Consistent Z lines - Solved
-
Thanks for the reply. It was one of the first things I checked and I can't see any play anywhere, to be sure I changed the hotend mount and it made no difference. I've also checked the nozzle and heater block.
I assumed it was not mechanical because the lines are so consistent on all 4 sides and through each layer.
-
What is your z axis mechanism?
-
2 lead screws either side of the X axis with anti-backlash nuts. I've coupled the lead screws with a belt but it's made no difference. I have a 6mm aluminum plate with a silicone heater underneath and a PEI sheet on top.
-
Experiment with layer height to see how the banding changes. Does the layer height evenly divided by steps/mm (sometimes an issue if using Imperial leadscrews)
-
I asked, as would suggest the same as Tony. You want your layer height to be a full step multiple of a full rotation of the motor. So if the screw turns 1 full rotation to give you 8mm of linear travel, and you have 200 steps/rev (1.8deg) motors then your layer heights should be in 0.04mm multiples, so 0.04, 0.08, 0.12, 0.16 etc… This is partly the reason why 2mm lead is preferable with 1.8 deg motors as the multiple is 0.01 which means any common layer height works great. Factor in any gearing down you have used between the motor and the screws. Or if you have 400 steps/rev (0.9) motors.
-
As DJ said except that from your config.g, I see you have 800 steps per mm which would indicate 4mm lead screws or 2mm with 0.9 degree steppers. In any case, you say that you are using 0.09mm layer height which is a really odd number to choose and means that you would be relying on micro stepping for positional accuracy which is not a good idea IMO.
In the other similar thread to this, the poster said that he has the problem when sliced using S3D but not when sliced using some other slicer ( Matter control or some such I think). As you too are using S3D, maybe that could be a contributory factor too.
-
Thanks everyone for the comments to date, to clarify a few points. I am running 1.8 degree steppers with 200 steps per revolution, x32 (256 interpolation) microstepping. Before the comments above I was certain I was running M8 leadscrews (8mm per revolution) but now I'm not sure. I measured 1 revolution of travel with my callipers and it was around 8mm, I'm guessing if it was imperial it would be close to that too, the antibacklash nuts are these http://www.makerstore.com.au/product/anti-backlash-nut-block/ and the leadscrew travels through without snagging. I tried printing again with 0.12mm layer height and the results are pretty much the same; see below.
I have the same result printing from either Cura or S3D.
How can I tell if the leadscrews are imperial?
-
Are you using PID or bang-bang control for the bed heater? Try PID if you are not already using it. I've seen Z banding cured by switching the bed heater to PID before.
Edit: I've just seen in your config.g file that you are using PID.
-
Ref steps per mm. At "normal" microstepping of 16X, 800 divided by 16 is 50 full steps per mm. I revolution is 200 steps with a 1.8 degree motor so that's a quarter of a revolution, therefore a full revolution would give 4mm, hence my reasoning that your lead must be 4mm. However, I've just noticed that you are using 32x micro stepping so my assumption was wrong. The maths is 800 divided by 32 = 25 full steps per mm = one eight of a revolution = so the lead is indeed 8mm. Apologies.
So we have 25 full steps =1mm. One step is therefore 0.04mm so you ought to choose layer heights that are multiple of 0.04. i.e 0,04, 0.08, 0.12, 0.16 etc. I've noticed that you say you tried with 0.12 layer height and the results were the same so we can now discount the fact that you may be relying on micro stepping for positional accuracy.
As I said before, the other OP resolved his issue by using a different slicer to S3D. Maybe try Slic3R? It's free so worth a shot. On the other hand, it could just be an extrusion issue - try printi9ng with the extrusion multiplier set to it at say 95% or 90%. If it makes it worse, try increasing it. For some reason, on my machine, I invariably get best results running at around 90%, even though the extruders are all calibrated.
HTH -
Thanks for that. The print in the photo above was done at 90% extrusion multiplier, I've tried 100% and 110% and the cube looks the same.
The only other issue I can think of, but it's a long shot, is that there is some pull on the hotend from the ptfe tubing. I've ordered a titan extruder and will try that out and see how it goes.
-
I've managed to smooth out the Z lines … mostly. I moved the Z lead screws further back which seemed to do the trick.
I'm now trying to eliminate the ringing, the strange thing is that no matter what I set Jerk and acceleration to, the ringing is unchanged.I would like to clear the ringing so I can fix the remaining Z issues
Jerk at 500 Acceleration 1000
Jerk at 1500 Acceleration 3000
Jerk at 4000 Acceleration 10000 -
What sort of extruder are you using (apologies if you have already said above)? I ask as it seems to be less noticeable at faster head speeds which the 4000 (very high) jerk and 10000 acceleration will give you. Might it be extruder pulsing?
What type of linear motion device do you have, I do not see this with rails, but rollers and smooth rods might be less good.
-
Also see this thread http://forums.reprap.org/read.php?1,757017 "war against ripples".
-
Do yourself a favor and download the solid 40mm cube from thingiverse then set bottom and top layers to 0 and set parimeters to 1
Itterate this print over and over again at 60mm/s this will allow you to get better visual feedback over what is happening. My Z lines are non existent because of doing this and it is helping me dial my pressure advance retraction settings in a way that just was not possible to see with a solid cube.
Also it makes it so it is not such a huge undertaking because it prints so fast.
The other tip i recommend is that your slicer is set to 5% overlap not 10 or 20. Set your extrusion multiplier to 1.00 and set your extrusion width to match your nozzle tip width. Then change your heat down or up until you get layer widths that match your nozzle width.
Doing that with your slicer if all your jerk, accell speed settings are right and your mechanics are tightly tuned you will not be able to see layer lines without close inspection (with pla you will need to turn the object sideways and view the layer lines in the light reflection) there will not be any ridges just smoothness the lines are only a reflection of light being diffused when everything is calibrated correctly
-
p.s. use a caliper on the cube walls to verify all is as it should be.
-
Solved!
I had changed the original H-Bar ends from a 4 wheel version to a 3 wheel version when building the D-Bot. The 3 wheel version moved well and there was no play at all. What I suspect happened was the H-Bar was twisting or flexing slightly as the belts applied force at either end.
Going back to the 4 wheel version solved the problem.
Thanks everyone for your help.
-
Solved!
I had changed the original H-Bar ends from a 4 wheel version to a 3 wheel version when building the D-Bot. The 3 wheel version moved well and there was no play at all. What I suspect happened was the H-Bar was twisting or flexing slightly as the belts applied force at either end.
Going back to the 4 wheel version solved the problem.
Thanks everyone for your help.
Can I ask which 3 wheel z carriage mod you were using?
