Very high E steps value

[Nicolinux 288]

I am trying to calibrate the extruder but I stumbled on a weird issue. My default E steps value was set to 836 (rev4 kit). I measured 150mm from beneath the extruder and marked it with tape. Then with Printrun (or Pronterface), I set it to extrude 100mm. The weird thing was that it extruded only 82mm. That's pretty far off from 100mm. I calculated the new E steps value to 1019.5 like this: (100mm/82mm)*836 = 1019.5. With the new value the new extrusion accuracy was way better - 98mm.

How can this be? Isn't this value way to high? I didn't take off the bowden tube though. I thought that under normal circumstances, the bowden tube is attached and the measurement should be made under "normal circumstances".

The reason I started to tweak the E steps was because some printed parts feature very prominent blobs. I thought that might be a sign for overextrusion.

With the old E steps value (836) I also tried to extrude directly via gcode following the advice in this post:

http://umforum.ultimaker.com/index.php?/topic/1647-extruder-calibration

However this failed because the hobbed bolt just ate up the filament and didn't move the filament very far. I used the following gcode(s):

 

G91  ;relative positioning
G1 E100 F300 ;extrude 100mm at half max speed
G90 ;absolute positioning

 

[illuminarti 18]

Were you pushing filament out through the nozzle when you were testing the extrusion amount?

If so, and you were moving it quickly, then you would have created huge back pressure, leading to filament slipping and under extrusion. Bear in mind that in normal printing, the filament is advancing at 1.5mm/s or even less. (1mm advanced pushes about 6.5mm³ of plastic, so 1.5mm in a second is the getting up towards the 10mm³/s limit). That's equivalent to an F rate of 90 (mm/min).

When you are testing to calibrate E, you need the nozzle hot (so as not to trip the prevent-cold-extrudes) logic. But then use fresh plastic, and insert it so that it's only just going into the Bowden - so that the extruder can advance it without having to actually push it through the nozzle.

 

I am trying to calibrate the extruder but I stumbled on a weird issue. My default E steps value was set to 836 (rev4 kit). I measured 150mm from beneath the extruder and marked it with tape. Then with Printrun (or Pronterface), I set it to extrude 100mm. The weird thing was that it extruded only 82mm. That's pretty far off from 100mm. I calculated the new E steps value to 1019.5 like this: (100mm/82mm)*836 = 1019.5. With the new value the new extrusion accuracy was way better - 98mm.

How can this be? Isn't this value way to high? I didn't take off the bowden tube though. I thought that under normal circumstances, the bowden tube is attached and the measurement should be made under "normal circumstances".

The reason I started to tweak the E steps was because some printed parts feature very prominent blobs. I thought that might be a sign for overextrusion.

 

With the old E steps value (836) I also tried to extrude directly via gcode following the advice in this post:

http://umforum.ultimaker.com/index.php?/topic/1647-extruder-calibration

However this failed because the hobbed bolt just ate up the filament and didn't move the filament very far. I used the following gcode(s):

 

G91  ;relative positioning
G1 E100 F300 ;extrude 100mm at half max speed
G90 ;absolute positioning

 

[gr5 2,172]

printing at 82% is about right for a very fast print. Illuminarti was able to get it down to 72% with very high pressures (fast extrusion rates). See this article for an explanation - especially look at the graphs (I like graphs) but if you really want to understand it you should probably read the whole article:

http://www.extrudable.me/2013/04/18/exploring-extrusion-variability-and-limits/

 

[Nicolinux 288]

@gr5: What du you mean with 82% - of what? I can't read the article right now because it seems to be down for me (from Germany). Will try later.

@illuminarti: You are right, I was pushing the filament through the nozzle. I will insert it only a little into the bowden tube and try again. By the way, I read somewhere that you shouldn't leave the nozzle hot for too long without extruding because it could lead to plugs. What would be the right temperature and the maximal time to leave hot?

About the image - is this really a sign of overextrusion? I printed this object with fairly standard settings (0.2 layer height, 50mm/s, 220°C, minimal layer time 5s). Retraction is always enabled with the defaults (40.0mm/s and 4.5mm distance). I have the "latest" official firmware from march 2013. I know about the bug with the halved retraction speed and I will take care of re-building the firmware later. For now I'd like to get some acceptable prints. After the z-wobble issue, the blobs are the biggest issue for me now. Especially because it is not consistent. Some prints do feature it, others don't. I guess I really need to take a more methodical approach when testing and write everything down...

