Weird behaviour after first print
-
Problem is still there
There is a new additional error message after the upgrade.
**Error: G32 has thrown an exception: [NullReferenceException] Object reference not set to an instance of an object.** Error: Probe already triggered at start of probing move -
And with G29 I now get
-
@martinnyhc There must be an exception or cancelled code somewhere in your bed.g which is why you get that exception. I've got a fix ready that will be included in 3.4-b3. Also you seem to call G29 without S parameter in mesh.g, which will invoke mesh.g over and over again until you get a stack overflow.
-
Yep, already fixed that. I need to do some further testing regarding the probe. In the meantime if did a fresh SBC install.
-
OK, spent the day with doing a clean SBC install (RRF 3.3.0) and rewriting all config files and macros. The main issue I have still persists. Gantry and bed leveling works like charm if it is the first print. As soon as a print finished or where cancelled and I start a new print it fails.
Fails means for example: The first thing of course is homing. Second thing is gantry leveling.
; bed.g ; called to perform automatic bed compensation via G32 if !move.axes[0].homed && !move.axes[1].homed && !move.axes[2].homed G28 ; home M558 K0 H10 F400 ; set trigger height M561 ; clear any bed transform M98 P"0:/macros/sys/undock_klicky_probe.g" ; undock clicky probe while iterations <=1 ; first pass G30 P0 X20 Y25 Z-99999 F9000 ; probe Z left front G30 P1 X20 Y320 Z-99999 F9000 ; probe Z left rear G30 P2 X330 Y320 Z-99999 F9000 ; probe Z right rear G30 P3 X330 Y25 Z-99999 S4 F9000 ; probe Z right front G1 H2 Z12 F2600 ; raise head M400 ; finish move, clear buffer M558 K0 H2.5 F200 ; trigger height to 2.5mm after first run while move.calibration.initial.deviation >= 0.003 ; Perform additional leveling if previous deviation was over 0.03mm. G30 P0 X20 Y25 Z-99999 F9000 ; probe Z left front G30 P1 X20 Y320 Z-99999 F9000 ; probe Z left rear G30 P2 X330 Y320 Z-99999 F9000 ; probe Z right rear G30 P3 X330 Y25 Z-99999 S4 F9000 ; probe Z right front G1 H2 Z12 F2600 ; raise head M400 ; finish move, clear buffer echo "Gantry deviation of " ^ move.calibration.initial.deviation ^ "mm obtained." G1 Z12 ; Raise nozzle M98 P"0:/macros/sys/dock_klicky_probe.g" ; dock clicky probe G28 ; Homing again for accurate Z offsetIt probes the four points for the first time, changes the probing height and probes for the second time. Suddenly it is just going to the probe points but don't lower Z to probe the bed.
Here's an archive of my complete config. I've spent so many days with this problem now and did everything which came to my mind, I'm definitely stuck and don't now what to do.
Help please!
Thanks in advance! -
Are there example configs from other voron users that use this probe?
Have you tested in standalone mode as a control?
-
There are many many other users and I took over the config from them. And it already worked for multiple weeks.
Nope, not tested standalone. But as said I just did acomplete fresh SBC install.
-
Well I'd say either test your current config files in standalone mode to eliminate the SBC for testing, or wait for chrishamms updated build.
Right now we're mixing 3 confounding factors, unproven macros, SBC mode, and beta firmware. When in doubt, try and roll it back to the most basic configuration and when it's working, start adding new things one at a time. -
@martinnyhc try
; bed.g ; called to perform automatic bed compensation via G32 if !move.axes[0].homed && !move.axes[1].homed && !move.axes[2].homed G28 ; home M558 K0 H10 F400 ; set trigger height M561 ; clear any bed transform M98 P"0:/macros/sys/undock_klicky_probe.g" ; undock clicky probe G30 P0 X20 Y25 Z-99999 F9000 ; probe Z left front G30 P1 X20 Y320 Z-99999 F9000 ; probe Z left rear G30 P2 X330 Y320 Z-99999 F9000 ; probe Z right rear G30 P3 X330 Y25 Z-99999 S4 F9000 ; probe Z right front G1 H2 Z12 F2600 ; raise head M400 ; finish move, clear buffer M558 K0 H2.5 F200 ; trigger height to 2.5mm after first run while move.calibration.initial.deviation >= 0.003 ; Perform additional leveling if previous deviation was over 0.03mm. if iterations = 5 abort "Too many auto tramming attempts" G30 P0 X20 Y25 Z-99999 F9000 ; probe Z left front G30 P1 X20 Y320 Z-99999 F9000 ; probe Z left rear G30 P2 X330 Y320 Z-99999 F9000 ; probe Z right rear G30 P3 X330 Y25 Z-99999 S4 F9000 ; probe Z right front G1 H2 Z12 F2600 ; raise head M400 ; finish move, clear buffer echo "BTC: " ^ iterations + 2 ^ " - Difference was " ^ move.calibration.initial.deviation ^ "mm" continue G1 Z12 ; Raise nozzle M98 P"0:/macros/sys/dock_klicky_probe.g" ; dock clicky probe G28 -
Did you solve your problem? I'm struggling with QGL and mesh with the klicky probe... My problem is, that I use the autoz calculation macro at first and then I try to perform a QGL... After the homing and autoZ macro, the Z0 is at the surface of the bed plate - as expected. But then my QGL fails...
-
@medicusdkfz For me it was a loose connection on the klicky probe which only occured randomly.
-
@martinnyhc Please would you share your RRF3-configuration with me? You'll find my mail adress in my profile...
Thank you in advance...
-
How is your process for creating a mesh from scratch? Do you also use a seperate z-microswitch for calculate the difference between the nozzle and the switch point of the klicky probe?
-
You have to call the probe deploy from within your homez and then again in your beg.g or other macros that need the probe. Conditonal gcode makes this an easy check and remedy.
The error is because the NC probe state is 1000 when the probe is not connected or triggered. you have to include in your macros the logic to address these states.
You should go look a the macros for Euclid Probe. We've been running this reliably for over 2 years at this point, both with fixed dock and moving bed pickups.
Also, we have found that inspecting the actual probe value is more reliable that checking the probe token in the object model. In the case of a mis-deploy, which is really rare with Euclid Probe, the probe value is more telling.
