Using G30 S0 inside homez.g fails

Ralms

Hi everyone,

Raising the issue which I believe to be a bug here instead of GitHub, as requested by the instructions.

In summary, I'm trying to use G30 S0 to home my Z axis and I'm running into inconsistent behavior.

Calling Home Z on DWC (which calls G28 Z and consequently homez.g), results in:
```
G28 Z
Error: in file macro line 5: G1: insufficient axes homed
```
However, calling the exact same Gcode, in the same conditions directly from Console, works and performs the homing of the Z axis.

Here is my homez.g:

var bedCenterX = (move.axes[0].max - move.axes[0].min)/2
var bedCenterY = (move.axes[1].max - move.axes[1].min)/2

M290 R0 S0 		   ; Reset baby stepping

G1 X{var.bedCenterX} Y{var.bedCenterY} F5000
M400                       ; added M400 as a troubleshooting step didn't make a difference 

G30 S0 X{var.bedCenterX} Y{var.bedCenterY} Z-9999 ; I've also tried doing just G30 S0, same result. 
M400

G1 Z20 F6000               ; Lift Z after homing. I've tested removing this, made no difference

The steps to reproduce the issue are:

Home X, called on DWC and wait it to finish
Home Y, called on DWC and wait it to finish
Home Z, called on DWC, fails with Insufficient axes homed error mentioned above.
Go to the console and execute G30 S0, works and homes Z, however, it still shows an error in the console:
```
G30 S0
Error: in file macro line 5: insufficient axes homed
```
After this, all 3 axis are considered homed (and with correct coordinates).

From the documentation, it says, that G30 without P, can be used to home the Z axis and the fact that it works outside of homez.g leaves me to believe to be a bug.

Hardware:

Duet 3 MB6HC running 3.5.2
Duet 3 Tool1LC running 3.5.2
DWC version 3.5.2

Please let me know if I can raise the issue in GitHub or if I'm missing something.

Also, if this is confirmed to be a bug, a workaround to unblock me would be appreciated, as I'm using the probe as a Z limit switch to home, no other option.
Thank you.

jay_s_uk

@Ralms you're using G30 incorrectly
it should be a single G30 with no other parameters
There are lots of us that don't have Z endstops and only use a probe.
I suggest you revisit the information about G30 in the gcode reference manual https://docs.duet3d.com/en/User_manual/Reference/Gcodes#g30-single-z-probe

gloomyandy

@Ralms said in [Bug] Using G30 S0 inside homez.g fails:

Error: in file macro line 5: G1: insufficient axes homed

The error you are seeing is being reported as being with a G1 command, not with a G30 command. Probably this line in your homez.g file:

G1 X{var.bedCenterX} Y{var.bedCenterY} F5000

This almost certainly means that for some reason when you are running the homez.g macro your X and/or Y axis has not been homed correctly. You should probably check your X and Y homing files (and possibly homeall.g) to understand why they are not working.

Ralms

@jay_s_uk said in [Bug] Using G30 S0 inside homez.g fails:

@Ralms you're using G30 incorrectly
it should be a single G30 with no other parameters
There are lots of us that don't have Z endstops and only use a probe.
I suggest you revisit the information about G30 in the gcode reference manual https://docs.duet3d.com/en/User_manual/Reference/Gcodes#g30-single-z-probe

I'm not using G30 incorrectly, I know the documentation and I've read it like 5 times troubleshooting this.
It says

G30, G30 S0 or G30 S-4 or lower are just normal probes. When the probe is triggered, sets the Z coordinate to the probe trigger height.

and as any other configuration in RepRap, you can either let it use the defaults or be specific, I'm being specific here.

Still, I've tested with G30 only, same error.

@gloomyandy said in [Bug] Using G30 S0 inside homez.g fails:

@Ralms said in [Bug] Using G30 S0 inside homez.g fails:

Error: in file macro line 5: G1: insufficient axes homed

The error you are seeing is being reported as being with a G1 command, not with a G30 command. Probably this line in your homez.g file:
G1 X{var.bedCenterX} Y{var.bedCenterY} F5000
This almost certainly means that for some reason when you are running the homez.g macro your X and/or Y axis has not been homed correctly. You should probably check your X and Y homing files (and possibly homeall.g) to understand why they are not working.

It's not, the printer executes this G1 command and stops only after when is about to run G30. I've tested this.
Also, at this point both X and Y are homed, so this G1 command should be valid regardless.

Lastly, doesn't explain me running "G30 S0" on the console executing and throwing the error on itself.
What I suspect is happening, is G30 calling G1 itself internally or something around those lines.

@gloomyandy @jay_s_uk To reduce any further discussion on your arguments, I've went ahead and
empty my homez.g, having only the "G30", as such:

; homez.g
; called to home the Z axis
G30

Literally, nothing else and it fails anyway with the same error.

@dc42 Could you please confirm if this is in fact a bug or some nuance? thank you

jay_s_uk

@Ralms you can't use X and Y coordinates in a G30 without a P value...

chrishamm

@Ralms I highly doubt there is something wrong with G30. Do you have a deployprobe.g and/or retractprobe.g? We also need to know what kind of probe you have configured.

Ralms

@jay_s_uk said in [Bug] Using G30 S0 inside homez.g fails:

@Ralms you can't use X and Y coordinates in a G30 without a P value...

It's irrelevant, I just showed you that running G30 without ANYTHING ELSE, also fails....

jay_s_uk

@Ralms ok, good luck then.
Works ok for me here with 3.5.2

Ralms

@chrishamm said in [Bug] Using G30 S0 inside homez.g fails:

@Ralms I highly doubt there is something wrong with G30. Do you have a deployprobe.g and/or retractprobe.g? We also need to know what kind of probe you have configured.

