UM2 Extruder knurled wheel slip issue?

[productdesign 0]

HI All,

I have not used the UM2 all that much since we got it. although it seems to have performed ok. not sure if its update / firmware related or cura related but part has really sparse layers. I upped flow and temp to PLA 230c. and reduced speed to %70. Build still came out patchy and weak.

I noticed a clicking sound coming from the extrusion drive while it was building. Was watching the knurled wheel and it almost seemed as tho the grub screw collar at end of knurled wheel was slipping on the shaft. Annoyingly you have to remove the whole assembly and have the motor fall off the inside wall just to tighten that little grub.

Tightened the grub and removed material > refed material and noticed that it was still slipping under load new material. Is this possible as its very hard to see with the naked eye. Also wasnt sure if UM2 has a built in load sensor or something that could make the extruder back off and auto slip?

I tightend the grub to the point where the grub stripped one time!

Any thoughts from the group as my prints on UM1 were way better.

Chris

 

[IRobertI 520]

Draw a line on the shaft and sleeve with a marker and you wont have to guess

What settings are you trying to print at? "70% speed" and "upped flow" tells us absolutely nothing. We need to know the speed, layer height and any additional tweaks you've done in the "Tune" menu during the actual print.

 

[productdesign 0]

Thanks for reply.

speed in cura was default : 85 mm/s a part at .25 layer .with 24% fill density.

The part was bad before any tweaks and not really helped by any tweaking of anything.

The knock sound sounds like a grandfather clock. defo coming from the drive. I assume this is not retraction but on the knocking sound the grub counter clock rotates 45 degrees. I think i can determine a difference between intended retraction.

Just disassembled the drive again and will put on some marker as you say. 95% certain this is the case.

Problem being if this knurled wheel IS slipping on the motor shaft and I have max torqued the grub what else should i do to ensure no slippage. might have been a better Idea for Ultimaker to have keyed the end of the shaft rather than rely on mild steel grub. :???:

 

[illuminarti 18]

The clicking is due to the extruder skipping steps when it can't push the filament hard enough. Several people have had this problem - there are several recent threads about it on the forum. It's not caused by the grub screw coming loose. That can occasionally happen, but when it does, the motor doesn't click, because then the motor has no resistance to overcome - it's free to turn, the filament just doesn't move.

That said, the settings you are using aren't any sort of default afaik - indeed, they're quite fast. 85mm/s with 0.25mm layers is equivalent to 8.5mm³/s; this is at the upper end of what a UM1 or UM2 can do, and are almost certainly the cause of your issues. Especially as the UM2 seems to need higher temps for a given flow rate, due to its smaller melt chamber.

The issue is that the extruder is limited in the volume per second that it can squeeze out. As the volume increases, so does the pressure, until it exceeds what the extruder motor can deliver - and then you can skipping. Increasing the 'flow' setting makes this worse, by directly increasing the amount of plastic per second that you are trying to print. Decreasing the speed helps reduce the volume per second.

What you need to do is to print slower, and/or thinner layers. Keep the temperature up at 230º. If you look at http://umforum.ultimaker.com/index.php?/topic/4586-can-your-um2-printer-achieve-10mm3s-test-it-here/, you can get a calibration piece that will let you see what the top end throughput is for your printer. Pay attention to your filament as well; tightly coiled filament off partly-used rolls seems especially susceptible to skipping (presumably due to the increased friction in the Bowden).

 

[productdesign 0]

Cheers illuminati,

That would make sense. I am only building at .25 to speed up the build. but working against me.

What your saying is it can build quicker using smaller layers at higher speed then?

Completely get the max volume throughput issue but seems odd the UM ships with 210 for PLA if we have to go over 230 to get a .25 build. why did they pull temp from the Cura UI?

So is the skipping intentional as a fail safe , I assume not if some folks get it and not others. if its not intentional, are you saying its the motor bugging out on load?

if the throughput vol is so intrinsically linked with build quality why does cura not show this as a quick calc. Just like the rev counter in ur car. you could see when your getting close to the red line. an analog dial would be good. as Im not going to have a clue otherwise. I know Cura does show red if your maxing it out. but there is a subtle difference in my suggestion.

 

[IRobertI 520]

You can put down lines faster at lower layer heights but the max throughput of the printer will still be the same. If you can print at X speed with 0.2mm layers you can then print at 2X speed with 0.1mm layers. So in the end the total print time will basically remain the same (acceleration, retractions and a bunch of other variables come into play, but simplified that's the case). That's why I asked for specific settings as that's the only way to know how much plastic you were trying to put through the printer.

