Creality Ender 3 Pro: Problem with Z-axis inaccuracy (squashed layers?)

Question

Creality Ender 3 Pro – Z-axis inaccuracy problem.

Before opening this new thread, I did read this question (Perhaps there is another question on this site I did not see!) Browsing through comments on that question, I don't really see any clear, verified solution for errors in the Z dimension when printing.

I recently bought a Creality Ender 3 Pro (migrating from a Kingroon KP3, to which I had migrated from a Printrbot Simple Metal).

Today I did a print of a small "testing" object which has a series of rectangular walls (1.5, 2, and 3 mm) running along both X and Y axis which are exactly 10 mm tall (in the STL file). I printed this with both the Ender and my older Kingroon. Kingroon had slight inaccuracy with the 2 mm wall (came out 2.2), but all other dimensions were correct. The Ender had perfect widths for all walls. But the accuracy of the wall height was TERRIBLE- instead of 10 mm, it was only 9.4!

I should mention two other observations:

1. Very often I print with rafts to avoid the "elephant foot" dimension problem. Each time I specify a raft with the Ender, the raft is more or less impossible to snap off!
2. During the print there are occasional "clicks" at the extruder feed gear (more or less throughout the print process). The only time I encountered that type of symptom previously was with the Kingroon on the very first layer if the height was not quite zeroed correctly and so it was trying to print too close to the bed (The whole leveling/height adjustment on the Kingroon was a constant challenge and needed to be redone each day).

In any case, these two symptoms could also point to improper Z movement during the print job. I should note that to create the G-code file for the Ender I used Cura with the default settings for the Creality Ender 3 Pro machine. Also, during setup I did follow the instructions on a video to "square" the frame.

That earlier thread I mentioned emphasized problems/inaccuracy was in the first few layers.

But in my case the "wall height of 10 mm" (which came out as 9.4) is not the overall height of the object from the bottom, but rather a measurement from the top of the floor of the object to the top of the walls. So this if well above the first few layers.

I have not yet tried monkeying with the eccentric "tightness" adjusters to the wheels on the two sides (I am a bit cautious about those adjustments because there does not seem to be a way to measure/quantify changes made or even be sure you return to the original state. Hence my use of the term "monkeying").

Also, unlike checking movement on X and Y, I cannot simply raise and lower the Z gantry manually to check for smoothness and freedom of motion).

Answer 1

Well, lacking any other suggestions, I took a deep breath and "monkeyed" with the blind eccentric adjusters for the wheeled assemblies for z-movement.
I say this as plural since there is one on the left side (power side) and one on the right side.
As I feared, because you cannot freely slide these, it is really difficult to tell if you are tightening or loosening the "movement".
Based on the potential cause as I identified in the original question,my goal was to slightly loosen them up.
On a positive note, I was able to measure the symptom by using a caliper to check the height of the horizontal gantry above the bed when "homed"(0) then up 10, 20, 30 etc in a non-printing environment. Indeed the height measurements verified that the progression of height did not correspond with the requested/reported movement as shown on the screen.
The variation was much worse on the (slave) right side... Counting on nothing more than dumb luck as I tried adjusting the two sides, I ended up with acceptable performance.
So for now I am done screwing around with it.
If anyone has a great suggestion as to how to better know whether the wheels are getting tighter or looser when making this adjustment, it could help anyone needing to make the adjustment.