SOLVED Horizontal bands

[Alex101 0]

Hi all,

 

I'm not sure if I am allowed to post here (as I don't own an Ultimaker printer). I use Cura 4.8 though, I hope that the issue could be resolved via  settings Cura settings, and this forum is likely to be my last hope.

 

The issue is that I'm getting horizontal bands on my prints - only when there is a feature (hole, groove, infill change, etc) that interrupts the continuous flow. These bands look like the thickness of the perimeter suddenly gets changed and then goes back to normal. E.g. it is wider when uninterrupted and narrower if there is a feature in place. If there is no feature the printing is smooth.

 

I have done all the calibrations (esteps, PID, flow, etc) the best I could, bed id level,  and also (this is a coreXY printer if it matters) adjusted and readjusted belt tension, carriage roller tightness, etc. The printer feels like it is tuned like a violin except this very issue. And I have so support from the vendor; many owners of the same printer I contacted have the same issue but no resolution or even idea.

 

May I please ask the experts here to have a look at atatched photos and suggest what to look for? Or perhaps redirect to another forum if this is only for Ultimaker owners. Thanks

Edited February 11, 2021 by Alex101

[gr5 2,167]

You also have retractions on those layers.  I'm guessing you enabled "z hop when retracted".  Turn that off.

 

z hop is meant for delta printers.  99.99% of printers are not delta printers so you probably don't have one.  z hop causes the z axis to move the nozzle farther from the print and then back and due to backlash you have inconsistent results where some layers overextrude (hence squish the next layer so it sticks out and makes a line) or underextrude (nozzle slightly too far from print so you get a layer that is thinner and sticks in and makes an inward line).

 

 

[Alex101 0]

Thank you very much. You are correct, I have Z hop enabled on the 1st print (the "in progress") photo, but I didn't have it enabled on the others (in fact today is the 1st time had it enabled).  Still I had issues.

 

Having said that, I will of course reprint without Z hop enabled, and yes my printer is CoreXY (Creality Ender 6) not delta.

 

I do believe that your suggestion is relevant and on the right track even while I have the issues even with Z Hop disabled - e.g. it may be something to do with retraction as this printer is Bowden setup and has a very long tube. I'm not arguing with your suggestions, just clarifying it further.

 

May I please ask you & the community  to keep helping me in troubleshooting this? Apart of disabling Z Hope, what else should I look at?

 

Many many thanks!

Edited February 11, 2021 by Alex101
Typos

[gr5 2,167]

Those horizontal bands are almost always caused by Z errors.  So try thicker layers.  Let's say the error is typically 0.02mm randomly.  If your layer height is 0.1mm then that's a 20% error and you will get 20% over or under extrusion causing the layers to stick outward (or inward) about 20% of line width.  If your layer height is 0.2mm then all the errors will be half that.

 

Cleaning the z screw(s) might help.  If your bed moves down (versus nozzle moving up) then putting a heavy weight like a brick on the print bed may help.  You might need to replace some of the Z hardware (vertical rods, bearings around the rods, z screw or Z nut).  The Z nut is the easiest to replace and moving from a 0.50 dollar part to a $3 part can make a huge difference in quality.

 

I have seen several printers with those horizontal lines and cleaning out the z screws helped a lot.

[Alex101 0]

Thank you. My bed is moving down. I assume Z screw is that long vertical threaded rod attached to Z-Axis stepper motor? And Z nut connects the screw to the motor? Thanks, I'll try that.

 

To illustrate - in the attached photo, I have a line as soon as the fillet ends, then it is pretty much smooth. Apologies, the photo uploaded upside down for some reason.

 

Edited February 11, 2021 by Alex101

[gr5 2,167]

Z nut is embedded into your bed - when the z screw rotates the nut pushes the bed up or down.

 

Note that the Z screw is triple helix on 90% of printers out there so you need to clean all 3 helixes.  Try just cleaning the top few 5 cm or so as that's probably where it goes through the nut when printing small parts like in your photos.

 

Try the "brick" trick also where you put some heavy weight on the bed out of the way - at least 3 pounds or 2 kilos.

 

If you get better results with the brick trick then I'd take the z screw out of the printer and clean it very well with wd40 over newspaper and then add 1 drop of grease.  Or less than 1 drop.  Only the Z screws get grease - all other rods get oil or even better possibly nothing.  If there are ball bearings in a bearing then maybe better to not use any oil at all.

 

For example the bearings for the Z rods probably have ball bearings inside.  Any oil will just help dust get in there and should probably be avoided.

 

If you do take out the z screw(s) then the added benefit is you can slide the bed up and down and feel how much resistance there is.  Sometimes the Z rods are not parallel or other issues.

[Alex101 0]

Thanks a lot.  I have tried all of that - totally cleaned the Z screw (the "working" part as you suggested), tried the brick on the bed, re-tightened belts and  pretty much everything I could, printed and installed a ball bearing bracket for the top of the Z screw, checked Z Hope and retract at layer change are off, even reduced retraction settings as a test.

