Slic3r 1.2.4 Settings & Terrible Prints

I am struggling with Slic3r settings and could really use some advice on what I have set wrong. First here are some results:




All three items were printed on the same printer with the same filament. The cube on the left was printed with an ancient Skeinforge and factory firmware. The thin-wall part on the back (_40x10.STL) was sliced with Slic3r 1.2.4 after upgrading to the latest Sailfish firmware. It is 0.39mm thick as expected, but had some defects. To tune it up, I used the same settings in Slic3r and told it to print the cube previously sliced with Skeinforge for reference. It was printing so badly, I didn't even let it finish.

I included my Config Bungle from Slic3r 1.2.4 with this post. The settings used for the thin-wall test and the second cube were as follows.
Printer settings: 0.1-LH-Fast
Filament: Dia179-H220B100-H220B100
Printer: Sailfish-LH-220C-100C

Here are some other data common to these prints you may ask about:
ABS
Head 220C
Bed 100C
All printed on the left head
FlashForge Creator Pro (clone) with controlled interior space (no wind) and MK8 Stepstruder dual head

What am I doing wrong? I'm happy to provide anything else you guys need to help me get this thing working right.

Thanks in advance,
Kyle

The unusual settings that caught my eye are:

1) You have set an extrusion multiplier of 0.57 which means that far less than the calculated volume of extruded plastic will be commanded. While some adjustment is often required, that seems to be quite a lot - how did you arrive at that figure and is your filament diameter of 1.79mm correct?

2) Your speeds are higher than my printer would be capable of printing (my extruder would be skipping like crazy), but maybe they are OK for your printer. e.g. Your first layer has an extrusion width of 200% and a height of .35mm. This means that the volume of plastic extruded over a given distance will be 7 times more than the same distance on a higher (0.1mm) layer. Yet your speed reduction is set at only 50%, meaning your extruder is pushing 3.5 times more plastic on the base layer at a given print speed as it does on higher layers. Pushing plastic through a 0.4mm nozzle with an extrusion width of 0.8mm or so (=200% of 0.4mm) and a height of .35mm (on the bottom layer) while travelling at 60mm/s (= 50% of 120mm/s) seems to me to be asking a great deal of the extruder & hotend, but I am not familiar with your printer and maybe it can handle that. OTOH your print perimeters have all the symptoms of massive extruder skipping, so I wonder whether it can. My thinking is that as you have chosen honeycomb pattern for sparse infill, that infill speed is going to be slower than your commanded speed (120mm/s) due to the frequent direction changes making it acceleration-limited - and so the honeycomb is printing OK, but your base layer solid infill (which the photos do not show) will be at high speed - if that infill is as broken as the perimeters it would tend to confirm my suspicion.

You mention that the defects occurred after a firmware upgrade. One possibility is that the older firmware had a firmware-imposed speed limit that is lower than the limit imposed by the new firmware. Thus previous speed settings that were too fast would have appeared to have worked OK - because you were not in fact printing at those speeds, but now that you are the extruder cannot take it.

I suggest try changing to an extrusion multiplier of around 0.83, print speeds of 70mm/s or less, first layer extrusion width of 100% and first layer speed reduction of 40% and see what result you get. If that's OK, you can try increasing your speeds.

Dave

Thanks for the thoughful reply, Dave!

On point #1, I too think 0.57 is quite low. After switching to Sailfish, I followed this guide for tuning your extrusion multiplier: Calibration The left head (used in the pictures above) measures 0.39mm. Using a multiplier of 1 yielded a 0.64mm wall thickness on the _40x10.STL model. Printing/measuring/calculating indicates 0.57 which seems to low. But with 0.57 it does print the _40x10.STL with a 0.39mm thickness as I mentioned in the original post. So we are in agreement that 0.57 seems way off, but I am unsure what to do about it.

You asked about the base layer solid infill. Here it is:


No argument... my numbers on the base may be way off. You can see there is no skipping however. There is no chattering, and the hot end is able to maintain temperature during the entire process.

It seems like I need to keep that extrusion multiplier, due to the calibration routine indicating that value. However, I will try your other suggestions on the two speeds and first layer width to see how that affects the print.

Thanks again,
Kyle

I forgot to answer one of your questions. Yes, my filament diameter measures 1.79mm.

