I have been working with auto-leveling and have three questions and one new approach

CAD & CAM Tool Chain ChiliPeppr
1

I have been working with auto-leveling and have three questions and one new approach others might be interested in.

Question 1: As shown in the attached video, during probing, the probe seems to move substantially in the XY plane. From what I can tell, it doesn’t affect anything, but I still suspect I am doing something wrong. Thoughts / Suggestions?

Question 2: Am I using the layers correctly? After positioning the spindle to 0,0,0 I zero out G54, then G92, first for X, then Y then Z. Is this proper? If so, I can’t see a purpose for the G92 layer. Perhaps this accounts for my problem in question 1 above?

Question 3: After I get the auto level data, the individual values shown on the display are obscured by the graphic of the actual Z values. Is there a way to easily see those individual values? Is there a way to change those individual values?

New Approach. As I described previously, I am using a fast relay to allow me to probe with 9v. When the probe makes contact, a light on the relay illuminates (which is very handy) and the relay closes to provide the tinyG with its signal. I think it is working well, but I think I still occasionally lost steps, as evidenced by the Z value at 0,0 being significantly off after 400+ test probes. I was also scratching boards (stupidly). To address both problems, I improvised a probe by taking a broken end mill (I have a big supply of those) and attaching the spring from a ball point pen. The bottom of the spring extends a couple mm below the endmill, so even if the probe travels beyond contact, the spring absorbs the excess, no steps are lost and my board isn’t damaged. And if I forget to raise the probe before jogging it, no harm is done. I can’t figure out how to post a pic but will do that separately.
https://youtu.be/4wDMamCQy9g

2

Here is a pic of the relay and probe:

missing/deleted image from Google+

3

The auto-level widget works in machine coordinates, so using G92 to establish a temporary WCS is what’s causing your problem. You’d have to fork/modify the auto-level widget to have it support other work coordinate systems.

4

Also, on the question of the values being obscured, yes that was a design flaw in the original design approach. Perhaps the Z level of that data being written to the 3D viewer should always look at the highest Z probed and then always position above it. That would take some new code to be written to look for that. You can always look at the JSON formatted probe data as well as a stop-gap. It’s in the upper right corner triangle menu I believe.

5

So I will not use / ignore G92, but it sounds like I should somehow be using machine coordinates instead of even G54?

6

If the G54 offset is 0,0,0 then you will have a good experience as it mirrors machine coordinates. That’s sort of assumed that it’s what people are doing to keep things simple.

7

Thanks it is working nicely.

8

Agreed with John, I autolevel every job and only ever need the G54. I use G92 on my larger machines between tool changes… if I mess up .

9

Thanks, Brandon.