Prints not being exact (enough)

Hello everyone,

i use my 3d printer for quite some time now (1.5 years), but i still lack some knowledge which i am not sure where and how to acquire.

Printer: selfmade prusa i3 with parts bought from ebay

My current problem is the following: I can't print stuff in an exact way (without underextruding).
I can print stuff that perfectly matches the stl i am feeding to my slicer, but the only way to get to that point is to lower the extrusion rate which makes my printer underextruding (well what else should it do).
Known facts for me:
- Perimeter walls are not touching? Underextrusion
- Size of ouput print does not match: Under-/Overextrusion

Anything wrong until now?
I used a guide to determine the correct steps per mm for my extruder (and used prusas calculator for x/y axis which match nearly perfect).
Steps i had done:
- Print a cube 20x20x20
- measure the cube
- fix steps per mm according the the measuring results
- repeat from step 1 untul measuring is correct

Following these steps i get underextrusion (because perimeter walls are not touching i guess this, wrong here?).

With the following steps i got to a point where i haven't had underextrusion anymore and prints looked good, but where the printer got not exact enough to for example print a part with a hole for an m8 nut to make it fit:
- Print a cube 20x20x20 without top
- while walls are not touching:
-- increase steps per mm
-- reprint

What am i missing or what am i doing wrong here?
Anything i can try?
Any help is appreciated!

Best regards from germany/hamburg,
s_s_

---

3D Drucker:
Prusa i3 (Alu "Rahmen", Heated Bed, Bowden Extruder, E3D-V6 Hotend (Klon), RAMPS 1.4, A4988 (extruder)/DRV8825 (x/y/z) Stepper Treiber, Marlin 1.1.0 RC2, Reworked Plastikteile)

ok, first you need to calibrate again your extruder so when you extrude 10 mm the filament on the infeed side of the extruder moves 10 mm, repeat that with 100 mm. adjust your steps per mm (using the biggest difference number) and repeat the process until you get exact results. Once satisfied you need to print single line extrusions in your built plate, with each line you are going to test your extrusion width and your extrusion multiplier in order to get the best extrusion line possible so:

if your nozzle is for example a 0.4 mm you will set one line with an extrusion width of 0.4 (manual) and one with auto (0.48), for each setting you will print several ones with different extrusion multipliers from 0.9 to 1.2 (although I never go higher than 1) in increments of 2 (0.90, 0.92......) repeat this process for every layer high you use (0.2, 0.3, 0.4, etc.) once you have collected all your extrusions measure all of them and select only the ones that will measure exact, from there you will have for each layer height the best extrusion width and extrusion multiplier and set those numbers in your slicer. Each material is different so you are better of set this test for each material you use and for each supplier you use (is a pain but worth the time).

Now few things you didn't mention and may cause your problems are: do you measure your filament in several places and set the filament width in the slicer? when you set your wall thickness are you calculating how many lines can actually fit in there? for a 0.4 nozzle with an auto width of 0.48 for 2 perimeters your wall needs to be set at 0.96, 3 at 1.44 and 4 at 1.92, anything different than that and you will have spaces in the perimeters forming the wall.

One last thing, circles will never be exact so you need to compensate for it, once you have your extrusion perfect you can print several test cylinders to find which is the best number for you, for me a 8 mm hole needs to be scaled to 8.2 so an M8 rod slide freely in the hole.

Getting accurate results requires accurate printer construction and good calibration. For a printer construction example, if the belts that move the extruder carriage on the X axis are not kept parallel to the X axis at all positions of the extruder carriage, prints will not be accurate, and the error will depend on where they are printed along the X axis. The closer they are to the ends of the X axis, the worse the distortion will be. Even if you go to all the trouble to calibrate your printer, you are handicapped by its poor design. The Z axis screws, driven by two motors, can easily lose sync, tilting the X axis relative to the Z axis, causing prints to be skewed. Your hard work of calibrating won't last long and you'll be doing it again very soon due to this instability. If you want the calibration to last, drive the Z axis screws with a belt from a single motor.

Use a larger object to calibrate, preferably one as large as the bed will allow. Whenever you make a measurement, there will be some error. If you print a small object, measure it, then use that measurement to set steps/mm, the error in the measurement is multiplied when you print larger objects. If you calibrate with a larger object, when you print smaller objects the error gets divided.

