Is it a retraction problem?

Hi guys, hope you can help me with this.

When printing multiple parts (or in this case one part that starts out with two legs) I always get a big mess where the hotend leaves and enters the part when it jumps between them. A notch forms at that spot and I can hear the hotend bump into it as it goes around.

I've been messing with the settings the whole day and nothings seems to help.
I have upped the retraction quite a bit. I'm now at 4mm @ 50mm/s, started with 2mm @ 20mm/s.
Changed the hotend temps between 190-210. Tried different bed temps. Changed acceleration from 2000 to 1500.

Prints are good except for this one thing. What else could I try?

I'm printing with PLA.





Edited 2 time(s). Last edit at 04/02/2016 02:14PM by Backdraft.

Tried lowering the print speed from 60mm/s to 40mm/s. This produced better result, but the bump is still there. I can actually see the printer shake a bit when it runs it to the bump.
The issue seems to be that the hotend sits there for just a fraction of a second while retracting. I guess some of the plastic still oozes out, creating these bumps.

Is there a way to set the retraction to happen slightly earlier?

When you see the shake it is moving to another perimeter. That is the seam and it looks like overextrusion or too high acceleration (or Vxy jerk?)

Quote
sigxcpu
When you see the shake it is moving to another perimeter. That is the seam and it looks like overextrusion or too high acceleration (or Vxy jerk?)

XYjerk is 20.0 and I've lowered the acceleration already. Haven't actually measured the feed rate, but calibration cubes print fine. Seems like maybe even a little underextrution at times.

What sort of extruder? If it's a bowden you might need to go with even more retraction.

Quote
JamesK
What sort of extruder? If it's a bowden you might need to go with even more retraction.

Direct drive. Replikeo Prusa i3

Hmm, I'm having a pretty low success rate with my suggestions recently!

What you have I call leaking.
Especially PLA at slightly higher temps tends to go almost liquid when not extruded out.
This means when you move from one spot to the next you end up with that drop hitting the start point.
It will smear a bit but the excess plastic remains in place.
Layer after layer this builds up.
Final result in your case: A big bump form that ruins the print.
Here is my suggestion of a fix:
1. With your current print settings print multiple small test cubes.
2. Check the layer bonding and if good lower your print temp but go up again once you notice the layers no longer bond properly.
3. Again start with a test cube (only one this time) but as a thin walled object.
4. Adjust your retraction until you no longer get a blob when a new line starts at the perimeter.
5. Check if your slicer supports different speeds for printing and non printing moves.
6. If it does use a high but stable speed for non printing moves (somewhere in the region of a fast homing speed).
7. Last but not least: If your slicer support cleaning pillars then use it and place the pillar between the objects to be printed but not in a line, let it form a triangle.

If you can at least address 1 - 4 your print should come out much better, 5 and 6 would make it almost perfect but this needs proper calibration of your acceleration and jerk.

Quote
Downunder35m
What you have I call leaking.
Especially PLA at slightly higher temps tends to go almost liquid when not extruded out.
This means when you move from one spot to the next you end up with that drop hitting the start point.
It will smear a bit but the excess plastic remains in place.
Layer after layer this builds up.
Final result in your case: A big bump form that ruins the print.
Here is my suggestion of a fix:
1. With your current print settings print multiple small test cubes.
2. Check the layer bonding and if good lower your print temp but go up again once you notice the layers no longer bond properly.
3. Again start with a test cube (only one this time) but as a thin walled object.
4. Adjust your retraction until you no longer get a blob when a new line starts at the perimeter.
5. Check if your slicer supports different speeds for printing and non printing moves.
6. If it does use a high but stable speed for non printing moves (somewhere in the region of a fast homing speed).
7. Last but not least: If your slicer support cleaning pillars then use it and place the pillar between the objects to be printed but not in a line, let it form a triangle.

If you can at least address 1 - 4 your print should come out much better, 5 and 6 would make it almost perfect but this needs proper calibration of your acceleration and jerk.

Thanks. Tomorrow I'll try printing some test cubes and lowering the temps.

With 3-4, a cube doesn't require retraction so a blob wont form. It only happens if I have multiple objects or an object with multiple perimeter spaced out in different parts. Heres a picture to illustrate what I'm doing. The messy parts are where the retraction happens and it jump to the other part.

The cube should be at least two cubes, sorry for that.
What you want is to first get all extrusion parameters spot on.
After that you want to adjust the retraction so that no blob will form.
If you can it really helps to have non-print moves going much faster than your print moves as it gives the nozzle less time to drool between parts.
And as said you want your temp as low as possible but not so low that you get bonding problems.
Bridging tests are good for this to find out what the best temp for stretching without too much dropping is.

