Lulzbot Mini : Cura : OverExtruding ~15%

To get parts that look clean and are dimensionally accurate, I always have to set the Flow Rate to 85% ( or even below! ) when working with ABS or HIPS. I think PLA may be a bit more forgiving, but I still have to set it down to 90%.

I calibrated my extruder, and set the steps per mm very carefully in the firmware so that 100mm of incoming filament is precisely input when 100mm is asked for. I always carefully measure and average my filament size ( 2.84mm in this case )

Does cura have a math issue when computing flow rates? This 10-15% discrepancy is consistent, which is pointing to a systemic issue.

I’m having a similar experience, with the flow% being too high. To calibrate my flow rate I set a cube to zero infill and turn off the top infill. Then I reduce the flow rate until the set wall thickness matches what I get with the calipers. The kicker that I haven’t figured out yet is that although I’m able to get the correct wall thickness (and like you it can be as low as 80% flow), the top infill is extremely patchy. So basically, I have to overextrude on the walls to get correct infill on top, and underextrude on top to get accurate walls. :question: :question:

have you guys got any example print pictures ?

Here’s a photo of my most recent calibration print. I have the extrusion rate calibrated so that the walls are almost perfectly 1mm when set to 1mm.

However, this results in the top having gaps as you can see. If I up the extrusion rate, I can get the gaps to disappear, but then I lose the dimensional accuracy of the walls.

I’ve since gone to the IRC channel and some suggested I try slic3r. I calibrated using 1 wall this time.

And here is the top infill once calibrated(I still don’t have the raft settings right)

EDIT (Again…)
I’m now trying to get it to work with Simplfy3d. I’m having similar issues with underextrusion on the top. However, it appears that the profiles provided by simplify3d are incorrect. For example, the fan is off for ABS entirely by default in the profile provided by simplify3d. Turning the fan on magically fixed raft separation issues and has substantially improved (but not fixed) the gaps in the infill.
Anyway, it’s nice to know it’s not a mechanical issue.

I have to set my flow rate to 60% for any sort of decent dimensional accuracy, so yeah, I’m with you. The sense I got from talking with Support is that they’ve optimized for ease of printing non-functional items, though it’s interesting that I have to go even farther than you for properly-dimensioned prints.

In the brief time I tried S3D and was having trouble with their supports sticking to the buildplate due to a lack of an anchoring grid, support suggested turning off the fans too which in my ABS tests lead to some pretty warpy stringy parts and only a slight improvement in support adhesion. They did honor the rumor I’d heard though, which is that you can return it within 14 days for a refund. More of a hassle than just not buying after a tryout period, but not too big a deal either.

With a 3d printer, there are three things that control your extrusion rate. Nozzle size, Filament Diameter, and extruder calibration. If your extruder is calibrated 100% dead on, adjusted in firmware, etc, and you have the nozzle diameter set correctly for the nozzle you have, the only thing that should be influencing over or under extrusion is the diameter of filament setting. Most slicing programs use 3.00mm as a default starting points. Most “3mm” filaments actually range from 2.85mm up to 3.12mm diameter. If you have that set properly in the program for the filament you are using, and the other two are set correctly, then you should get near perfect parts. If it’s off by 10-15% every time then something is off. It could theoretically be in the software side, but I would suspect it might be the extruder calibration. Depending on the calibration method you use, speed of extrusion is critical to getting a proper calibration reading if you left the extruder attached. If you are pushing 100mm of filament through it all at once and measuring off of that, you will get a different reading than if you slowly extrude 1-2mm of filament in chunks and measure that way.

Yes, optimized indeed.
^ default HIPS settings, lulzbot mini. The layers are so overfilled that the print head crashes into the lumps making awful noises during fast moves. I should probably put in the time to adjust the settings so that it doesn’t break the extruder. After the first 20 layers, the prints usually look decent. For what this printer costs, these results suck.

I too just got a Mini a couple of weeks ago and have experienced this very same result. I blame it on some environmental factor (room temperature, air pressure, etc.) I have yet to identify. Try these settings as a starting point:

Inland (Microcenter’s house brand) PLA:
Start with the default Lulzbot profile and change the following:
Print temperature: 180 - 205 C
Flow rate: 85%

Colorfabb nGen:
Only change the flow rate: 90%

I discovered these on my own and they work with my specific filaments with my specific printer in my specific environment. As with many things, your mileage may vary.