Thanks,
Kyle

Quote
dmould
I suggest try changing to an extrusion multiplier of around 0.83, print speeds of 70mm/s or less, first layer extrusion width of 100% and first layer speed reduction of 40% and see what result you get. If that's OK, you can try increasing your speeds.

Dave

Here it is with the extrusion multiplier 0.57, print speeds 70mm/s, first layer width 100%, first layer speed 40%:




I attached the latest config bundle. I used
Print settings: 0.1-LH-Medium
Filament: Dia179-H220B100-H220B100
Printer: Sailfish-LH-220C-100C

I'm not worried about the issues on the top of the calibration box at this point. I'd just like to get the sides back to working the way they did before.

Thanks, Kyle

Your problem is not slic3r. Reduce the speed or increase your temp.

Only increasing the temp you should need 10 to 20°.

---

Triffid Hunter's Calibration Guide

--> X <-- Drill for new Monitor

Most important Gcode.

I agree that this is a printer issue not a Slic3r issue. The extrusion is tearing as it is laid. A small tear with ABS tends to curl up and cause the layer above to tear, thus snowballing. Here are things that IME can cause that:

1) Bad filament quality. Filament that contains impurities or bubbles or has not been stored in a very dry atmosphere can have voids or can "popcorn" when extruded.
2) Print speed (still) too fast. As proof of concept try the print extremely slow (20mm/s or even less). This can also compensate for the other issues.
3) Hotend not quite hot enough. As Wurtnase says, try a 10 or 20 deg increase. Your temperature readout may be inaccurate.
4) Partially clogged nozzle - perform a cleaning operation (allow hotend to cool until plastic is thick but not yet solid and retract fully - debris hopefully sticks to end of filament which is then cut off).
5) The bottom of the nozzle is rough or contaminated - it should be mirror-smooth so melted plastic doesn't stick to it.

Can anyone think of other possible causes?

Dave

Quote
Wurstnase
Your problem is not slic3r. Reduce the speed or increase your temp.

Only increasing the temp you should need 10 to 20°.

This morning when I turned on the machine, both heads reported 24C, which indicates their temperature is accurate at the low end. For these prints, I changed the head temperatures from 220C to 240C and nothing else. The box is now square, so that is an improvement. But it still is not good like what I was getting from Skeinforge on the same printer with the same roll of filament:



The nozzle and Stepstruder drive gear are clean/good. I had to check this to verify the nozzle diameter during calibration. Next I will try printing extremely slowly and report back my results.

Thanks,
Kyle

Latest results using 20mm/s (config bundle attached) with these settings:
Print settings: 0.1-LH-Slow
Filament: Dia179-H240B100-H240B100
Printer: Sailfish-LH-H240B100



The bottom is nice and shiny at the really slow speed. But the sides are still all torn up even at the 20mm speed with the 240C head.

Dave and I both think the extrusion multiplier of 0.57 is too low. The bottom layers all look good at 100% width (i.e. we have the right amount of material). Slowing down and working hotter have made little effect. Should I just increase the extrusion multiplier and see if things get better, regardless of what was determined through calibration?

Thanks,
Kyle

I increased the extrusion multiplier to 0.83 as Dave suggested in his first reply:

Thanks,
Kyle

Can you test whether the Skeinforge file still prints OK? I cannot see that the sliced G-Code is responsible for the failure after the changes you have tried, but there is a possibility that the printer firmware is to blame, so I suggest reverting the firmware and try again. Also it is very often possible to see what may be causing it by looking carefully while it is printing - are there any unusual moves (e.g. jerky) or does the extruder stop or start at unexpected places?

A couple of additional points:

1) An accurate low-end temperature is no indication that the high temperatures are accurate - in fact firmware is often set up so that the high temperatures read OK, but the low temperatures are way out - my hotend is out by 15 degrees at room temperature but it is accurate at 230 deg. I have checked my hotend temperature by replacing the filament with a wire thermocouple probe so that the end of the thermocouple was sitting right in the puddle of molten plastic in the nozzle. If you have access to a digital thermocouple thermometer it may be worthwhile checking (an IR thermometer is no good for checking hotend temperatures for a couple of reasons). If you have a multimeter with a thermocouple input, you can buy a flexible wire type thermocouple pretty cheap (multimeters usually need a type "K" thermocouple).

2) ABS filament can degrade over time - especially if it has not been stored in a sealed bag with desiccant. It absorbs moisture from the air which can collect in small voids inside the filament. Upon reaching the nozzle the water boils causing the extruded plastic to "popcorn" which can cause the breaks that seem to be happening with your print.

