Autoleveller

After many more attempts and running out of PCB boards I have finally ran out of options as the auto levelling will not work correctly and I am unsure where to look next.
I have attached the gcode files so you may find what is wrong.

I am using mach3 USB CNC machine which works with the G31 probing code as the manufacture also supplies a macro so you can level the tool tip which is using the G31 code so I can assume that’s not my problem I have also created a macro called M2002 and set the value SetVar(2002,getoemdro(802)) and im using the custom option with zero word set to G52.
Every time I run the PFG file on my machine it appears to probe perfectly each time then when I try and load the PFG file in the Autolevel software it always fails and I have to use the repair option which produces the fixedbottomprobe.txt file this file of course loads and then I add the levelling code to bottomcopper.tap which produces ALbottomcopper.nc I then run this gcode on my cnc machine but every time the engraving gets messed up by either getting open air or too deep engraving tracks which destroy the copper tracks. I am probing with a 0.1mm 10 degree V tool tip and then I engrave with the same tool tip. Currently I cannot engrave a track any smaller than 20th im using proteus to create my circuits GBR files then I use Flatcam 8.5 to create the Gcode. I have noticed that the probe file bottomprobe doesn’t have the X,Y,Z in front for the coordinates but the fixedbottomprobe file dose I don’t know if this has anything to do with it!.
any help would be appreciated at this stage.
I am using version AutolevellierAE 0.9.4u2
Regards
Nigel

Nigel,

First of all, thank you for the details and attachments, the more information, the easier it is to find the problem.

USB is almost always the issue in these cases. However, looking at your RPF, the probe values look perfectly normal to me so the USB is probably not the problem (I cant say for certain how accurate these values are however).

The next question is, how are you holding the board down? Clamps, double-sided tape, vacuum table?

The fact that you have such thin traces means you will need a very shallow Z depth as you likely know. Looking at your original GCode, Z is set to Z-0.0010. I am wondering if this is too shallow a depth. For example I have my depth set to about 0.1 - 0.2mm but this is not SMT work, still 0.001mm seems overly shallow to me. Maybe someone else with more experience than me with SMT can share their numbers?

I know Art uses a vacuum table but what engraving depth? im sure he will see this and tell us more. BTW, 0.001mm is ~0.00003937"

Nigel,
I'm here and will attempt to fill in some of the details that James mentions.

Yes, I use a vacuum table to hold my boards and still use Autoleveler due to the "waves" in the board structure itself.

As to the depth of cut, we first need to know the thickness of the copper layer and of course want to cut to a depth greater than that. From:
https://www.pcbuniverse.com/pcbu-tech-tips.php?a=4
we find

1 oz
1.37 mils (thousandths of an inch) 0.00137 inch

0.0347 mm 34.79 µm (micron/micro meter)

(1 oz Copper Thickness Conversion
*To determine the thickness of 4 oz,
simply multiply 1 oz thickness x 4.
1.37 mil x 4 = 5.48 mils = 4 oz)

So if your trying to cut at .001 mm depth of cut, you will NOT be cutting through the copper even at 1 oz thickness! And this assumes the copper is evenly distributed over the board surface. (which it very well may not be)
The absolute minimum I have been able to get away with is .002" (0.0508 mm) and that is on a small board and the maximum is usually around .005" (0.1016 mm) . Again, there can be tolerance errors due to the mechanics of the system. One big source is stepper motor resolution (especially with micro stepping) is not guaranteed to be exact but I don't remember the tolerance at this time. Then we need to consider the overall rigidity of the machine and fixtureing of the pcb. Backlash in the lead screw can also cause a problem as I have had engravers that for whatever reason would not initially penetrate the copper and simply "skate" across the top of the copper. This occurred on my original machine where I had backlash in the Z axis lead screw.

Next, we have to address the "actual" width of the tool doing the engraving. Without testing, we don't know as the actual width of the tip of the engraver is not exactly "0 mm". To do this, I use the system developed by Poul-Henning Kamp and can be found here:
<http://phk.freebsd.dk/CncPcb/calibrate.html>
to do the calibration of my tool. Another positive effect of doing this calibration is that it will also include any run out that you may have in the spindle and/or the tip of the tool itself.

I have found that each of my engraving bits (even from the same source and of same "specification") will very slightly and this can make a big difference in the depth of cut needed.

I probably have added more confusion to the problem at hand, but these are some of the factors I have found that go into trying to make decent boards by this method. Hopefully this will be of some help in determining how to solve your particular problems. If I can be of further assistance, please ask away as I monitor the board most of the the day (Eastern timezone in the United States).

Regards
Country Bubba