Why did my steppers suddenly going backwards and nozzle start scraping the bed on a subsequent print?

General Discussion
1

Ender 3 Max
Creality 4.2.2 with TMC2208/HR4998 steppers
Genuine BL-Touch Smart v3.1
Mainsail
Running klipper on a laptop connected via usb
klippy (1).log (2.9 MB)

What went horribly wrong that caused my printer to start extruding backwards, and completely forget about the bed mesh?
Prior to this I had just finished a screws_tilt_calculate, z offset adjustment and bed mesh, then I started my pressure advance tuning box per exactly as the docs state to, except a few layers in after the base, I realised I didnt set all the speeds to 100 mm/s so I cancelled the print, re-sliced with only changing the relevant speeds to 100 mm/s, preheated, cleaned the nozzle and the whole show goes tits up from there, its shooting filament out the back of the extruder and moving all cockeyed, scratches the bed before I realized what was going on and cancelled the print.

Also, is there any info repo available for instructions on how to read and understand the log file in order to troubleshoot myself?
The docs only mention the logextract.py but I’m not sure thats what im looking for…?
Anyway I’d greatly appreciate any help in figuring out why this printer decided to go haywire out of no where on me.

2

Can you share the file you wanted to print?

BTW: The homing speed for X and Y is way too fast. Default is 5 mm/s. You should not exceed 20 mm/s. Homing is a precision process and should not be done with a sledge hammer.

1 Like
3

gcodes-2024116-03151.zip (368.4 KB)
thats the gcode print file directly from mainsail
square_tower.stl (65.3 KB)
this is the stl if thats what you’re after though just in case, but I got it directly from the link in the klipper docs pressure advance tuning section

Thanks for the tip on homing speed, I was kind of iffy about the potential for repercussions by setting it high but this was much earlier on in my experience.
Could it be that its not homing correctly?

I have my bed mesh speed at 200 mm/s is that also an issue?
at least for this particular problem that cant be the case since the first print started fine, but based on your the homing speed tip im guessing that applies to the accuracy too?

I will set the homing speed to 20 mm/s and try it again. Will let you know how it goes.
Thanks for the troubleshooting references.

4

Just ran a screw_tilt_calculate again and somehow the right side front and rear both need a full turn CW.
Im gonna check the bed roller tensions because i’ve been having all my bed meshes with a tendency to rise on the left side.
Will get update if I have any luck.

1 Like
5

Just wanted to write don’t use relative extrusion mode but you begin and end your gcode with absolute extrusion command already.

1 Like
6

Thanks for your insight

7

ok well, the print just started, so far so good, things have gone back to normal, guess im never gonna figure out why it shot the filament all the way out the back of the bowden tube but it was weird that the bed level was that far off within such a short time of calibrating it.

I did the bed mesh at 100 mm/s and im thinking of setting the homing speed to that too because i almost did the dishes between steps it takes way too long.
The way I see it if I think about it logically, im gonna be printing at 100 mm/s, so if the calibration goes out of adjustment because of that speed, it only means the adjustment will go just as far out as soon as the print starts…
to make matters worse I have the travel speed of moves between printing set at 250, so realistically, there’s no dramatic effect from the high movement speeds, because i’ve already printed 4 calibration towers with very high speed settings, albiet this doesn’t discount the presence of a gradual creep in misalignment, sure I mgiht have an issue forming over time, (that being no less than 8 hours of print time) but I cant make reason to put the speeds that slow when I only stand to lose time and gain nothing

EDIT:

OK well as I was writing this I came back to a mess so somethings definitely going on…

IMG_20240216_025437
IMG_20240216_0254371920Ă—2560 162 KB

For reference, I redid the screw calculate adjustment, z offset and bed mesh all over again, atleast last time the first subsequent print got well further than this.
The bed rollers were fine, back one needed a little nip up maybe 3 degrees but no way near enough misalignment to cause this mess.
So I’m back at sqaure one.
Given the random extruder reversal filament ejection job I’m guessing something had to have gone wrong software related.

If you have any ideas im more than greatful.

8

You may double check the hardware, pulleys etc.

Check the Vref of the stepper drivers.

9

I spent a whole day going over everything mechanical, recalibrating tilt, x offset, bed mesh, retraction settings, travel settings, i’ve tried everything I can think of, but I cant get a print to lay the first layer smoothly, it keeps stringing and blobbing and catching itself, I got close to almost a succesful print but it still had issues, I cant figure out what im doing wrong, there’s a million variables and every “feature” feels like its just conflicting with every other feature, I’m trying to understand what my printer.cfg conflicts with in mainsail, in cura, in anything and everything I try, I’m just going round in circles this is a mess

10

You have a quite simple printer and if the mechanical parts and extruder is ok, it should work fine with the right slicer settings.

Maybe upload your current printer.cfg and check slicer start and end code and what Klipper macros might be involved.

You could just load a pre-defined Cura profile and calibrate the printer for the used filament.

1 Like
11

Thats what I was trying to do, I just put the start and end gcode into the printer.cfg because it was recommended but it seemed to work fine until it didnt, I dont understand where I went wrong, the printer.cfg has been tweaked back and forth in a lot of ways but this is as close as I can get it to almost printing
printer.cfg (6.3 KB)

12

I think you should at least load the bed mesh after homing all axes.
So put it after the G28.

But maybe others using bed mesh already for longer time can share some tips here.

Otherwise to not overload the chinesium PSUs you should first heat up the bed as this takes the most time and then heat the hotend.
Otherwise filament might already cook in your hotend while the bed still heats up…

Advantage for splitting the heat processes is that while the hotend heats up the bed has some time to settle its temperature.