3) Cleaning the nozzle is not always successful and perhaps there is still a possibility that it is partially blocked. Over on the Ormerod forum a poster said he had to clean 6 times before a blockage was finally removed.

4) Check your mechanics - I once saw a similar symptom (though not as bad) when my hotend tilted due to a slipped bearing so that the nozzle was no longer vertical.

Dave

Edited 1 time(s). Last edit at 01/27/2015 08:02AM by dmould.

hi,
you seem to be going too fast.
try 30mm / s on the perimeters and 45mm / s for infill

---

Enrico

[www.dapa3dservice.it] [www.studiodapa.it]
repstrap dApA, Ramps 1.4, MarlinKimbra / scanner 3D Cubify Sense / Formlabs Form 2

Quote
enricodare
hi,
you seem to be going too fast.
try 30mm / s on the perimeters and 45mm / s for infill

He has already tried 20mm/s with no joy :-(

Dave

I don't think it is the machine or Slic3r itself. I still suspect I have a setting messed up somewhere in the Slic3r setup.

I dislike Skeinforge (its GUI is horrid), but I see the logic in your request. Unlike Slic3r, my current Skeinforge settings are putting out too much ABS. You can see here that everything is running smoothly. It even has a "hump" in the middle where the plastic is building up. I did not run the print to completion for fear of damaging my printer from the head moving through solid plastic. You can see the dimple where the head came to a rest in the part at the back:


I increased the extrusion multiplier to 1, sliced, and printed:


It looks just as bad as before. I double checked to make sure I did not re-print an older version. I am considering increasing the extrusion multiplier beyond 1 to see if I can get Slic3r to put out more plastic.

I'm using 1.2.4. Can anyone verify that the extrusion multiplier works on Win7 Pro 64 bit? Looking at the picture history, once I slowed the print down, it looks like I get the same results regardless of which extrusion multiplier number (0.57, 0.83, 1.00) I use on the 20x20x10 cube.

Thanks so much,
Kyle

Quote
kda406
I'm using 1.2.4. Can anyone verify that the extrusion multiplier works on Win7 Pro 64 bit? Looking at the picture history, once I slowed the print down, it looks like I get the same results regardless of which extrusion multiplier number (0.57, 0.83, 1.00) I use on the 20x20x10 cube.
Thanks so much,
Kyle

I've had a sudden thought. Is Slic3r set for absolute or relative extruder moves? How about the printer?

I am suspecting that Slic3r is outputting absolute moves while your printer is set for relative moves. It would explain many things. Each move would progressively over-extrude worse and worse until the Slic3r code does an extrusion reset (G92 E0), which it does at regular intervals, at which point the printer's extrusion will suddenly decrease before building up again.

The "Custom G-code" in Slic3r should set the printer to either relative (M83) or absolute (M82) mode, and there is a tick box under "General" printer settings to set relative E distances (checked) or absolute (unchecked). Either have M83 in the custom code and the box checked, or M82 and the box unchecked.

If that's not the problem, post the first 100 or so lines from your Slic3r G-code file and your Skeinforge file and I will try to identify what is causing the big difference.

Dave

Still new to 3D printing, but I was thinking the extrusion multiplier was for minor adjustments and that anything beyond 0.05 (0.95-1.05) was an indication of a poorly calibrate extruder motor.

Did he calibrate the stepps per mm in his firmware or just in his slic3r settings, thus the 0.57 multiplier?

Apologize if that is a stupid question.

Quote
Kurzaa
Still new to 3D printing, but I was thinking the extrusion multiplier was for minor adjustments and that anything beyond 0.05 (0.95-1.05) was an indication of a poorly calibrate extruder motor.

You would think so, but a thin wall test showed that I needed an extrusion factor of 0.83 (5/6) to get the extruded width correct despite the fact that I know my extruder is correctly calibrated and my filament is the correct diameter, and that seems to be a common finding of many people. I have no idea why that should be. Maybe Slic3r's volume calculations are out, or maybe the density of extruded plastic is different to the filament's original density so that (volume out) != (volume in). Obviously the volume of the hot molten extrusion is higher than the input volume because of heat expansion, but that should quickly reduce to the correct volume as it cools.

Dave

Quote
dmould
The "Custom G-code" in Slic3r should set the printer to either relative (M83) or absolute (M82) mode, and there is a tick box under "General" printer settings to set relative E distances (checked) or absolute (unchecked). Either have M83 in the custom code and the box checked, or M82 and the box unchecked.