Make sure you calibrate the extruder. I like this guide for getting good results: [reprap.org] Pay careful attention to the stuff about fine tuning the extruder steps/mm.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
ggherbaz
ok, first you need to calibrate again your extruder so when you extrude 10 mm the filament on the infeed side of the extruder moves 10 mm, repeat that with 100 mm. adjust your steps per mm (using the biggest difference number) and repeat the process until you get exact results. Once satisfied you need to print single line extrusions in your built plate, with each line you are going to test your extrusion width and your extrusion multiplier in order to get the best extrusion line possible so:

if your nozzle is for example a 0.4 mm you will set one line with an extrusion width of 0.4 (manual) and one with auto (0.48), for each setting you will print several ones with different extrusion multipliers from 0.9 to 1.2 (although I never go higher than 1) in increments of 2 (0.90, 0.92......) repeat this process for every layer high you use (0.2, 0.3, 0.4, etc.) once you have collected all your extrusions measure all of them and select only the ones that will measure exact, from there you will have for each layer height the best extrusion width and extrusion multiplier and set those numbers in your slicer. Each material is different so you are better of set this test for each material you use and for each supplier you use (is a pain but worth the time).

Now few things you didn't mention and may cause your problems are: do you measure your filament in several places and set the filament width in the slicer? when you set your wall thickness are you calculating how many lines can actually fit in there? for a 0.4 nozzle with an auto width of 0.48 for 2 perimeters your wall needs to be set at 0.96, 3 at 1.44 and 4 at 1.92, anything different than that and you will have spaces in the perimeters forming the wall.

One last thing, circles will never be exact so you need to compensate for it, once you have your extrusion perfect you can print several test cylinders to find which is the best number for you, for me a 8 mm hole needs to be scaled to 8.2 so an M8 rod slide freely in the hole.

Thanks a lot for this, i will try that, though i am not sure yet whether that will fix the problem.
I usually just set how many perimeters i want, not how big (0,48;0,92;etc.) they should be.
In the beginning i used slic3r, but recently i tried simplify3d which is very nice to use - but doesn't have a separated Filament Setting section.
I personally think that one should not have to scale holes because the design tells us how big the hole should be. The Slicer should be made able to give you setting for this (since they are able to fix quite a lot printing issues like retracting, ooze shield etc.).

Quote
the_digital_dentist
Getting accurate results requires accurate printer construction and good calibration. For a printer construction example, if the belts that move the extruder carriage on the X axis are not kept parallel to the X axis at all positions of the extruder carriage, prints will not be accurate, and the error will depend on where they are printed along the X axis. The closer they are to the ends of the X axis, the worse the distortion will be. Even if you go to all the trouble to calibrate your printer, you are handicapped by its poor design. The Z axis screws, driven by two motors, can easily lose sync, tilting the X axis relative to the Z axis, causing prints to be skewed. Your hard work of calibrating won't last long and you'll be doing it again very soon due to this instability. If you want the calibration to last, drive the Z axis screws with a belt from a single motor.

Use a larger object to calibrate, preferably one as large as the bed will allow. Whenever you make a measurement, there will be some error. If you print a small object, measure it, then use that measurement to set steps/mm, the error in the measurement is multiplied when you print larger objects. If you calibrate with a larger object, when you print smaller objects the error gets divided.

Make sure you calibrate the extruder. I like this guide for getting good results: [reprap.org] Pay careful attention to the stuff about fine tuning the extruder steps/mm.

I haven't had a real problem with my z-axis screws yet. They are tight and only in the beginning it had a little problem, but not anymore.
I don't understand how 1 motor using a belt driving the z-axis could be better than 2 motors that do exactly the same, can you explain?
For how i understand it: The belt can always have the same problem (turn over) as loose screws on the z-axis motor coupling.

---

3D Drucker:
Prusa i3 (Alu "Rahmen", Heated Bed, Bowden Extruder, E3D-V6 Hotend (Klon), RAMPS 1.4, A4988 (extruder)/DRV8825 (x/y/z) Stepper Treiber, Marlin 1.1.0 RC2, Reworked Plastikteile)

When you use microstepping to drive the Z axis motors, every time you shut off power, the motors will try to snap to their closest full step detents. One motor may go one direction, the other motor may go the other direction. Any electrical problem, or any binding of the mechanism on one side or the other can lead to skipped steps and tilting of the X axis. A loose screw on a motor or screw shaft can also cause the X axis to tilt. A curious child or wrong-headed operator (thinking they are releveling the bed by tilting the X axis to match it) can manually twist one of the screws with their fingers and throw the alignment off.

