Wierd lines on the test cube sides

skayred

@mrehorstdmd I don't have a photo of the final one but it is based on 3060 t slot with metal nuts, angles, and cnc plywood frame about 30mm to keep towers.

I made a test cube with Matter slicer, and the problem seems to be a bit different - fat lines after retract or layer change are still there but a bit shorter.

I also tried to experiment with pressure advance but it doesn't affect at all for my 40cm Bowden

mrehorstdmd

A more detailed description of the machine, such as the architecture, size, Z axis construction, etc. would be helpful. This type of problem is often due to less-than-ideal mechanical construction. So far, you haven't provided enough information to allow any sort of diagnosis. Use the same camera that you used to take pictures of the prints to take pictures of the printer and post a few here.

skayred

@mrehorstdmd please find the pictures below:

So, just a Delta printer with lead screws and PTFE nuts

dc42

Here are some possibilities:

Your machine has play in the joints (causing backlash), and the slicer has decided to print the perimeters in opposite directions on alternate layers.
You have not run heater tuning, so the bed heater is running in bang-bang mode

skayred

@dc42 hello, here is what was printed with Matter Control:

so, walls quality is more or less okay so in the previous example, I suspect, the problem is overextrusion after retract. According to S3D troubleshooting guide, it could be solved by the negative extra restart distance (or I have problems in the extruder - incorrect low speed or jerk).

But, according to Duet docs, extra restart distance is generally bad practice and pressure advance should be used instead.

denke

Hi @skayred !

It's hard to see from this angle, but it seems to me that your part is solid?
If so, any (even a very very very slight amount) of over-extrusion can cause effects like this, because there is no other place the excess material can go but off the side of the part (or build up on top and cause the head to eventually collide with it) Plus as DC said maybe each layer is printed in different direction, and there you have it.

If the part is indeed solid, then please try to print it not 100% solid, or decrease the extrusion factor a few percent (10% to be sure for the test) and show us the results

skayred

@denke hello, it is not even close to be solid, it was 25% infill But perimeters are printed inside-out, could it be the problem?

@dc42 thanks for the advice! I just found that bed was for some reason in bang-bang mode indeed, doing PID tuning right now, let's see how it will go after

denke

@skayred Probably not ... maybe if you have many perimiters, but unlikely

skayred

@denke in that case, maybe I am missing some calibration models? Like, I didn't calibrate extrusion width and so on, just started printing test cubes with 3 perimeters at once. If you know what to do first, it would be nice to hear

mrehorstdmd

I think you need to fix the backlash in the mechanism before you can calibrate the extruder. Once the layers stack properly on top of each other you can adjust extruder calibration based on line/wall width measurement.

skayred

@mrehorstdmd hello,

just checked all towers - it slides properly, and I have preloaded PTFE nuts there, so I didnt find any play there.

@dc42 after PID tune, I repeated very first G-code, it seems to be just a little better (no random lines) but overextrusion right at the beginning is still there. Going to experiment with negative restart distance

skayred

Small update - I just printed 66mm diameter turbine impeller and identified interesting thing - one arm ball at one exact position was going out of magnet's center. I have added some bearing grease, and it prevented the balls from escaping, so impeller was printed without any issues, just really small bowden scars. Still not dimensionally accurate but I think it was the problem

skayred

Another update - seems to be fixed. What I did:

First of all, I have replaced all steppers - old ones were faulty, their shafts were possible to move along, about 3-5mm. New ones don't have these issues and work normally on 2A
Replaced 6-factor calibration with 8 factors
And finally, adjusted cartesian scale

Edgars Batna

@skayred said in Wierd lines on the test cube sides:

First of all, I have replaced all steppers - old ones were faulty, their shafts were possible to move along, about 3-5mm. New ones don't have these issues and work normally on 2A

Move in the direction of the shaft? This might be by design. All steppers that I've disassembled so far were not constricted to carry any load on the shaft other than rotational. If a stepper motor is new, then it might just have enough friction in the internal parts for it to appear to work, but this will change the longer it runs or depending on the temperature. The coils were centered vertically only by the magnet from my observations.