Gaps/Holes in the 3D Print walls

  • I'm new to 3D printing, I bought an Ender 5 Pro just recently. Trying to print with 1.75 mm PLA, but the results are very bad, unfortunately. When I got after bed leveling I got a decent result from the demo dog and started to print small things, where the quality was acceptable. Then I tried to switch to bigger items, but the printing failed half-way due to not enough adhesion.

    There were other quality problems too with this print, and you can already see that in the walls there are holes:

    enter image description here

    I did another round of bed leveling and Z alignment to make sure the adhesion is good and a test print came out quite fine in all corners and the center. The one layer rectangles and not perfect, in some places the lines separate, but they are mostly fine. But now I have a different problem: all the prints are very messy, less clear, they are not as strong as before and also there are big holes now in the walls.
    The same statue base (as I had to stop after getting messy earlier) looks like this now:

    enter image description here

    Another try failed after a few hours (like printing stopped and was printing nothing in the air), but also there is the holes/gaps problem even more visible:

    enter image description here

    Finally as a test I printed an object with the same G-code as previously used and the result is much different. The object on the left is the new one. It is weak, you can feel by pressing that the walls are not solid, they bend. The rectangle "eyes" are also not clear:

    enter image description here

    There is a difference though, I changed the extruder's nozzle between, the new one was also in the Ender package, it is also 0.4 mm as the original should have been, I changed to try with a new one.

    Because of the last test with the same gcode, different result, I think the problem might be in hardware adjustment as well, not only software. Anyone has an idea what I'm doing wrong?

    EDIT 1: after calibrating the extraction amount and reducing the print temperature from 200 C to 190 C, now I get the below result. The values used:

    • Bed temperature: 65 C initial, 60 C for the rest

    • Print temperature: 200 C initial, 190 later

    • Print speed: 80 mm/s

    • Wall speed: 40 mm/s

    • Retraction: 10 mm

    • Retraction speed: 80 mm

    • Wall thickness: 0.8mm

    • Layer height: 0.2 mm

    • Initial layer: 0.2 mm

    • Line width: 0.4 mm

    enter image description here

    EDIT 2: Based on the comments, some changes were made and here are the results.
    XYZ cube,

    • print speed reduced to 60 mm

    • layer height 0.12 mm

    • Z Seam Alignment is Sharpest Corner

    • Infill density 30 %

    • Retraction distance 8 mm

    • Retraction speed 40 mm

    It looks good, the layers are visible though, some ghosting right to X and Y.
    The sizes are not correct though: X = 20.07 mm, Y = 20.03 mm, Z = 19.84 mm
    enter image description here

    Are X/Y acceptable? What should I do with Z, increase steps/mm ?

    Finally here is the 3D benchy too, although looks mostly fine, there are some bumps in the walls and small strings in open areas. This was printed earlier and with different settings though:

    • Layer Width 0.2 mm

    • Print speed 80 mm

    • Retraction 10 mm

    • Retraction speed 80 mm

    • Z Seam Alignment is set to Random

    enter image description here

    EDIT 3 I tried now to print the PolyPearl, that has thin curving lines. The first try failed after 2 hours, a knob developed on the nozzle that ruined the print. This was printing with 190 C. I gave a new try and printed with 200 C, and amazingly it completed the job. See below the pictures, here are my settings for it (changes for Cura 4.6.0's default Super Quality):

    • Layer height: 0.08 mm

    • Initial layer height: 0.12 mm

    • Wall thickness: 1.2 mm

    • Z Seam Alignment: Sharpest Corner

    • Print speed: 60 mm/s

    • Infill acceleration: 1000 $mm/s^2$

    • Print acceleration: 300 $mm/s^2$ (default is 500)

    • Print jerk: 8 mm/s (default is 10)

    • Retraction distance: 8 mm

    • Retraction speed: 40 mm/s

    The model sticks well to the glass plate even without glue or hair spray, maybe a little too well. I see some problems though, not sure how normal they are:

    • outside area of first layer is not nice

    • there is some oozing, thing lines on the surface and between the columns

    • the top end of the tower is somewhat messy and there is a horizontal line attached to it (sure, can be removed easily)

    • the bottom is very smooth, I can see the glass' texture (the Creality glass top is textured) and the texture of the very first failed print, when I didn't take into account the extra height of the glass after leveling, and the print head hit it hard and the nozzle got completely damaged. Beginner's fault.

    Here are the images:

    Are this problems normal/acceptable?

    Hi, welcome to 3DPrinting.SE! It seems that you are facing more problems. Possible problems are: inconsistent STL file, filament flow oozed prior to Z seam start, temperature too high, travel too slow, slicing error with incorrect filament diameter, volumetric mode enabled for linear commands, clogging, incorrect retraction settings and maybe more. Basically, we could use the settings (temperatures, speeds, retraction, which material, etc.) you used to create the G-code files.

    hmmm, looks to me like... ringing or ghosting? to make it easier to analyze, please print a cube with 2 flat sides and 2 with indented X and Y. As a second test print, I suggest a cylinder and one with a cut in the outer shell.

    Unfortunately I was not clever enough to save the whole project for the tries, only the gcode, so getting the setting is harder. I read around and found the under extraction can be a problem. Indeed for some reason instead of 100m it extracted 95mm. After a lot of calibration now I'm close to 100m and generated a new gcode for the same slime (cube) model. The result now is close to the original one. The surface is not clear enough though. I will ammend the original question with the values used for this and a pic.

    ok, I see one parameter that might result in the pattern: "Print speed: 80 mm/s" - can you try slowing that one to 60? Also, the massive underextrusion in some intermediate prints might be a clogged nozzle - cleaning might also be in order.