Some dual motor Z axis people don't notice that the X axis is slowly tilting over time and they keep releveling their print beds to match the X axis position (I've seen this at the makerspace), and if you use autoleveling, that is exactly what it will do. Eventually they show up here or on G+ and ask why their prints are coming out skewed. People and autoleveling both assume that the axes are orthogonal and remain that way forever, even though most people have never bothered to check orthogonality or even know what it means. With dual motor Z axis, you can't guarantee orthogonality for any number of printing cycles.

When using a single motor with a belt, the screws are locked in sync at all times. The only way they can get out of sync is a loose screw on one of the pulleys which is easy enough to prevent with a drop of loctite. You can't change the XZ alignment without using tools. I built my printer using a belt drive on the Z axis screws from the first day. After aligning the X and Z axes I literally never had to touch the alignment again, ever.

Using a single motor also has a very useful advantage over a dual motor set-up. If I want to raise lower the X axis to work on the extruder, or clean the nozzle off, zero the Z axis, etc., you can do it with the power on or off just by pulling on the belt. In a dual motor set-up, you must power on the machine and maybe a computer, step through multiple menus, jog the Z axis, and wait for it to stop. It doesn't sound like much, but I've worked both ways and believe me, you can't beat the convenience of a single motor Z axis.

If you want to check for orthogonal alignment of your printer's axes, print a cube (a large one) and measure the diagonals. If the diagonals are equal length, the axes are square. I developed a calibration cube and companion spreadsheet for the purpose of aligning axes. You can DL them here: [www.youmagine.com]

You print the cube, measure the sides and diagonals in each plane, and enter the measurements into the spreadsheet. It will tell you the angular error and how to correct it. You can use it before or after calibrating your extruder, though you might want to calibrate the extruder first, and you can use it to calibrate the steps/mm for each axis.

With a dual motor Z axis, don't count on the XZ relationship to last very long...

The printer/slicer can't make holes the correct size because it doesn't know the correct size. When you convert your CAD model to an STL file, all curved surfaces become polygonal approximations of the curves (and if you use SketchUP, even your CAD model is full of polygonal approximations of curves). When you're laying down molten plastic, it shrinks as it cools. Think about the forces on the cooling plastic when you're printing a circular hole perimeter. As the plastic shrinks, it pulls itself toward the center of the hole, where there is no plastic to support/prevent it from going very far. It tries to become a smaller hole. The effect will vary with the settings you use when you convert to STL (you can control the number of segments in a circular approximation), the plastic, and the temperatures involved, and probably with print speed, too.

Edited 1 time(s). Last edit at 10/16/2016 08:40AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

I made a modification to my printer's X axis yesterday. I had to take the entire X axis assembly off the machine, which is pretty easy because it is attached to the Z axis screws with just 4 screws and all the wires go to a screw terminal block. After I made the modification, I put the X axis back on the machine and checked it- I had to adjust the Z=0 position, but the alignment didn't change at all (due to Z axis screws being locked in sync by the belt).

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

@the_digital_dentist: Thanks a lot for your explanations, that makes stuff more clear to be.
I have seen your machine (from your signature) and it's very interesting, also made clear how stuff then has to look like if one wants to set it up with one z motor.
Should you ever come to hamburg, germany come and grab yourself some beer with me Feel invited

EDIT: Since following Triffid Hunters Guide the printed part with the M8 Nut fit much better (event though tight), while the part that should be fitting for an LM8UU still doesn't fit (but is much closer to what it should be like).
I have now remodeled the not fitting part to be able to un-/tighten the LM8UU holder.

Edited 1 time(s). Last edit at 10/23/2016 05:22AM by s_s_.

---

3D Drucker:
Prusa i3 (Alu "Rahmen", Heated Bed, Bowden Extruder, E3D-V6 Hotend (Klon), RAMPS 1.4, A4988 (extruder)/DRV8825 (x/y/z) Stepper Treiber, Marlin 1.1.0 RC2, Reworked Plastikteile)

Thanks, I will look you up the next time I am in Germany. Hacker-Pschorr FTW!

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]