Solved Z-Axis Auto Leveling Routine Inconsistantly Starting

I've had my new printer for a few days and have a good success rate with my first prints, but I'm getting rather confused by the autoleveling routine.

I'm not using the software autoleveling, I'm using the 9-point routine.

Most of the time, the autolevel happens before each print, but sometimes it skips it and just starts printing. When that happens, the skirt is placed down smashed on one side and very loose on the other. In order to get the printer to level again is to turn it off, unplug the USB port, and reboot the computer.

Is there something that I'm doing wrong that prevents the routine from taking place before every print job?

FIX: The solution is to move the extruder temperature heat command (M109) to the top of the Start G-Code list.

For some reason with the current firmware, waiting for the bed/extruder to heat up causes the bed probing to fail. The bed probe seemed intermittent to me because sometimes I manually preheat, and sometimes I do not.

 

Here's my start gcode:

G28 ; home all axes
G1 Z5 F5000 ; lift nozzle
M109 S[temperature] ; set the extruder temp and wait
G28 Z0 ; Home Z again in case there was filament on nozzle
M565 Z0
G29 ; probe the bed

​

G29 is there, but for some reason, it doesn't always take. It's rather frustrating, because the prints come out very well if it happens at the start. I've let it go when it doesn't but the print usually doesn't make it past the first couple of layers.

It started happening again after I posted, and even rebooting the computer, unplugging the cable, and cycling the power on the printer wasn't making it do the 9-point autolevel. After a few attempts, it finally did it, and I was able to start the print that I was working on.

I'm going to put another instance of G29 at the end, to see if it might catch on the second one...

 

I'm not sure how to check my firmware version, but MatterControl prompted me to update the firmware when I started it up earlier this week. It says in the bottom-right corner that it's latest version.

I'm pretty sure that it's not waiting for the temperature to come up. There are two, distinct start patterns that I'm seeing. When it's working, it follows this routine:

1. All three axis go to the home position
2. Bed begins to heat
3. Home positions are maintained until the bed comes up to temperature
4. As soon as bed temp target is reached, the extruder begins heating and moves to the center position, then touches the bed and raises up
5. When extruder reaches it's temp target, the extruder raises and lowers to the bed again
6. The 9-position autolevel begins
7. Printing begins and usually is completed barring any other issues.

Occasionally, when I start a new print, the printer follows steps 1-4 then jumps right to 7, skipping 5 & 6, as soon as the extruder reaches temperature, then my prints are jacked because the Z-axis doesn't move at all unless it is moving to the next layer. There is no leveling being done at all.

When this is happening, I don't know how to force the machine to "reset" and start the bed probe commands again. At first, I thought I could unplug everything and turn it all off, but that doesn't seem to reliably kick it back into shape. It'll keep doing that, then out of the blue, it will start the 9-point search again on a new print.

I've removed the part out of the queue and re-sliced it, to see if it was something in that code, but that doesn't make a difference, either.

I've added the second G29 command, and each time I've tried it since, it's done a double-bed probe, but the extra time it takes doing that twice is a lot less than the time spent pulling up all of the bad plastic and starting over...

 

Thanks a lot for helping me out with this.

Here's the start of my gcode for the print I just ran. The machine is back in the skip-autoleveling mode.

; Generated with MatterSlice 1.0
; filamentDiameter = 1.75
; extrusionWidth = 0.4
; firstLayerExtrusionWidth = 0.6
; layerThickness = 0.2
; firstLayerThickness = 0.3
; automatic settings before start_gcode
G21 ; set units to millimeters
M107 ; fan off
M190 S50 ; wait for bed temperature to be reached
M104 T0 S210 ; wait for extruder 1 temperature
T0 ; set the active extruder to 0
; settings from start_gcode
G28 ; home all axes
G1 Z5 F5000 ; lift nozzle
M109 S210 ; set the extruder temp and wait
G28 Z0 ; Home Z again in case there was filament on nozzle
M565 Z0
G29 ; probe the bed
G29 ; probe the bed again
; automatic settings after start_gcode
T0 ; set the active extruder to 0
G90 ; use absolute coordinates
G92 E0 ; reset the expected extruder position
M82 ; use absolute distance for extrusion
; Layer count: 1
; LAYER:0
M107
G0 F9000 X134.20 Y147.20 Z1.00
G0 X134.20 Y147.20
; TYPE:SKIRT
G1
​

It appears that there is a bed temp set, M190, that happens before the extruder, as I'd noticed. There are also the two G29 commands, which are not being performed.

It's very strange, because everything else seems to be working.

 

Commented out the lines; no bed probe

 

I'll give it a shot tomorrow, as soon as it gets back into the "skip" mode. I started a 3-hour job, since it's currently working right.

Very odd. I do a lot of automation at work, and these situations are the worst, kind of like some kind of race condition. Almost certainly caused by a software bug.

I might try to reload the firmware, on the extremely unlikely possibility that something's wonky.

I'll let you know how it goes, and I appreciate the help. I'm loving the printer - when it behaves! I still need to learn some tricks, like how to print tall, thin objects. They go fine when they're printing layers with other things, but as soon as they're the only one left, they start warping. I'm sure it's because they get too hot, so I've got to figure out some way to keep it from happening, besides making extra copies on the same print! Maybe a better fan?

