Improving Print Quality

Not sure if my expectations are inline or not with what the SM2 should be capable of, advice appreciated:

Layer Height: 0.12mm
Default Print Speed: 50 mm/s
Retraction: 5mm @ 60 mm/s

PLA 210, 60

STL: 3D-Voronoi with openScad is possible

Up close, looks like maybe it was a little too hot? Perhaps go slower to improve stringing?

Hey, my thoughts are the following:

  • Have you tuned the print temperature well to your filament?- Overall it looks not so bad.
  • Print speed should be less.- I print the most things with about 40mm-s or less
  • Try another layer height.- I had similar problems with a few filaments at low layer heights.- maybe you try 0.2mm layer

Overall i think you should print such geometries at lower speeds. You can see your layers are not very accurate and very strong varying (depending on 0.12mm layer) on the pictures.

Thanks for the advice, have made a few changes, temperature I’ve dialled down to 195, there wasn’t much in it vs 210. Thought I would give it a try:

Dialled down the retraction, I’m sure Luban is set to 5mm retraction with 60mm/s, so I used this in S3D However, this looks terrible, set to 1mm retraction now.

Have turned down print speed to 40 mm/s.

Once I’ve finished with Dimensional accuracy on holes

Will try the tealight print again.

1 Like

Be sure to calibrate your E-steps as well to ensure you arent over/under extruding.

This is a great guide for it:

For small objects like your tea-light holder, printing slow enough to allow for adequate cooling is very important, I would suggest either printing two at once or reducing your speeds to <40mm/s as the previous poster suggested.

Another thing worth experimenting with is the number of shells, generally more shells takes longer but can improve overhangs as it supports the outer wall better.


200 degrees is too hot. 195 or 200 degrees is OK for printing. And for the model you printed, a lower speed would make the surface look better.

About the retraction, I did not use the S3D and can not tell the settings. But I use Snapmaker Luban and use the default settings, 5mm, and it looks good.

For fast testing, I set the speed at 60mm/s so the surface looks not good enough. :wink:

1 Like

@C.Harris That guide was good, I had already dialled in the extrusion itself. However, good to check it’s still accurate, which it was.

The calibration cube in vase mode prints out at about 6mm/s. Have to say the layers look so good at this speed. Layer Height: 0.16

This really helped dial in the dimensions. However, I’m left with one frustration as all of the sides are very close after calibration, apart from one. I have 0.46,0.47,0.48,0.53?. Bit frustrating, I can perhaps go next level with the calibration cross and fiddle with the axis steppers, just more time consuming and not sure it’s worth it, unless others have or would recommend another guide for this?

First I would like to preface this with the fact that I too see small variations in wall thicknesses (~0.05mm). Also you mention values around 0.48 but your nozzle is 0.4mm so you may still wish to reduce flow rate.

Changing your settings for your axis steppers wont have an effect on the width of the line extruded, this is purely down to how much material is being pushed out by the extruder and how consistent that volume of material is over time.

the inconsistency you noted could be due to a number of factors:

  • Walls not being laid perfectly on top of each other could easily lead to a 0.05 mm disparity as you are not measuring a single layer but across the full height

  • Diameter/roundness/general consistency changes along the filament you are printing with could also cause more/less material to be extruded at different layers or even within the same layer

  • Small inconsistencies in the shape of your extruder gear teeth

I feel that the variations you are seeing are completely normal and you would be hard pressed to reduce them much further.

1 Like

Did you tidy this up? I don’t see any stringing :+1:

Hey, after you have calibrated your E-Steps you are gone right to calibrate your flow with this little
“just outline in vase mode cube”
Now you have to setup your flow to get the walls to the size you want it to.

I have read you are using simplify3d and i don´t know if you have set the preferences of your process right…

Tab “Extruder”
Nozzle Diameter is 0.4mm right?
Set the extrusion multiplier to 100% for your first print (this will be adjusted, this is flow in s3d)
Set the extrusion width to manual at 0.45mm (its your choice, bigger will also work but less than 0.4mm would not work good enough)

print it slow in vase mode and be sure not to messure the wall with/including the bottom layer.

Now if you messure 0.48/0.5/0.47/0.52 or anything else, adjust the flow by a few percent and print again till your print matches your wishes.

Hope this helps!

(here is another youtube guide:

So in S3D the diameter is auto calculated to 0.48, I switched that to Manual and set to 0.48.
I did the first vase and reading were all over 0.50 so then I adjusted as per instructions so it’s much closer to 0.48 apart from 1 wall, which is a little higher.

I don’t think I can adjust again as the other 3 will probably get smaller to try and fix the outlier.

I did not tidy it. I took the photo when the printing was finished. There are few stringing actually. But I think it is good enough.

1 Like

It’s interesting, out of the box the A350 was under extruding by 10% which I corrected. Interestingly the dimensional accuracy suggests that I need to lower my extrusion rate by 10%, which does help with accuracy. However, it leaves me with gaps in the top layer again… I’m going to up the rate and see if I can find a happy balance.

1 Like

Gaps in your layers between walls and infill may be due to linear advance. Linear advance or K value is set at marlin default of 0.22, however myself and a number of other users have tested and found a k value of 0.08-0.1 to be much better.

I copy below a message I sent to another forum member on how to calibrate for K value:

"As of this moment it is a slow process to calibrate K value but possible if you have the time to do it. I used the gcode generator at this link to make a test file

Unfortunately as the SM2.0 currently doesnt respond to M900 Kx.xx commands, you cannot use the gcode generator above to automatically test multiple K values at once. However, it is still useful as a test print. Because of this, it does not matter what k values you set using the generator.

Once you have the g-code generated, try printing it. This will be done with K0.22 as this is the SM2.0 default. You can evaluate the print performance with reference to the visual guide at this link

Once you have determined whether you need to increase or decrease your K value from the initial test you will need to change the setting on your printer before running the test again. To do this, connect to the printer over usb. Once connected, type into the command line M503. This brings up the printer saved settings and you should see one of the lines says M900 K0.22.

To change your K value type into the command line M900 K0.xx (I suggest jumping right to K0.08 as this is what I found was the best setting). Type M500 to save this value. Restart your printer by power cycling and re-connect over USB and type M503 again to double check your new K value has been set. Now re-print the g-code and re-evaluate.

Rinse and repeat until you are satisfied that you are printing consistent line width!

Hope this helps."


Circling back on this, finally got round to playing with k-factor, it made a big difference to final output.

I had issues similar to yours. I found that the hot end was not holding a consistent temp. I added thermal paste to the thermister and that made a huge improvement.

Happy printing,

1 Like

Is the thermal paste an adhesive? The thermister can be quite loose, once you shove it back into the print head hard to say if it is lodged in properly or not.

if you buy a thicker thermal paste it will help secure its location (as it hardens a little bit as it dries, more like mud than a traditional adhesive). the fact that it seats so loosely is why i recommend thermal paste. it will help with both keeping the thermister in place, as well as improve the thermal conductivity for less temperature fluctuations. i recomend Artic Silver brand (available on amazon) but really any anti-oxidant compound that states “improved conductivity” will work. i used a brand called “ox-gard” (also available on amazon) that i got as a big tube from my local hardware store. but that is much harder to apply to the tiny hole then the nice syringes.

Sorry for hijacking this thread, I printed the gcode from the k-calculator today. I manually changed all “K” to “k” within the annotations. But I can’t see any difference between the line?!

I don’t think Snapmaker can change K value on the fly during print. So the K value is what you have set in firmware. If all lines look good then you probably have a good K value set.

I though exactly that was the generator for?!?!