 

No luck, effects are still there.

[gr5 2,167]

More pics please.

 

Well I didn't mention it but very rarely the problem is that the temperature of the nozzle is not stable.  You can see this by just watching the temp as it is printing the "bad" layers.  The temp will be at least 5C hotter or colder on those rows than the rest of the print.  I don't think this is it because it is only happening on specific layers.

 

Another issue is play/backlash (if you don't know what this is look on wikipedia).  It really helps to have closup pics of the problem from a few angles.  Play can come from many sources like a loose head, too much friction, loose belts,  Basically with the nozzle cold and power off push gently on the nozzle left/right and print bed forward back.  Push as hard as possible without moving the stepper (not very hard).  If the head or bed can move back and forth by say 0.2mm (easy to feel that amount - not sure how our fingers detect but they do) then that could be the issue.  Loosish belts are hard to detect with this method and instead you pluck the belts like a guitar string and use a guitar tuner to test.

 

The symptom of backlash is that it's perfect on identical print paths (goes around the same corner before printing a perfect always from left to right (or always the other way).  Edge of wall has same exact shape each time - no other changes from one layer to another.  Those walls will be consistent (same backlash) but if the wall corner is now sharper or longer or some small difference then a layer sticks out but typically only on left side or only on back of model or whatever.  It's hard to define "identical print paths".  I know it when I see it.

 

more pics of the problem parts since after getting rid of "zhop".

 

how to test belt tightness:

 

[Alex101 0]

Thank you, doing some test prints and will post pics. Thank you very much.

 

Update. I have tested the nozzle, the bed, etc as you suggested, all seem tight.

 

Pictures of the latest test print attached, please let me know if the do tell the story.

 

Thanks

 

Edited February 14, 2021 by Alex101

[gr5 2,167]

Wow.  It really starts at the holes!  It's really obvious.  Are you absolutlely 100% sure you don't have zhop on?  I mean zhop would explain this perfectly!  Excessive retraction (say 10mm instead of 1mm) could also explain it.  Something related to retraction.

 

Please share your project file: go to menu "file" "save...".  The resulting file contains not only your STL but the scaling, positioning.  Also it has your machine (printer) settings, your material settings, your profile used, and settings that you overrode.  Please post that here.

[Alex101 0]

On 2/15/2021 at 5:38 AM, gr5 said:

Wow.  It really starts at the holes!  It's really obvious.  Are you absolutlely 100% sure you don't have zhop on?  I mean zhop would explain this perfectly!  Excessive retraction (say 10mm instead of 1mm) could also explain it.  Something related to retraction.

 

Please share your project file: go to menu "file" "save...".  The resulting file contains not only your STL but the scaling, positioning.  Also it has your machine (printer) settings, your material settings, your profile used, and settings that you overrode.  Please post that here.

 

Thank you very much. Will absolutely do and appreciate your help. 100% Z hop (Z Hop when retracted) is not on,  however there are 2 very valid points you raised:

 

1) my retraction is around 7mm x 50mm/s (down from 10mm that came with Creality Version of Cura/profile for my printer) - its is a very long bowden tube setup and prone to stringing.  I would be very grateful for any hints here,

2) What I noticed after you mentioned that "backlash" (I have checked the temp, the fluctuations are within 0.1 degrees C): lets say I'm printing the hole or the arc (top of 3d Benchy) on the wall. The "edge" of the layer where the print head temporarily stops extruding kind of curls up/blobs as soon as it gets printed. Then the print head kind of hits that curl up/blob on the next extruding move within the same layer and (potentially) knocks itself off. I have observed hitting today and could hear the bump. If it makes sense. No clue what to tune to avoid that (maybe retraction as you point out earlier) but this likely being a problem.

 

Attached is the curl up and "fillet" (narrower layer) I was talking about - taken mid print.

 

Thanks and I will post the project file later on.

 

Edited February 16, 2021 by Alex101

[gr5 2,167]

Yes I call that "raised edges" on overhangs.  But that wouldn't explain the bottom half of the hole.  There's not much to do about raised edges except make sure your parts sticks like hell to the print bed.  That's a different topic though.

 

Is the feeder on the (1) print head?  Or is it on the (2) back of the printer on the other end of a long bowden?  If (1) you typically want about 1 or 2mm of retraction.  If #2 then you want about 7mm per meter of bowden.  7mm may be too much - if you retract too much then air gets in the nozzle and creates all kinds of problems with over and underextrustion which could explain what you are seeing.  You might want to try just 4mm retraction.  Look at the top of the arc of the bowden while printing - during retraction it should drop off from the top of the bowden and almost rest on the bottom of the bowden in that top arc.  Any extra retraction will cause the filament to pull upwards out of the nozzle which is bad.

 

When you are printing it's normal to have 100psi of pressure in the nozzle (maybe 1000psi - I forget exactly).  Retracting just enough to get that to 10psi is plenty of retraction.