 

Alright, I've flashed the printer with the latest firmware on the Firmware page. MatterControl says there is an update available, but that's the one that I'd been using that has the weirdness. If this one works better, I'll just ignore the update notifications unless I hear that there's a better one released.

I'll post if this solves my problem or not.

 

Even with the latest firmware from the Robo3D page and Ziggy's simple gcode, I'm still getting the weird problem of the bed probe being skipped.

I'm not sure what's causing it, because when it's in that mode and I'm sending G29 commands during the terminal, it goes through the routine just fine. I don't know why it is skipped during the actual print gcode, and even when I load a gcode file that worked once, it still gets skipped when it gets in that mode.

In any case, I've done so much tweaking and leveling that even when it's skipping the G29 instruction, it's now pretty straight, only slightly smooshing the first layer a bit more than normal, and I can still get a good print out of it. I'm pretty convinced that it's some software/firmware/hardware issue that is particular to my machine or computer. I see this enough at work, where code written on one computer simply won't work right on another exact setup. Sometimes you find the problem, and sometimes you don't.

I'm just going to chalk this up to what you sometimes get from an inexpensive printer, I guess. I've got it tuned in pretty well, and I've been printing ABS like a madman. I must have gotten the magic spool, because with only Suave Max Hold hairspray on the heated glass, I've been getting perfect adhesion and almost textbook prints. With these results, I can easily work past the G29 problem. I want to calibrate my extruder next and see if I can squeeze a little more performance out of it, but overall, I'm very pleased with the printer.

 

Do you have G29 in your startup gcode

 

Yep, even put it in again for good measure, and when it works, it does it twice as expected.

It usually - but not always - works the very first time I turn on they system for the day.

For example, I'll load up the 300microndisk file and start the print. The G29 code is executed on the machine, and I get a perfect print. When it's done, I can print the very same file again, and there's a 50/50 chance that it will either do the G29 again or not on the subsequent print. I can use the exact same code or slice another, but it won't do any more automated G29s until it decides it wants to again. Every time I manually send G29, it works, even when it's misbehaving.

I was hoping that the printer was just keeping the G29 settings saved, but when it prints without it, the z-axis is dead steady when printing the layers. Before I got things manually leveled, it would result in disastrous prints. Now, at least, I can get a useable print, but the first layer is much nicer when the G29 is executed.

 

Are you loading Gcodes or STLs?

 

I copied the terminal output from two consecutive prints using the same gcode. The first time it didn't do the G29, but it did the second time. I noticed a couple of differences between the two, but I don't know how to interpret them.

Excerpt from the "good" print:

<-T:235.1 E:0 W:2
<-T:235.1 E:0 W:1
<-T:235.2 E:0 W:0
<-ok
->N12 G28 Z0*106
<-ok
->N13 M114*21
<-X:120.00 Y:127.00 Z:0.00 E:121.99 Count X: 120.00 Y:127.00 Z:0.00
<-ok
->N14 M105*18
<-ok T:235.4 /235.0 B:80.9 /80.0 T0:235.4 /235.0 @:74 B@:0
->N15 M565 Z0*91
<-ok
->N16 G29*37
<-Bed x: 15.00 y: 20.00 z: -0.00
<-Bed x: 107.00 y: 20.00 z: 0.11
<-Bed x: 199.00 y: 20.00 z: 0.18
<-Bed x: 200.00 y: 130.00 z: 0.11
<-Bed x: 108.00 y: 130.00 z: -0.02
<-Bed x: 16.00 y: 130.00 z: -0.27
<-Bed x: 15.00 y: 240.00 z: -0.04
<-Bed x: 107.00 y: 240.00 z: 0.13
<-Bed x: 199.00 y: 240.00 z: 0.33
<-Eqn coefficients: a: 0.00 b: 0.00 d: -0.15
<-planeNormal x: -0.00 y: -0.00 z: 1.00
<-ok
<-echo:endstops hit: Z:0.34
->N17 M105*17​

Excerpt from the "bad" print:

<-T:235.0 E:0 W:2
<-T:235.0 E:0 W:1
<-T:235.0 E:0 W:0
<-ok
->N14 G28 Z0*108
<-echo:Home X/Y before Z
<-ok
->N15 M114*19
<-X:120.00 Y:127.00 Z:5.00 E:121.99 Count X: 120.00 Y:127.00 Z:5.00
<-ok
->N16 M105*16
<-ok T:235.0 /235.0 B:80.3 /80.0 T0:235.0 /235.0 @:74 B@:0
->N17 M565 Z0*89
<-ok
->N18 G29*43
<-echo:Home X/Y before Z
<-ok
->N19 T0*2
​

It looks like something in the printer is preventing the G29 from being executed, and returning "-echo:Home X/Y before Z" to the computer. I'm sure it's an important clue, but I'm not sure how to get past it. Perhaps I can adjust the order in which the start code is written to make it work?

 

I added a new G28 command before the second z-drop. I'll see if it seems to work better.