How to ensure all dimensions of an object

Hi everybody!
I am having difficulties in ensuring the dimensions of objects in my Graber.
I thought that it was only a matter of steps by mm, but it is not just external dimensions.
At the end, inner diameters and distances between holes dont fit.
Is there any kind of checking list or something?
Thank you all!

Steps/mm are for physical moves. these should be calculated on machines hardware.

When parts when sliced the plastic is put the inside the boundaries of the object. This requires that the extruded filament is exactly the diameter expected or it becomes to large or two small.
This is based on E steps/mm and the filament diameter and the mechanism used to pull in filament and nozzle diameter. So E steps/mm is a starting point but you need to tune it for reality.

Distance between holes is a concern, as that should be accurate if X and Y steps/mm is accurate.
In saying that all plastics shrink when cooling.

PLA: 0.3-0.5%
ABS: 0.5-1.5%
PETG: 0.2-1.0%

So if your part is larger, the more it shrinks and positions can move.

Edited 2 time(s). Last edit at 08/09/2025 03:16AM by Dust.

Holes come out small because the plastic shrinks as it cools, and that pulls the plastic into the open hole. You can either use an offset- try 0.2-0.3mm larger diameter in your CAD file, or use holes smaller than needed and drill them out when the print is finished.

On any given axis, dimensions and hole spacing should be right if the steps/mm is right. But if the axes aren't square, you'll have all sorts of problems getting printed parts to fit together or to fit with non printed parts. If the axes aren't square, you won't be able to print threaded parts that screw together because round objects won't be circular. You also won't be able to print boxes with tightly fitting lids. Test squareness of the axes by printing the largest cube (NOT a 20mm "calibration cube"- they are too small to be of any value) you can measure accurately. Measure the diagonals in each plane. If the axes are square, the diagonals will all measure the same. If not you have to figure out how to adjust the positions of the linear guides in the printer to square them up. Some printers may have some software compensation for out of square axes, but I recommend getting the hardware set up correctly.

---

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

Thank you!
I will print a large cube as you suggested.
But now my problem is that dimensions at X and Y axis are ok, but not in Z.
Anyway, I will print the cube.

Hi!
The diagonals are equal.
But I still cannot reach good precision on the z axis.
I have tried to lubricate the fuses, improved the microsteps, bed levelling and so on.
In 20mm I could only reach 19.62mm!

Hi everybody!
Trying to understand if there were a problem of steps/mm in calibration of the z axis of my 3D printer, I printed this ladder in the pdf file attached.
Notice that the steps are equally spaced in height by 10mm each.
After printing, the dimensions of height of each step are:
1st = 9.80mm (should be 10mm), so, if I should calibrate steps by mm the correction should be 1.020
2nd = 19.68mm (should be 20mm), so, if I should calibrate steps by mm the correction should be 1.016
3rd = 29.60mm (should be 30mm), so, if I should calibrate steps by mm the correction should be 1.014
4th = 39.56mm (should be 40mm), so, if I should calibrate steps by mm the correction should be 1.011
5th = 49.50mm (should be 50mm), so, if I should calibrate steps by mm the correction should be 1.010
6th = 59.48mm (should be 60mm), so, if I should calibrate steps by mm the correction should be 1.009
7th = 69.44mm (should be 70mm), so, if I should calibrate steps by mm the correction should be 1.008
8th = 79.40mm (should be 80mm), so, if I should calibrate steps by mm the correction should be 1.008

Analyzing these results I believe there is not a problem of steps/mm, but I have verified lots of possible problems, and I could not find the problem:
I have checked lubrication, aligment of the both trapezoidal spindles, microsteps, and so on.
I have attached a picture of the printer, also.
Does anybody have an idea of what can be?

... set a box above/around the printer and let the air inside get on averaged 30-40°C temp (by heating the bed to 100°C and waiting some ten minutes) -- then your results should be more consistent ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I confess that I am lazy to install an anti backlash nut in z axis.
Do you think it is possible to correct these results with an anti backlash nut?

... could help -- but too, "it depends", as I'm not knowing all the other issues, you're tapping into

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

How thick were the layers you printed? If you're trying to print 10mm high intervals using layers that are 0.3mm thick, it isn't going to work. Be sure that you use layer heights that can give you the 10mm intervals you're trying to measure. Likewise, when you design things to be printed, you'll get best results if you design the objects for the layer height you intend to print. If you're using 0.2 mm layers, don't put features at x.5 mm.

Also, you usually want to print in layers thickness that is a whole number multiple of the displacement per full step of the motor. You will get most accurate/precise results when using a full step multiple. For example, if each full step of the motor(s) gives 0.035mm of movement, don't try to print in 200 um layers. Z axis calibration really should be a simple function of the lead of the screws, steps/rev of the motors and the microstepping used.

Are you using Z hop on layer change? If so, antibacklash nuts might help with precision. If not it shouldn't matter. Turn off z-hop and test again. Normally, if the mass lifted on the Z axis is sufficient, and the guide rails and lead screws are properly aligned, you shouldn't need antibacklash nuts on the Z axis.

Like all calibrations, using small things to calibrate is going to introduce errors. Every measurement you make has some error because of the way you make the measurement and interpret the measurement. If you use a caliper, did you hold it perfectly straight? If you used a ruler, did you interpolate between markings accurately? Lets assume you use a 20 mm "calibration" cube and you make a 0.1 mm error when you measure its Z axis. When you're measuring a 3D print, a little over extrusion can result in a rough top surface that will likely introduce some error in the size of the print. 0.1mm error is very easy to make. Now you bake that error into your printer by setting steps/mm based on the size +/- the error you introduced. Now you print an object that is supposed to be 200 mm high. You will then find it is actually off by 1 mm because your 0.1 mm error multiplied by the height of the object relative to the calibration object. Yes, the % error stays the same, but the magnitude is now 1 mm.

Always calibrate with large prints or movements. When you calibrate the Z axis, home the z-axis and put an accurate ruler on one of the Z axis verticals in the printer with the "0" of the ruler at the home position. Now command the Z axis to move up 200mm or however high it can go without hitting the mechanical end of the axis. Now read the position from the ruler. If you told it to go 200 mm and it went 195 mm, your steps/mm is too small. Correct it by multiplying whatever steps/mm value you have set by 200/195=1.026. If it reads 207mm, multiply current setting by 200/207=.966. If you made the same 0.1mm error reading the measurement when you calibrated using the 200mm movement, when you print something that is 20 mm high the error will be divided down to 0.01 mm.

Is the bed temperature controlled using bang-bang or PID control? It should be PID control.

Finally, in addition to rough top surfaces due to over extrusion, you're printing hot plastic that exhibits die swell and shrinks as it cools. There is a limit to the accuracy you can expect to achieve. Sometimes, 3D printing just isn't accurate enough and you have to use some other process/material to make the object.

Edited 1 time(s). Last edit at 09/12/2025 12:15PM by the_digital_dentist.

---

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