour either, as it doesn't appear for me (opened another thread on that, which Keith is looking into).

[adatesman]

At the moment, I'm simply trying to quantify mechanical error in probing/digitizing. Part is a surplus 2.2190" master ring gauge, and it's looking more and more that X-Y probing has too much error due to backlash and kinematic error in the probe. End goal is measuring a logarithmic spiral, and likely writing a custom probing routine to use X and A instead, as that would eliminate backlash errors. Had opened a thread on that last week, but it fell from the front page.

[adatesman]

Here’s the part I’m trying to map. Spiral is quasi-logarithmic, and I’m trying to back into the modifiers they used, if any. I need radius over ~270 degrees. Guy who invented it made mention to me of using modifiers to the spiral ages ago, and I'm finally getting around to checking it.

[adatesman]

Had been doing it manually with an indicator and rotary, but a royal PITA.

[adatesman]

Yup, Wall Following works fine. Single pass, one Z depth.

At least, within the limits of my machine. Between backlash and kinematic error in the probe, I'm looking at 0.0056" error over 360 degrees.

Too much for what I need.

Edgefinding with X or Y and A would do just dandy, as I can cut the mounting boss on the rotary and it come in under 0.0002"TIR, at which point screw mapping takes care of error.

Does not look like there's a canned script for X/A edgefinding.

Guessing writing it won't be too bad, but have not dug that deep yet.

Hints appreciated.

[frijoli]

Yup, Wall Following works fine. Single pass, one Z depth.

At least, within the limits of my machine. Between backlash and kinematic error in the probe,

What probe are you using?

[adatesman]

Drewtronics.

Played around more last night to quantify error. First I verified the ring gage using a 0.0001” Mitutoyo dial bore gage set with gage blocks and a micrometer, ring is on diameter and circular under 0.0001”. Stylus is 1mm ruby, under 0.0002” TIR.

The bore probe cycle comes out at 2.2190” +/-0.0001” consistently with a probe diameter set to 0.0331”, regardless of starting position of the ring.

Ran Wall Following several times with different stepover, error seems to repeat with a max deviation of 0.0056”. Error always in the same section of the ring, hence thinking kinematics of the probe plus backlash in the screws. Gonna poke more tonight. “Corrected” column is due to it apparently not taking probe diameter into account, which isn’t surprising.