Depending on what speed you're printing at 210C can work without issues.

Settings were moved so that you could easily tune things without having to re-slice.

The skipping is intentional. Instead of grinding the filament to dust or causing other issues the motor skips back.

 

[illuminarti 18]

The skipping is intentional, as Robert says, and is a mechanical rather than software effect; once the required torque to move the filament increases beyond what the motor can deliver, the motor doesn't move on command, but instead skips back to the last stable position of the coils.

The issue is that for some machines, the limit is reached at lower volume per second than for others. Exactly why this is the case is not completely clear at present.

Also, on the UM2 when the limit is reached, it's more obvious because the extruder clicks back, momentarily halting extrusion, and making it very obvious. On the UM1, the filament would just move slower, leading to some percentage of underextrusion, and eventual grinding of the filament (see my tests http://www.extrudable.me/2013/04/18/exploring-extrusion-variability-and-limits/). So it was less of an issue at first, and harder to spot.

On the UM1, the effective extrusion limit before prints started to fail was somewhere in the 8-10mm³/s range. It varied for those machines too, in practice. The extrusion test piece I created that is linked in that other thread was designed to offer a comparison with the UM1's capabilities. Lots of UM2 printers - including mine, and a bunch of others I've had access to test - are able to print at 10mm³/s without a problem.

 

[productdesign 0]

Hmm. well my point on Volume throughput is it would be good for cura to calculate this and reflect it in an analog dial with amber and red line. I noticed it does flash yellow and red if you go high on settings but those I was using did not flag either.

I have stated on the thread illuminati gave, that this could be update related. I did not used to get this issue. Many folks saying UM2 used to not skip. and does now.

On removing the drive case its clear there are areas rubbing, which is bad design. leading to people discovering over tightenin the case or loosening had an impact here. the concept of applying tension on that load wheel can only increase the offset in x-y and increase rubbing if you look through the thing its all very tight.

If I may throw this back at Ultimaker, as mine is under warranty. Lets get away from user generated problems and slicer impact. Just using standard UM filament and a UM2 with the latest firmware, on changing filament this generates the skips. When you use the auto feed / change material as the material starts to flow from the nozzle. I suggest those who need to fix this, use that as datum. what ever the vol UM is set to push through on reload, is this expected to cause skips? would seem odd if that were the case. Im looking for a review of the firmwares involvement in this issue, who do I need to speak with, dya reckon?

Cheers

Chris

 

[IRobertI 520]

During filament change/load the current to the stepper is reduced to 2/3 of the normal current, so yes, skipping during the filament loading procedure is expected and intentional.

 

[productdesign 0]

right....ok. What is the conclusion as to a fix for me then? I will say although the spool im using orange UM PLA. is full ish. there does appear to be an undulation in it. also note their is a fair amount of friction just in spooling round that turn adapter ( material binds assume its abs? )

 

[illuminarti 18]

Again, as Robert says, it is intended to skip, to prevent the filament from stripping. The filament feeds much faster than normal printing speeds - it would take an age to load otherwise. So the current is lowered, so that the motor will definitely skip, rather than damage the filament.

 

[illuminarti 18]

I'm not yet convinced that you even have a problem, Chris. Did you do the extrusion test print? If your extruder is only failing when printing over 8mm³/s then I'd say your printer is operating in the normal range - towards the lower end of the range, yes, but its still not disastrous.

Tightness of the filament on the spool and any unevenness in the material can definitely lead to binding, making the extruder work harder, and so lower throughput. You can mitigate this by putting the filament on a low friction spool, or simply putting it on the floor under the printer, with lots of loose uncoiled plastic for the printer to feed on.

 

[productdesign 0]

will do the test when i have a mo. but i didnt follow how folks were varying the speed was this via tuning on the machine? i always wonder how people do those multi setting prints in one. we often see in troubleshooting. i take it its a file and you have to watch the layer change to the indicated vol and tweak it as such...im not that clever

 

[illuminarti 18]

No, you just print the gcode file. I already built the speed changes into it. Just copy the gcode to an SD card, and print it. The only thing you need to do is tweak the temperature setting on your printer - I recommend printing the test at 230º, if that's not the default already on yours.