The tips that helped me the most were:

1. Use a set of calipers to measure your filament and get an average over a few meters/feet. I just took the coils of filament off my spool and measured every third one until I had about 5 to 7 measurements and then averaged that and used the result for Cura’s filament diameter setting.

2. You do indeed get what you pay for. The filament from Microcenter was around $15/kg USD and the actual print temperature was significantly lower than the label (180-190 C) if I wanted to reduce the stringing and oozing, but for best layer adhesion and model strength, higher temperatures (190-205 C) are better. The filament diameter was very inconsistent. The Colorfabb nGen was $46 for 0.75kg, quite pricey, but worth it. The diameter was consistently within 0.01mm of the specified 2.85mm, and the print temperature was dead on 230 C.

3. Most bad prints can be traced to bad extrusion rate settings. ( I verified that the Mini is extruding the correct amount of filament specified in the gcode as best I could, but being a brand new printer, I would expect this to be calibrated properly. Other factors that play into extrusion rate are filament diameter and the flow rate setting in Cura. I verified the filament diameter as best I could and adjusted the diameter settings in Cura accordingly. So that left me with the flow rate. Physical inaccuracies have been eliminated so that left me with either errors in my measurements or some unknown environmental factor. The flow rate setting in Cura can help fine tune and account for these. Start by reducing 5% at a time using a small calibration print, something that takes less than 15 minutes to print. I used OpenSCAD to create a 10x10x1mm “cube” for this purpose. It literally takes one minute to print.

4. Settings often need to be changed depending on what you are printing. Need a model that is one solid piece that is just going to be for display? Consider using no infill and go into “expert settings” under “black magic” and check the box for “spiralize outer contour”. This makes the model one continuous line of plastic rather than separate layers. Need a model that will stand up to some abuse such as load bearing, tension or torsion? Consider using 100% infill so that the model ends up being one solid piece of plastic. Also, if possible, scale up the model. Very small or thin pieces will not be very strong, no matter what settings you use.

5. Keep in mind that 3D printing is far from being mainstream, regardless of what printer manufacturers would have you believe. If you own a 3D printer, you are an early adopter and are essentially beta testing and bug hunting for future generations. Also, it is not an exact science as evidenced by point #3 about extrusion rates. If it were an exact science, we’d have figured it out by now.

I also reduce flow rate to 15% with high quality PETG filament from Formfutura, for a balance of strong, quite accurate and beautiful parts. Correct filament diameter 2.85mm is set default in Cura, and I measured it with a calliper repeatedly.

With my lulzbot mini there was no E-Steps calibration label for the included extruder, so a default of 833 is used.

I bought a flexystruder v2, and with that came a E-Steps calibration value on a label on the extruder. The flow rate of 100% seems spot on there.

So I guess Lulzbot unfortunately did not bother with calibrating the standard extruder, at least at the time I bought the printer, but went with a generic value, and overshot to keep parts together and stuck to the plate with a wide variety of materials.

I also had to raise the print head 0.15mm (z-offset) above the default setting, so the first layer would not be squashed too much into the print bed (PLA, ABS, PETG). For the flexystruderV2 a z-offset of 0.1mm has been great.

Calibrate your extruder:

Make sure you measure your filament with calipers. Very very very rarely is a 2.85mm filament actually 2.85mm. If you don’t set this diameter correctly it will create poor quality parts.

Just to chime in on this I was printing with about 85% using Cura Lulzbot Edition and it was still over extruding by a lot on the first 15-20 layers. It would create a ‘slump’ or bulge on the bottom of all my prints and look quite ugly. Not to mention making it difficult with mechanical tolerances.

Yesterday I tried out the new Cura 2.5 and imported the same profiles but with 95% extrusion rate, and for some reason the slump is completely gone! I am wondering if it is some issue with Cura, rather than the firmware of the printer.

FWIW I used Simplify3d for quite a while with about 90% extrusion, and it was better, but S3d has other issues.

So if you’re having overextrusion issues I recommend trying Cura 2.5 - it worked for me, I don’t even have an elephants foot on my prints anymore.

I did the same with mine, I also wrote an esteps value of 810 (empirical, not measured) into the firmware because I was tired of always modfiying the profiles to use less flow.

I’ll have to try Cura 2.5. I did try slic3r but the resulting print quality was far behind Cura, even with my vain attempt to match profiles exactly.