Can't print at low extruder speeds

[jeremy 0]

Hi all,

I've had my Ultimaker for about a month now and have managed to make some pretty good looking prints with skeinforge at larger layer heights (.2 mm). However, with Netfabb, I can't seem to successfully print at the low extruder speeds that are required for high quality prints.

When I try the calibration cylinder on .4 width, 20 mm/s speed and .075 mm layer thickness, all I get is tiny blobs strung together. It looks as if the extruder speed is way too low.

When I try the same cylinder at the same settings except for 75 mm/s speed, it prints successfully and is .4 mm thick as it's supposed to be.

My extruder rpm for the successful 75 mm/s speed is 1.21 rpm, so it would seem reasonable to set the extruder rpm to 1.21 x 20/75 = .32 for the 20 mm/s speed, right? I've tried that setting and all I get is the mesh of tiny blobs and strings (I can upload a pic if its helpful). Even if I increase it to .6, I get the same result. At first I thought the nozzle might be jamming, but if I stop the print, I can push the filament through by hand fairly easily.

Any thoughts? My firmware is Marlin build 2 if that is relevant.

-Jeremy

[ddurant 0]

Have you tried viewtopic.php?f=9&t=194 ? The latest version is on page 4..

[owen 19]

Hi all,

I've had my Ultimaker for about a month now and have managed to make some pretty good looking prints with skeinforge at larger layer heights (.2 mm). However, with Netfabb, I can't seem to successfully print at the low extruder speeds that are required for high quality prints.

When I try the calibration cylinder on .4 width, 20 mm/s speed and .075 mm layer thickness, all I get is tiny blobs strung together. It looks as if the extruder speed is way too low.

When I try the same cylinder at the same settings except for 75 mm/s speed, it prints successfully and is .4 mm thick as it's supposed to be.

My extruder rpm for the successful 75 mm/s speed is 1.21 rpm, so it would seem reasonable to set the extruder rpm to 1.21 x 20/75 = .32 for the 20 mm/s speed, right? I've tried that setting and all I get is the mesh of tiny blobs and strings (I can upload a pic if its helpful). Even if I increase it to .6, I get the same result. At first I thought the nozzle might be jamming, but if I stop the print, I can push the filament through by hand fairly easily.

Any thoughts? My firmware is Marlin build 2 if that is relevant.

-Jeremy

I have had a similar problem. Try backing off the tension screw on the filament drive a little or raising your temp. I've went from 193 degrees to 213 and it seems to be consistent now.

[jeremy 0]

Thanks for the responses guys.

owen - didn't work for me. I actually had better luck when I dropped the temperature. go figure.

ddurant - I tried your utility and it recommended an extruder rpm of 1.06 for my settings (20 mm/s). That seemed high to me because it was almost as high as my successful 1.21 rpm @ 75 mm/s setting. However, I tried it and it worked - leaving me very puzzled. Then I re-tried the 1.21 rpm @ 75 mm/s setting and it also worked - however as I watched it, I realized that the extruder motor was actually turning way faster than 1.21 rpm.

So this explains why it prints successfully in both cases, but it leaves me wondering what the "rpm" value in Netfabb means if it is not the actual extruder rpm. Is there some hidden multiplier?

I looked through the gcode (specifically the last line of each file) and found:

"F4500 E3311.1677" in the 75 mm/s / 1.21rpm file and

"F1200 E2900.7219" in the 20 mm/s / 1.06rpm file.

...so this makes sense in a way because 4500 / 1200 = 75 / 20

and 3311.1677 / 2900.7219 = 1.21 / 1.06.

So it seems to be correcting the rpm for the additional speed... but what is the assumed uncorrected speed? I seem to have some misunderstanding about how Netfabb is calculating things here... maybe I should contact their tech support?

[ddurant 0]

So this explains why it prints successfully in both cases, but it leaves me wondering what the "rpm" value in Netfabb means if it is not the actual extruder rpm. Is there some hidden multiplier?

Yes and no.. And no.

Although there is indeed a hidden multiplier in my code (which is 12.461 - don't ask how I got this), the real issue is that the old-style number that netfabb and SF pre-40 use didn't really mean something specific. It was supposed to be the number of extruder steps that would cause 1mm of filament to come out of the nozzle.

The problem with this is that the length of filament you get out the nozzle varies according to nozzle diameter and temperature and what sort of filament it is, even different types of PLA have different characteristics and how fast you were extruding (aka: die swell) and 97 other variables. It's a very fuzzy value..

The new style of things (SF40+, slic3r, etc) specify filament by the length of raw filament stock to pull into the extruder. If the slicer knows the filament diameter, it can calculate the volume of a line it wants to lay down, do a bit of math then figure out, quite exactly, how much filament needs to be pulled in to make that line..

The magic 12.461 is what I came up with to convert new style to old style in my program.. It's actually doing new style calculations based on thread width & height and filament diameter then throwing in this magic number to make it work with old-style slicers.

The extra "and no" is because 5D machines (which we have) don't actually have a concept of extruder RPM any more.. That's more of a pre-stepper-extruder idea that's (thankfully!) long gone now. Netfabb hasn't updated yet, though. Eventually, once netfabb gets updated, there'll be no need to mention RPM any more.. It will tell (via gcode) the firmware how much volume it wants and the firmware will do the right thing to get the right amount of filament in the right amount of time.

[jeremy 0]

I appreciate your taking the time to explain this. I'm sure you saved me at least a few hours of head-scratching.

So if I understand correctly, your program does the volumetric calculation that SF40+ would do, and then spits out an old-school calibration based on that for Netfabb...

[ddurant 0]

Yup.