 

[productdesign 0]

I see. could you point me to the G code your referring too. and your saying save it from something like notebook?

As you may have guessed only used cura in my experience to date.

thanks for the help,

PS dropped the speed and layer height and printed better. Does anyone get the benefit of the volume calc being visible as an on the fly instrument. i have never as yet performed a vol calc while 3d printing and kinda feel things should be progressing the ease of use way. im not a hobbyist, i need ease of use and reliable prints. Tho I take my hat off to all yall who have helped make UM what it is.

 

[illuminarti 18]

Here's the original post with the test piece:

http://umforum.ultimaker.com/index.php?/topic/3976-almost-always-missing-layers-underextruding/?p=33427

Save the file, and copy it onto the SD card to print it.

I totally agree with the value of having the volume calculation monitored in the slicer: it's something I've advocated for in the past. In Kisslicer, for instance, you can set a maximum volume per second limit, and the slicer will adjust the gcode (by slowing down linear speed) to ensure that the limit is never breached, even when it's doing complicated things that increase the throughput in less-than-obvious ways, such as printing multiple layers of infill at once, using a different speed from the regular one.

 

[productdesign 0]

I have performed the test. Thanks for making that possible. I attach the image. The knock sound I referred to is occurring and causing this under extrusion. So someone said they are not convinced there is a problem, is that still the case?

It looks like Under extrusion is occurring before the 4mm 3/s

This is way off the 8-10 referred to as normal...would appreciate the help guys getting this chap working? If its a voltage thing can someone help me sort that?

 

[illuminarti 18]

Yes, that's definitely not right - you should be able to go quite a bit faster. Did you raise the extrusion temperature to 230º?

The knock sound isn't really the problem, but a symptom; there's some additional resistance somewhere that your filament is having to overcome, and that's overtaxing the extruder motor, and making it skip.

One thing I would do is review http://umforum.ultimaker.com/index.php?/topic/4586-can-your-um2-printer-achieve-10mm3s-test-it-here/?p=40399, and then relax the 4 thumbscrews on the head slightly, and reseat the Bowden tube, as I describe there . Then also relax the torque on the screws holding the extruder on the back - they should be holding the black plastic parts against the body of the printer, but not squeezing or deforming it.

Try that and then retest it (making sure you're at 230ºC).

 

[productdesign 0]

Thanks. Yes printing at 230. Have disassembled the drive unit and as described the design has areas which rub slightly. reseated the bowden. made sure the mount screws are not tight. have not seen these head thumb screws yet so will try that... What about the motor voltage discussion.

 

[illuminarti 18]

While the current setting for the motor could potentially cause some quality degradation, I'm not convinced that it's the most likely cause of significant under-extrusion. Start with the mechanical issues, and we'll work up to electronics.

The screws in question are the four long ones with a knurled bit on top that run the length of the print head. You need to adjust those before seating the bowden.

 

[productdesign 0]

the thumb screws were tight. so loosened off. The resulting print starts under extrusion in the middle of the #4 . so a fractional improvement perhaps. I am reluctant to mess with the hot end unless you think i have too? Bit confused why this wasnt picked up before shipping as its a quick calibration test you have the gcode for.

 

[illuminarti 18]

If you remove the filament from the printer, and then remove the bowden tube from the head, can you run the filament up the Bowden tube by hand from the 'wrong' end, easily, without encountering any obvious resistance?

I'll also send you a different version of the extrusion test to try.

It wasn't picked up before your printer shipped, because I only developed the extrusion test fairly recently, and it's only more recently that UM has been playing with it. However, it is something they are testing with fairly extensively now, I believe.

 

[illuminarti 18]

See PM in top toolbar of this page.

 

[norsac 0]

I also have a very similar problem and have had ongoing underexrusion. I have pulled apart the entire feed assembly 3 times now and cleaned to see if I can get it to work.

I am attaching a example photo using illuminarti's test gcode at 230

 

[joecross 0]

I am having a similar problem, but only with certain filaments. At 40mm/s and .2mm height with 1.75 PLA, some filaments slip intermittently. I thought it was a mechanical problem with the extruder except that the filament sent with the printer NEVER slips.

Is it possible that the filament I bought is inconsistent in width and sometimes is more than 1.75mm so that it overloads the flow rate of the extruder? Or maybe it is less than 1.75mm and is too thin for the extruder to grip onto?

Any help appreciated.

Joe