So maybe something like that:

    M190 S{params.BED_TEMP|default(60)|float} # heat to bed setting

    #maybe home axes here and load bed mesh...

    M109 S{params.HOTEND_TEMP|default(200)|float} # heat to hotend setting

Your macro does not take into account the temperatures provided by the slicer.
Depending on your used filament the first layers can fit or not.

If using my snippets you should change your Cura start code to pass the parameters:

;Filament weight = {filament_weight}
;Nozzle diameter = {machine_nozzle_size}
;Filament type = {material_type}
;Filament name = {material_name}
;Filament Scaling Factor = {material_shrinkage_percentage}
PRINT_START BED_TEMP={material_bed_temperature_layer_0} HOTEND_TEMP={material_print_temperature_layer_0}
1 Like
13

Thanks for the help, also I do usually manually preheat it just so I dont have to wait and it has the meanwell psu that are supposed to be good but idk
I changed it to heat without waiting after these issues started though just because I was getting impatient, i’ve probably had about 20 failed prints today just trying to do pressure advance

;Nozzle diameter = {machine_nozzle_size}
;Filament type = {material_type}
;Filament name = {material_name}
;Filament weight = {filament_weight}
; M190 S{material_bed_temperature_layer_0}
; M109 S{material_print_temperature_layer_0}
START_PRINT

This is the snippet I had in my start gcode on cura, I’ve also been trying to use Ellis’s pressure advance tool to get the gcode directly
https://ellis3dp.com/Pressure_Linear_Advance_Tool/

14

Impatience is not the best when 3D printing, but I know what you mean.
If you put the heating into the slicer start code you should adapt/remove the respective commands in your START-PRINT macro.

However this is only part of the whole process.
Did you calibrate feeder esteps and flow of the used filament?
For a good first layer you will need a clean bed and nozzle, the right temperature of both and the right distance to the bed, respectively the correct filament flow and depending on the print temperature of course the right printing speed.

Z=0 means the nozzle barely touches the bed. At least in theory as in real life the bed often is not 100% plain, hence bed mesh is needed…
If the first layer is not closed or over extruded though using the correct bed distance you can adapt the (first layer) flow (up or down) until it is fine.

Stringing and blobbing there might be over extrusion maybe coming from your purge line where too much filament is being pushed through the nozzle.
Best is to show a photo of such a print attempt.

3D printing is a cycle with dependencies and if you change one parameter it will affect others.

2 Likes
15

E steps are calibrated, I printed a temperature tower and flow tower for this filament, at the start of the day I washed the glass bed with a sponge and dish soap thoroughly, it air dried, I have been keeping it up to standard with a lint free cloth and 98% ethanol.

I completely dismantled the hot end today and to verify there were no clogs, it was fine, the ptfe tube has a clean cut hard up against the nossle, I have a good set of accurate toledo feeler gauges that I know I can trust as i’ve only used them to check valve clearances and this is what I use for z offset checks, I’ve also tried the using the “light method” when the bed is at full temp.

I set the temperature to 67 specifically because I didnt get good enough adhesion early on with 60, and I was going for 65, but when I checked the actual physical print surface temperature with an infrared laser thermometer it showed a consistent 65 when mainsail reports 67.

I was trying to stay cognicient of the cycle of paramaters but I was just using my logic to determine what I was adjusting since I dont have the experience to know what to look out for. I ended up finding that 60 mm/s retraction speed with a distance of 0.5mm seemed to do the best, and setting a higher travel speed (200-250 mm/s) gave me a better chance and stopping a blob from forming

1708081628522974474109657915148
17080816285229744741096579151481920Ă—1440 359 KB

17080816822763741890462656235124
170808168227637418904626562351241920Ă—2560 463 KB

17080818330064496580057337282357
170808183300644965800573372823571920Ă—2560 205 KB

16

This just looks like bad bed adhesion.
Did you try lowering the bed temp to slightly below 60°C?
Those ultra bases have good adhesion until they don’t have. :smiley: Maybe you can grind it a bit with very fine sand paper. But that can also destroy the surface as well.
So better play with temps at first.

I don’t like those heat towers as they often do not show the material under full printing speed as the head is always accelerating and decelerating and often not reaching its commanded print speed for some time.

Maybe you can chose a hollow tower with same appropriate size.

How fast is your first layer print speed?
If your print head is still stock maybe some predefined templates for your printer will work fine to start with.

17

My first layer print speed was 20 mm/s

I lightly scuffed the glass with the scrubby side of the sponge but it was an abrasive safe for non stick pans so it might not have done anything

Im just gonna start again this is a mess

If you have any resources for good heat towers I can download I would appreciate it greatly

I purchased the califlower stl pack from vector 3d which I have been trying to work my way up to but if I can’t even get anything to stick let alone tune pressure advance I don’t see it happening any time soon…

18

Until you can use the Califlower all the rest has to be set up properly.

You could insert a cube with maybe 10x10x10 cm and print it in spiralized/vase mode with either big brim or 1-2 bottom layers.
With post processing change at layer you can change the temp every 5 mm about 5 °C starting with the upper limit.

19

Thanks for that suggestion Brian,
I ended up doing some investigating into the gcode files of the successful prints to backtrack till I could find what changed that caused everything to go wrong.
Unsurprisingly, the temperature towers started at 230C since that was the highest starting temp, but turns out I didn’t change the 230 degree starting temperature for the flow towers so the only reason I had good adhesion was because it was set at 230C, which is a bit excessive, so im going down in 5 degree increments to find out what temperature the bed adhesion starts to fail at again

20

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.