[Alex101 0]

1 hour ago, gr5 said:

 

Yes I call that "raised edges" on overhangs.  But that wouldn't explain the bottom half of the hole.  There's not much to do about raised edges except make sure your parts sticks like hell to the print bed.  That's a different topic though.

 

Is the feeder on the (1) print head?  Or is it on the (2) back of the printer on the other end of a long bowden?  If (1) you typically want about 1 or 2mm of retraction.  If #2 then you want about 7mm per meter of bowden.  7mm may be too much - if you retract too much then air gets in the nozzle and creates all kinds of problems with over and underextrustion which could explain what you are seeing.  You might want to try just 4mm retraction.  Look at the top of the arc of the bowden while printing - during retraction it should drop off from the top of the bowden and almost rest on the bottom of the bowden in that top arc.  Any extra retraction will cause the filament to pull upwards out of the nozzle which is bad.

 

When you are printing it's normal to have 100psi of pressure in the nozzle (maybe 1000psi - I forget exactly).  Retracting just enough to get that to 10psi is plenty of retraction.

 

This is a long bowden setup (extruder at the back, approx 50cm of bowden tube).  I think this is the retraction issue as you've pointed out (ignoring the arcs), I'll be printing the test with retraction  disabled (ignoring strings), to see if its solves the "bottom half of the hole" problem. I may be having too much of it in my chase of no stringing.

 

Thanks and I will post the progress.

 

 

 

[gr5 2,167]

No extrusion is bad also.  Try 3mm of extrusion.

 

If you look at the bowden - usually it's transparent enough that you can see the filament inside as the printer is printing.

 

Look all printers are different - sometimes the extruder is at the top of the printer - a meter above most of the printer - maybe it would be good if you tell us exactly which brand you have.

 

Anyway the bowden arcs?  If you look inside the bowden while it prints it's hard to see what is going on until suddenly when there is a retraction you can see the filament move.

 

Usually there is a lot of pressure from the feeder - the feeder is pushing very hard - often hard enough that if you tried to push just as hard you would easily lift the printer off the table.  Something like 20 pounds or 10kg of force.

 

So think of the inside of the bowden as a tunnel with a diameter significantly larger than the filament.  When under pressure, at the top of the bowden arc, you can see the filament is at the roof of that tunnel.  when the feeder retracts the filament should gently rest on the floor of that tunnel (again - at the top of the arc).  It should retract the absolute minimum such that it is no longer pushing at the top of the tunnel.  If it retracts so much that it actually pulls out of the nozzle then that's too much retraction because now air has to go in somewhere - possibly through the nozzle tip.  Air in the nozzle will cause problems.  Even if the air goes in through the bowden that's bad as well - you can get air bubbles in among the molten filament and those will cause overextrusion as they expand and then underextrusion when they escape through the nozzle.

[Alex101 0]

Managed to get from this to this. Better, not ideal. As you have mentioned, the issues was in retraction. Where the wall is interrupted by 2 holes, it is getting sliced into 3 rectangles (not sure if I'm explaining it properly) and because 2 outward rectangles and the gap between small, the retraction didn't have enough time to push back enough material and the layer was getting "starved".

 

Managed to improve with 3.5mm @25mm/s back and 15mms/s forward plus 0.8mm3 if extra priming amount.

 

Now I'm having blobs on 3D benchy hull sides. 😕  Will try to find a balanced set up, however I wonder if this is the only way?

 

BTW I asked another person (more experienced) person who has the same printer to print my test STL with his printer/setup and  he came back with the same issue.

[gr5 2,167]

So the one on the left is the latest?  If so then you don't need any more help with retraction!  Quality is getting so much better!

 

If the left one is the newer version and you want to refine retractions even more you can create experimental parts where you don't have to print for hours to be able to see if there is a problem or not.  Then play with retraction amount.

 

We call those rectangles you mention "islands".  When you think of only one layer and ignore that the part is 3D then you have islands and you retract when going from one island to another.

 

 

[Alex101 0]

7 hours ago, gr5 said:

So the one on the left is the latest?  If so then you don't need any more help with retraction!  Quality is getting so much better!

 

If the left one is the newer version and you want to refine retractions even more you can create experimental parts where you don't have to print for hours to be able to see if there is a problem or not.  Then play with retraction amount.

 

We call those rectangles you mention "islands".  When you think of only one layer and ignore that the part is 3D then you have islands and you retract when going from one island to another.

 

 

 Thank you very much, the left one is the latest, the right one is when I started experimenting. I still need to do some tweaking, as I do not think using extra priming amount is the right long term way forward (it may create an issue with other prints, that don't have "islands"), but at least there seems to be a light.

 

Thank you so much.

 

I wonder if Cura might have some sort of dynamic compensating setting for islands (maybe in a future)?

 

Many thanks.

[gr5 2,167]

You are using extra priming?  I've never used that.  You shouldn't need that.  I thought you fixed it by reducing the retraction distance?