 

[makingzone 2]

@Nicolinux

I've got the same kind of calcul when I wanted to calibrate my Ultimaker. The mistake is coming from the extrusion speed during your calibration. If you extrude too fast, the nozzle resistance cause some sliding between the filament and the extrudor.

For better calibration, you have to extrude at the same speed (approximatively) as when you print.

 

[illuminarti 18]

George is referring to to 82% of the 'expected' extrusion rate - your 82mm when requesting 100mm. That much under-extrusion, and more, can be caused by pressure alone - even if your actual non-extruding steps-per-e is right.

The website should be up and working fine - please let me know if you're still having problems, and I'll beat up my hosting company. :-)

Regarding prolonged heating - it certainly can cause some problems - particularly with very high temps - and/or exotic filaments like PVA or Laywoo-D3. With some types of PLA, I have occasionally seen tiny amounts of cooked PLA be enough to stop oozing, but not enough to resist being pushed through easily when the head is primed, but some PLA's might be more unhappy about sitting, I guess - depending on the colors, additives etc. In general, i think it's more a case of 'nothing good is going to come out of it, so avoid it when you can' - although I often preheat my head to change plastics, and then end up leaving it sitting for quite a while when I get distracted by shiny things before the SD card can make it back to the printer.... and I've never had problems.

A more likely source of problems related to all this - and one that I HAVE had happen several times, until I learned not to... is blockages due to trying to feed way too fast. This can happen as in your calibration test above, or just when using the Ulticontroller to advance the feed to prime the head. It tries to move waaay to fast, and the pressure spike can squeeze molten plastic back up into colder parts of the hot end where it solidifies.

 

@gr5: What du you mean with 82% - of what? I can't read the article right now because it seems to be down for me (from Germany). Will try later.

@illuminarti: You are right, I was pushing the filament through the nozzle. I will insert it only a little into the bowden tube and try again. By the way, I read somewhere that you shouldn't leave the nozzle hot for too long without extruding because it could lead to plugs. What would be the right temperature and the maximal time to leave hot?

 

[gr5 2,172]

I guess I really need to take a more methodical approach when testing and write everything down...

 

Yes! I keep a notebook and pen under the print bed. When I start heating the nozzle I also write down the date and what I'm printing. I usually mention temperature and feedrate as those are the main things that aren't obtainable from the gcode (the gcode always stores all the settings in the last line of the gcode file).

My Cura print temp is always 0 which makes it much easier - I set the temp manually and can start the print when the temp is "close enough". I don't have to wait for the temperature to settle.

170C is the minimum temp for the extruder to work so I would set the nozzle to 175C or 180C. I wouldn't recommend leaving it like that for more than 5 minutes but I have many times and nothing bad every happened.

 

[gr5 2,172]

But none of this is what you want to know!!

What is wrong with your print???

I'm not sure.

1) Was this printed by itself or with other objects at the same time? If the answer is "other objects" then almost guaranteed that is "stringing". You can verify by looking at the gcode view in Cura and there should be a blue line right at the exact same location of each blob. You can fix this by printing one object at a time. There are other ways to fix stringing but the best is method in this case is to print one object at a time.

2) It could possibly also be caused by over extrusion or over pressure - sometimes the pressure is too high and the UM keeps printing just fine until suddenly it all leaks out in a burst/blob. This could be caused by a bad Z movement - maybe it is sticking and some layers are less than desired so there is less space for plastic and over pressure (over extrusion).

3) It could be infill. This also can be checked by looking at the layer view and see if each blob lines up perfectly with the infill pattern. Even better, print it again, look at it when it is half way through and see if the blobs line up with the infill. If this is your problem, increase the shell thickness to .8 or 1.2mm to prevent the infill from reaching past the outer skin. This "extruding infill" is a problem with the latest cura (steam engine). I have some details on it here:

http://umforum.ultimaker.com/index.php?/topic/2501-is-this-slicing-settings-or-hardware-problem/?p=17683

 

 

[Nicolinux 288]

@gr5: Thank you for the ideas. I have created another post and added your answer there in order to keep the topics separate.

http://umforum.ultimaker.com/index.php?/topic/2786-please-analyze-my-prints

A little update:

I measured the extruded filament while it moved through the bowden tube only (and not the nozzle). It was only 1mm off so I calculated the new value and set it to 844. I have also calibrated the x, y and z steps and now my calibration cubes are spot on when measured with a digital caliper So this problem is solved now. Thanks for all the help guys!