Printquality, backlash

Hi
So I noticed a dropp in the printquality of my parts, it turned out that it was caused by several factors, one of them was that one of the z-axis bearings got broken. After fixing that, I can still see some irregularities in my prints, which seem to be caused by backlash. I tried to measure it, and it looks like that the position of the nozzle depends on the direction it moves from. If I move the nozzle to position for example X=100 from the left and from the right, the difference in the possition seem to be 0.1-0.2 mm. I checked the alignment, I have both tried to loosen and tighten the belts, I checked the pulleys, lubricated everything, I have checked the motors, and nothing seem to help. I heard that the 2.5T belts are not great for lineary motion and may cause backlash so I have plans about changing them to GT2 belts and pulleys. But I was wondering how big backlash you guys have on your printers? Is this normal (I dont think so because the printquality seemed to be better)?

Hello Tytonjara,

please can you explain how you did the measurements on the backslash. Furthermore it would be interesting to know how your prints look like with the mentioned backslash, and for commenting a picture would be helpful.

Hi
So I measured the backlash using a caliper

Here are the prints:

here both parts are printed with random seam position, but the part on the left was printed alone, while the one on the right with other parts, that caused that the nozzle sometimes started the layer coming from the “outside” of the wall before starting printing, and that is where the layer sticks out, notice how the same layer does not stick out just before it comes back to the seam, that it because the nozzle then comes from the “inner” side of the wall.

Here is a picture of an other part from under a microscope:


where you see the layer which sticks out in the red ring, from overneath in magnification. This thing should not stick out at all.

Just wondering.
What kind of filament are you using?
PLA or ABS?
Have you checked your retraction settings and speed in Cura?
Retraction = 4mm Speed = 110mm.
PLA prints best at 190C
ABS = 220C

This particular print is ABS, PLA shows no difference, I print this particular ABS at 223 degrees, and all the settings were exactly the same for both prints, my retraction is 5 mm and speed 110. So I would like to know what backlash is typical for this machine, is mine an exception, or is this normal for the k8400?

Dear Tytonjara,

you are taking these issues really serious, so I will try my best to give you some useful comments.

First of all I’m not convinced from your measurement approach, I don’t believe an accuracy of 0.1 to 0.2 mm in this configuration.

However, the quality of the prints is the main topic. Let us assume that we have no mechanical issues, then we must discuss about the fine tuning of your print settings.
On your left picture you can see a lot of little holes. That is one effect which is quite unusual for me. Here it seems that material is missing. From my experience up to now this is in indicator of slipping in the extruder (feed rate too high or extrusion speed too high). This will cause local variations in the print width, which can be interpreted as backslash. For such extreme holes I would assume a very hot extrusion temperature.

The right picture them confirms this impression. When the printhead is moving from one part to another, the pressure in the nozzle is removed but the filament is heated up above your setting. this makes it very liquid, and after the restart of the extrusion much more material will be printed for a short time. This can cause this impression of backslash once again.

So my first impression is, that the printouts are not that bad, but some optimization on feed rate and/or extrusion speed as well as the temperature can improve. Especially the fan control can have a significant impact on the quality of the prints, because it will significantly disturb the thermal equilibrium at the hotend if it is changed during printout. The temperature control algorithm in the firmware is not capable to handle this disturbance properly without large variations of the hotend temperature.

Since I was long time in doubt about the proper results on my printer too, I summarized my results and findings here: http://ddd.k8400.hoehnemann.net/. I invite you to share my findings and give some feedback based on your conclusions.

Hi hoh61

This were just testprints to ilustrate the problem, and the problem becomes more annoying when printing detailed part with for example vertical columns.

The little dots on the left part are the seams, it is just where the nozzle starts to extrude, I chose those to be at random positions, but I could align them on an edge anth then there would be no dots visible on the part, i’ve already testes that. The material is not missing there I can asure you about that, I can even post som pictures under magnification if you are interested. I can also assure you that the nozzle temperature is not to high, I’ve done same tests on lower temperatures, and also with lower feedrates, lower speed, and so on.

An other argument for that it is backlash is that the problem appears only when the extruder moves from the outside of the wall, I will try to design a test part to convince you about that. The problem is not that the extruder retracts, and the heat builds up, because when that happends, and the movement of the nozzle is from the inside of the wall, the problem is not visible, it is visible only on a movement from the outside if the wall.

What software and version are you using?
Since you are using PLA have you tried 190C

I use slic3r 1.3.0
I have also updated printer to the current version.
Yes, 190C is my default temp for PLA, and I tune it individually from there for each PLA plastic.

Have you tried the software that was customized for the K8400.
http://www.vellemanprojects.eu/downloads/files/vertex/software/vertex-3d-printer-repetier-host-1-0-6.zip
This is a step back from what you are using but my prints look very good.
You may need to go in and change the printing temp retraction and speed.
But it is something to try
This one also uses Cura.
Cura is faster when it comes to slicing.

Do you use the option “External perimeters first” ? It would be interesting if a switch of this option will cause a difference in the appearance of the sidewall.