    I would bring the temperature up, for better bonding and to make the PLA more fluid (easier to extrude) and reduce by 1-2% the extrusion multiplier (see that on the top layer there is too much material and the nozzle left a line on the top right?)

    Also, try linear advance to reduce blobs in the corners (where there is significant deceleration) or reduce acceleration overall by 500 mm/s^2. Printing time won't be affected much but you'll get better quality

    Thanks for the hints. I have printed the XYZ cube and the 3D benchy, I will upload those two. It is definitively much better, not sure if something can be improved still, or this is how it is. Details will be in the edit.

    The Z steps/mm must be ONLY calculated, not adjusted to compensate. Check your Z screws and that's the value, I think 400. If the cube is smaller, maybe the first layer is too thin (bed too high?) or something else. However, given the bottom surface of the Benchy, the bed is fine, since you can see the different extrusion tracks

    Measure the vertical size of the Benchy features. Are they correct or too short, like the cube?

    The top of the box is about right (15.48 mm instead of 15.50 mm), the whole height is a little smaller 47.88 mm vs 48 mm. In general the measured sizes are smaller with about max. 0.2 mm, but this was printed with a 0.2 mm layer height, if that matters.

    0.16 sounds like you lost one layer thickness... hey, is the first layer thickness 0.2 or 0.24? if it's the latter, it does *not* add the needed last layer.

    For the cube, the initial layer is 0.12 mm, for the benchy I used 0.2 mm.

    If vertical sizes are offset by a fixed amount, not "scaled", then clearly that's a proof that the Z steps/mm should NOT be changed. In fact, it never needs to be calibrated. Never.

    Ok, I did not change them, I changed only the extruder steps.

    try a simplified version of the cube: 0% infill, 0 top and bottom layers, 1 shell. All layers - including first 0.2 mm. That should be rather fast and should come in to less than 5/100th of a mm difference.

    BTW, I highly congratulate you on providing us with enough picture material and settings to properly troubleshoot with you!

    The cube with those settings (0.2 mm, 0& infill, no top/bottom layer, 1 shell) gave interesting results: X = 20.00 mm +/- 0.01 (0.02 is the caliper resolution), Y=19.90 mm, Z=19.85 mm, so Z/Y got smaller, Z is the same as with the finer settings. Thanks for all your help, and I hope all this discussion and my post is useful for other beginners too.

    Forgot to add: for X there is no ghosting with these settings, the only strange thing is a line at the Y/opposite X corner, with the height of the Y character, you can even feel it is a bumb along the edge.

  • Trish

    Trish Correct answer

    2 years ago

    Let's see what kind of errors I see (accumulated from what I commented):


    The legs seem rather underextruded. This could be a clogged nozzle as much as too fast printing or too low a temperature as well as too little pressure (e.g. miscalibrated extruder). Try a cleaning or a fresh nozzle, other playing with the parameter can wait after you fixed the next big issue.

    To fix the under extrusion if it persists after fixing Ringing, experiment with the following settings:

    • Check the mechanical system. Make sure that the extruder

      • has no defect. I had seen a lot of under extrusion when my extruder lever broke. The gear turned but did no longer push the filament.

      • does press the filament well with the idler bearing into the hobbled gear.

      • is properly calibrated according to the physical parameters. Use the Extrusion multiplier to account for different materials.

    • Temperature. If extrusion comes bad, increasing it in 5 °C steps can help.


    The cube with the indentations shows it best: There is a massive ringing happening. This is an effect that happens when the printer changes directions fast - the printhead can't accelerate and decelerate with infinity, as it has momentum, and as a result oscillates around the new path a little, like a sinus ring. This creates a "ghost" of the preceding path changes on a flat face, which is why I asked to print a cube.

    To fix Ghosting and Ringing, there are 2 software fixes:

    • reduce print speed.

      • this comes at the cost of longer print times

    • work out better maximum acceleration and jerk

      • Reducing the maximum acceleration some (down to between 500 mm/s² to 1000 mm/s²) can massively better the quality at little to no effect on total print time, and keep your print speed up.

      • Jerk is the derivate of acceleration; finding out the two values that might be best for your printer setup can be a lot of experimentation.

    There is also a hardware design way to reduce it in many designs by shaving off weight from moving parts or stiffening up the construction:

    • Get a lighter printhead

      • remove unnecessary addons

      • swap for a lighter printhead design

      • changing to a Bowden design

      • in CoreXY: use lighter rails

    • Stiffen the construction

      • Adjust the mounting of the carriage in such a way that it is better constricted

      • change out bearings and bushings for ones with tighter tolerances

      • bracings can massively stiffen a portal and cube design

    Height inaccuracy

    This could stem from any of several sources. In descending order of likeness:

    • the first layer is not a multiple of the actual layers (e.g. 0.12 instead of 0.2 mm) leading to a partial layer at the top being cut. in the example this would result in a fixed negative error about 0.08 mm.

    • a miscalibrated first layer. Depending on your leveling skills, this can account for a fixed error of up to 0.1 mm for a bad calibration to 0.025 mm for a really good calibration. This error can appear both positive and negative.

    • Z-Axis inconsistencies from a misaligned leadscrew, a sticky leadscrew, or insufficient Z-stepper voltage. Both would usually create a systematic error - as in printing a given percentage loss of height. Such inconsistencies can be fixed by solving the mentioned issue: check orthogonality first, then gently oil the screw with light machine oil (not WD40!), finally increase motor voltage slightly. Only if your results are massively out of whack, you should check if your steps/mm are set correctly. But I don't see the need in this case.

License under CC-BY-SA with attribution

Content dated before 7/24/2021 11:53 AM