Interesting thought. My custom G-code does not call out either. "Use relative E distances" is not checked on my Slic3r settings.

Unfortunately, neither the Slic3r nor Skeinforge gcode files call out M82 or M83 anywhere in them. I have attached both the Skeinforge gcode which is smooth and puts out too much plastic, as well as the Slic3r gcode file. I zipped them so they would fit on the forum.

Thanks,
Kyle

Quote
kda406

Quote
dmould
The "Custom G-code" in Slic3r should set the printer to either relative (M83) or absolute (M82) mode, and there is a tick box under "General" printer settings to set relative E distances (checked) or absolute (unchecked). Either have M83 in the custom code and the box checked, or M82 and the box unchecked.

Interesting thought. My custom G-code does not call out either. "Use relative E distances" is not checked on my Slic3r settings.

Unfortunately, neither the Slic3r nor Skeinforge gcode files call out M82 or M83 anywhere in them. I have attached both the Skeinforge gcode which is smooth and puts out too much plastic, as well as the Slic3r gcode file. I zipped them so they would fit on the forum.

Thanks,
Kyle

OK, both are using absolute positioning, and also Slic3r does not re-zero the extrusion at frequent intervals so it is not possible that the printer is in relative mode (you would soon notice if it tried to extrude 500 meters of filament in one move! It may nevertheless be an idea to try switching to relative mode, just in case it makes a difference (there should be no difference at all). Simply put a M83 command at the end of the custom start code and check the relative E distance box. The Slic3r code is extruding far less than the Skeinforge code - Skeinforge uses nearly 2m of filament whilst Slic3r is using under 700mm (less than half as much) which indicates that your extrusion multiplier (0.57) was used. Layer height is 0.1mm for both. Slicer has a first layer of 0.35 and Skeinforge appears to use 0.05mm first layer height (seems awfully low to me). Skeinforge speed is 60mm/s, Slic3r 70mm/s.

I suggest that you check your extrusion calibration. Put a mark on the filament as it enters the extruder, and another mark 100mm further away. Manually extrude 100mm into the air (I assume the printer has a control to allow you to do that), and check that the second mark you made is at the same position the first mark was at. If not you should change the steps per mm in the printer's configuration (though you can also use the difference to calculate the theoretical extrusion factor to use).

At this stage I also suggest trying the same print with a layer height of .25mm to see whether a thicker layer of extrusion prints better.

Dave

Hello,

I have the same issue with slic3r.
I have a CTC with sailfish and when I use skeinforge everything is working fine.

What I have noticed is that a single wall object is good, but dual perimeter or with infill are terrible.

@Kyle did you find a solution?

kind regards
Alex

After many months I have the solution.

Dave: printing thicker did not help. My printer uses the x3g protocol, so extruding 100mm was difficult. I eventually set up an empty print and wrote my own custom g code header in Slic3r to extrude 100mm. It extruded 99.9mm as close as I could measure. Skeinforge continued to work, but put out a little too much plastic.

Alex: I'm happy to report I just yesterday found the solution.

All the pictures above (5 months ago) were sliced with 1.2.4 for Windows 64 bit.

Last week I upgraded to 1.2.7. I reset my filament setting to 1.76mm (measured), extrusion multiplier of 1, layer thickness 0.3mm, and got this result with the other settings the same as above:


My boss suggested that I check to see if there were any bug reports on 1.2.7 and when I did I found that 1.2.8 had just come out. So I updated to 1.2.8 and did a test print with no Slic3r setting changes from the 1.2.7 print above:


Then I Slic3ed and printed a 6 hour print overnight of a large tool-fixture. It also printed just fine using 1.2.8.

There were no changes between the 1.2.7 and 1.2.8 prints above except loading a new version of Slic3r. The printer, the filament, the settings, the environment were all the same. I conclude that it was Slic3r all along..

Thanks to all those who tried to help. It has been a very frustrating few months. I'm glad Slic3r 1.2.8 works.

Thanks,
Kyle

Hello,
Although you said you didn't change the setting, I still hope you can post the 1.2.8 setting for analysis. Maybe the default value has change and you didn't realize.

I've read that sometimes slic3r gets stuck, and changes in settings do not affect it any more until you start with a new configuration from scratch. Might this update to 1.2.8 have just done this?