Well, I stand corrected, with you asking about deployprobe.g I remembered that I had some old draft gcode there and was the source of the error.
Removing the gcode from that file, allowed G30 to execute.

Still, doesn't explain why G30 will execute when called from the Console, ignoring an error at deployprobe.g.
Which is concerning, as it leaves the possibility of having the printer probe the bed without the probe being ready.

Regarding your question, about the probe, I forgot to share it but, essentially its a Klicky probe, running as type 8 due to being wired on the 1TL.

M558 P8 C"^121.io2.in" H5 F200:100 K0 T2000

@jay_s_uk Just to finish on your insistence in how I was running G30.
Running any of the mentioned configurations for G30 works and probes the bed successfully, that being G30, G30 S0and G30 S0 Xnnn Ynnn Z-9999.
However, specifying S0 parameter will execute anyway, probe the bed and set Z, however throws error Error: Number of calibration factors (0) not equal to number of leadscrews (4).
The same thing can be said for the parameters X, Y and Z, they are all just ignored, but return no error.

So yes, you were right that S, X, Y and Z parameters should not be used with G30 in this scenario but they were not the source of the issue, as I showed that even running without them would fail anyway.

droftarts

@Ralms Thanks for reporting this. I can confirm that G30 S0 does seem to be interpreted by the firmware as having a P parameter, and causes the error messages, even though the actual probing/homing completes successfully. I'll raise an issue to get this fixed, as G30, G30 S0 and G30 S-4 and lower should be synonymous.

For now, don't use G30 S0 for homing, or S0 without a P parameter, until this is fixed, to avoid the error message. Use G30 on it's own.

@Ralms said in [Bug] Using G30 S0 inside homez.g fails:

So yes, you were right that S, X, Y and Z parameters should not be used with G30 in this scenario but they were not the source of the issue, as I showed that even running without them would fail anyway.

I think the code in your deployprobe.g was probably unhoming the X and/or Y axes. Homing by probing in Z (as opposed to homing Z by hitting an endstop) requires X and Y to be homed. This caused the initial 'Error: in file macro line 5: insufficient axes homed' message.

Ian

Ralms

@droftarts Thanks for confirming.

The issue I had in my deployprobe.g was that I was trying to raise the gantry in absolute mode without having Z homed yet.
The goal of raising it is to avoid the probe hitting the bet when its attached.

Regarding G30, yeah, I will be using it without any parameter, its achieving the same end-goal, so I'm ok with it.

I just hope that someone doesn't get caught off-guard by G30 ignoring failures in deployprobe.g when called from the console, as it could mean the probe is not even deployed/attached and than the printer rams the hotend into the bed waiting for a trigger.
This is a very niche scenario and would require the probe to be setup incorrectly (normally open instead of the more reliable normally closed and opens on trigger), but still, its a possibility.

engikeneer

@Ralms note that when you call a homing file it first flags the axis as unhomed. My guess is when you did it from command line you had already homed the z axis?

Ralms

@engikeneer said in Using G30 S0 inside homez.g fails:

@Ralms note that when you call a homing file it first flags the axis as unhomed. My guess is when you did it from command line you had already homed the z axis?

When calling from the console, is just doing G30.
G30 only resets the Z axis and doesn't touch X or Y.
And yes, before calling G30 from the console, I already have the X and Y homed and with the probe in place.

engikeneer

@Ralms so basically, when you called G30 from console, Z was already homed, so the problematic G1 line to move an unhomed axis in your deployprobe.g wasn't an issue? But when calling homez.g with just a G30 in it, the z axis is marked as unhomed before deployprobe.g is called and hence you got the error. Your problematic command was to move z anyway so x and y don't come into it.

My point is that I think the behaviour you're seeing there is fully expected and 100% explainable

Ralms

@engikeneer said in Using G30 S0 inside homez.g fails:

@Ralms so basically, when you called G30 from console, Z was already homed, so the problematic G1 line to move an unhomed axis in your deployprobe.g wasn't an issue? But when calling homez.g with just a G30 in it, the z axis is marked as unhomed before deployprobe.g is called and hence you got the error. Your problematic command was to move z anyway so x and y don't come into it.

My point is that I think the behaviour you're seeing there is fully expected and 100% explainable

You miss understood.
Z is never homed in any of the scenarios I've mentioned (doing Home Z in the UI or calling G30 in the console), only X and Y are.

So yes, my G1 move in the deployprobe.g was a problem always as it was trying to move Z20.

The issue I'm raising now, is that G30 will execute from the console, even if deployprobe.g fails, it just ignores it and that should never happen.

oliof

@Ralms if you want to retain the clearing move on the z axis, you could do an unhinged move like this

G91 ; relative moves
G1 Z10 H2 ; move up 10mm
G90; absolute moves

droftarts

@Ralms said in Using G30 S0 inside homez.g fails:

The issue I'm raising now, is that G30 will execute from the console, even if deployprobe.g fails, it just ignores it and that should never happen.

G30 works from the console or from within a macro, whether the Z axis is homed or not. The X and Y axes must be homed. When G30 is called, it will run deployprobe.g if the probe type requires it. When and why might deployprobe.g fail, and how would it know if deployprobe.g has failed? If running G30 from the console, it could possibly be too low to fully deploy, eg BLTouch, or off the bed, but I'd say this was operator error rather than the firmware's fault. You could have a check in deployprobe.g for current Z height, if it is below say 2mm, or if the axis is unhomed, move up 5mm. The stock homez.g from the configuration tool moves the Z axis up 5mm before homing to avoid this.

Ian