Some extruders actually work better with slower retraction.

There is something called pressure advance, where the extruder starts underextruding shortly before retractions and slow-downs. It can be implemented either in the firmware or at the slicer level. Some hot ends like that better than others.

I've been tweaking my setting the whole weekend and now I'm getting better results. The bumps are still there, but much smaller now.
Dropped the hot end temps as low as 185. Reduced print speeds and acceleration on Y and X as low as 800. Dropped XYjerk to 10. I more than doubled Z acceleration (from 20 to 50) so it would act quicker and spend less time idling and oozing. Also calibrated the extruder. It was actually under extruding and still I was getting the bumps
The biggest thing that made a difference is just placing table fan in fornt of the printer. I know filament cooling does wonders, but I've seen many people print with out fans and get good results....I don't know how they do it. I just ordered a radial fan that I'll later install.

Next I'll have to play around with the pressure advance. I read that it is supposed to be used with a bowden setup, because of the lag in the feed. This might help, but I feel this is just adressing the symptom and not the real cause of the problem.

What other setting can I change to help the layer change happen quicker? Is there a down side to upping the Z acceleration to a very high number?

You don't need to move up meters
The main culprit is the bowden setup.
With them it really is best to print parts that won't need retraction LOL

Do check on how much your filament actually jumps back in your bowden system:
Set the temp so you can just extrude without jamming or chewing the filament.
Extrude so you get a nice and constant flow then stop, mark the filament at the extruder and open the pressure plate.
Measure how far the mark has moved and this value is your starting distance to play with the retraction lenght.
So instead of risking to loose steps on the z-axis make sure you can retract the length of filament required fast and safe enough - same for the way back where you might have to add a little extra like 1mm so get the flow going right away.
Your extruder won't really do any work during these retractions until it needs to build up pressure again, so it is important to have the pressure plate adjusted properly as otherwise the filament will start to slip or the bolt fill up with dusted filament bits.

Sorry, I'm not using a Bowden setup. It's direct drive. I did mention this earlier in the thread.

Thats what I meant by not really adressing the problem by using the pressure advance since a direct drive setup shouldn't really need that.

Quote
Backdraft
Thats what I meant by not really adressing the problem by using the pressure advance since a direct drive setup shouldn't really need that.

I'm not completely convinced that's true, Clearly bowden is more likely to have problems, but you get plenty of pressure related issues with nylon even with direct drive. I see similar behaviour with 3mm pla which surprised me. I keep meaning to try the advance settings, but all the 'not stable' warnings put me off.

Quote
JamesK

Quote
Backdraft
Thats what I meant by not really adressing the problem by using the pressure advance since a direct drive setup shouldn't really need that.

I'm not completely convinced that's true, Clearly bowden is more likely to have problems, but you get plenty of pressure related issues with nylon even with direct drive. I see similar behaviour with 3mm pla which surprised me. I keep meaning to try the advance settings, but all the 'not stable' warnings put me off.

I don't know, I'm just going by what I read from slic3r in the settings. I'm happy to try it if I just knew how.
Do you know how to use the pressure advance to get less filament? I can't add a negative value there. I tried 0.1 and it made the problem worse as expected. That means it should fix the problem if I can just get it working in negative.

Edited 1 time(s). Last edit at 04/05/2016 01:27PM by Backdraft.

Like I said, I haven't tried it so I'm no expert. From what I've read the algorithm aims to start increasing pressure a little before an increase in extrusion rate is needed, and correspondingly reduce pressure a little before extrusion reduces - trying to compensate for lag in the system. I'm not sure what the difference is between running the algorithm in the slicer vs in the firmware, but in all cases there seems to be relatively little reports of it being used, and a lot of warnings about possible instability. Shame, because it sounds like a good idea.

How much retraction are you using?

I've been using firmware retraction in Marlin. It has to be compiled in configuration_adv.h . I have these commands in my Slic3r startup.

M208 S.05 F180 ; Firmware retraction restart setting
M207 S2.0 F1200; Firmware Retract Speed ;

If you have too many extruder steps per mm, like I had 3200 with 32 u-stepoing, you can't set F1700 or so. It has to be dialed down. If you tell it to go too fast it will just sit there and ooze. The good part about firmware retraction is you can change it during the print with the M207 to tune.

Are you using z-retraction as well? How long does that take vs. the filament retract? Or you might want to try adding some z-retraction.

Edited 2 time(s). Last edit at 04/11/2016 11:21PM by CTCHunter1.