Random Pauses for no reason causing blobs

[leovinci 0]

Hi,

 

I have a weird issue that I cannot seem to resolve on my Ultimaker 3 Extended. I now verified it is not due to the slicing as I used Cura as Simplify3D and both have the same issue:

I want to print as smooth as possible vase/spiral mode, and it is crucial that the movement is as smooth as possible. Now I have at random a very brief pause of the movement which causes some extra material to leak out and create a blob. This seems to happen a bit at random , sometimes every layer onces, other times only after a couple of layers. There is NOTHING in the Gcode that could cause the pause. Believe me, I spent a considerable time investigating this.

 

It seems it is like a buffer thing or something, that he is waiting for new instructions. I print pretty fast, but when slowed down it is the same issue.

This is really causing me a headache and another dissapointment from using the Ultimaker 3 Extendend. I thought when I bought this machine I did not have to face these issues, and I paid the premium price for a reason... Even my 300$ Creality printer can do this kind of thing smoother.

Please help in guiding me. As I mentioned, I have the same issue with Cura or Simplify3d, so I don't think it is related to settings (which I already played with).  

I have the GCODE from simplify3D attached, which is very similar to Cura. For the issue during the spiral, scroll down to the middle. In the beginning I have different settings because I wanted a hole in the bottom.

monochromatic_lamp_150mm.gcode

[chexov 0]

Are you printing via USB wire?

[leovinci 0]

no, I print via the default memory stick

[robinmdh 100]

On 4/28/2018 at 10:08 AM, chexov said:

Are you printing via USB wire? 

This does not matter/USB would slightly more limiting.

 

The issue is likely with the high poly count of your model, which generates very short segments of gcode?

In a way all printers will have problems with this at some resolutions which is why a lot of firmwares discards too small moves and slicers have some limits, but short bursts of high detail are still very possible.

 

If so we're actually working on improving this but it's not a simple solution, bandwidth is arguably a finite resource in a system like this and it's made worse by EMC interference which is not the same for all users, so it'll take a while yet.

 

You could test with different firmware versions as we have been making changes to the communication protocol. perhaps V3.6.x is a good point to start in that case.

 

 

 

[chrisgr99 1]

I am also having this problem with my new UM3.  I just updated the firmware and am using the most recent Cura with default settings.

 

The head pauses randomly every few minutes for 5 to 10 seconds and extrudes a small lump of material at that spot.  My models are small with quite a few features but don't seem excessively complex.

 

I also notice that when the head pauses the display gets rotated horizontally so the right side is on the left and the left side on the right.  The display corrects itself after a minute or two until another pause.  Would a model being high poly cause this?  It seems like a bug.

 

The first time I noticed the problem was when I switched to 0.1 layer height.  Thicker layers had not shown the problem.  Could this be the cause?

 

What settings can I tweak to try to eliminate this?

 

 

[Nicolinux 288]

@chrisgr99 If I understand Robin's answer correctly, the model you are printing "might" be too high res. If you have created the model yourself, try exporting it with a lower resulution to an .stl file.

If it is a pre-made .stl file, then try another model (completely different model - just pick something on Thingiverse or YouMagine).

[gr5 2,173]

I agree - it's too high poly.  Most likely.  For example if you print a 1cm cube with rounded corners and there are 20 points on the corners - well that's a problem because marlin (the firmware inside the UM3) can only look at up to 16 points at a time and has to slow down enough so it can stop within the next 16 points.  If the next 16 points are within 1mm then you get these pauses where the points are too close together.  Plus the computer is working so hard weird bugs pop up like the display jumping.

 

If it's not your STL you can decimate it - try cutting the polygons by 3X:

http://www.shapeways.com/tutorials/polygon_reduction_with_meshlab

[robinmdh 100]

2 hours ago, chrisgr99 said:

I also notice that when the head pauses the display gets rotated horizontally so the right side is on the left and the left side on the right.  The display corrects itself after a minute or two until another pause.  Would a model being high poly cause this?  It seems like a bug. 

I've seen all kinds of shifts happen on the display, it is usually caused by a powerfull Electro static distcharge(ESD) but it can also be caused by constant EMC.

The fact that it happens every time is strongly indicative of nearby electrical disturbances, could be a microwave or anything that generates sparks, broken CLF tubes, that kind of stuff.

 

That same disturbance is also wreaking havoc on the communication channel between the real time controller(marlin/arduino) and the high level controller (griffin/olimex/linux) this lowers the potential throughput a lot, throughput required when printing more high-poly models but your EMC disturbance could make it problematic for even simple medium poly prints.

 

This can most likely not be fixed by firmware only more shielding. And any other 3d printers in that environment would quite possibly skip moves (not stop though, as most don't have error recovery on their read  from serial or SD card).

 

I suggest moving the printer to a different room or building and seeing if the problem persists.

 

P.S: (the display would likely shift left to right or top to bottom and then wrap around so that the edges of the screen are now in the middle, not rotated).

Edited May 21, 2018 by robinmdh

[chrisgr99 1]

Thanks everyone!

 

In my case it probably was that the model was too high poly .  I tried importing it into Meshlab and it crashed Meshlab.  It's a model I created in OnShape and exported in fine resolution.  I exported in medium and coarse and they also crashed Meshlab!  However a very similar fine resolution model has been successfully CNC milled by Protolabs

 

I then slightly simplified my model (removed some fillets), exported it at medium and loaded it into Cura.  I changed the layer thickness from 0.1 to 0.15.  It's now almost finished printing without a problem.  I'm not sure which of the changes solved the problem. 

 

I'm surprised even a coarse export of my model (as STL) would not import into Meshlab, considering that Meshlab is for simplifying and repairing models.  My model is not really complex compared to some I've seen although it does have a few sub-mm features.

[chrisgr99 1]

1 hour ago, robinmdh said:

I've seen all kinds of shifts happen on the display, it is usually caused by a powerfull Electro static distcharge(ESD) but it can also be caused by constant EMC.

The fact that it happens every time is strongly indicative of nearby electrical disturbances, could be a microwave or anything that generates sparks, broken CLF tubes, that kind of stuff.

 

That same disturbance is also wreaking havoc on the communication channel between the real time controller(marlin/arduino) and the high level controller (griffin/olimex/linux) this lowers the potential throughput a lot, throughput required when printing more high-poly models but your EMC disturbance could make it problematic for even simple medium poly prints.

 

This can most likely not be fixed by firmware only more shielding. And any other 3d printers in that environment would quite possibly skip moves (not stop though, as most don't have error recovery on their read  from serial or SD card).

 

I suggest moving the printer to a different room or building and seeing if the problem persists.

 

P.S: (the display would likely shift left to right or top to bottom and then wrap around so that the edges of the screen are now in the middle, not rotated).

I don't think I'm near any electrostatic discharging machines.  It's in a quiet residential workshop.

 

The screen wrapping always happened at the same time as a pause so they must be related.  The wrap would clear after minute or two (during the same print job) only to reappear on the next pause.  Could it be that a complex model is overflowing the processor's memory which is corrupting the display?  I also noticed when I aborted the problematic print job, the display was even more fractured than just being wrapped. I shut the machine down and restarted to clear it.

[robinmdh 100]

TL;DR -> @chrisgr99 If the model was to high poly for Meshlab then I guess that was it and the screen is unrelated.

 

Don't worry too much about the screen, if it goes back to normal after a minute or so then everything is fine.

 

You can be a great source of ESD event yourself, just FYI, your clothes etc can generate charges well over 5000Volts just with very few joules of energy. The sudden discharge of that can still break and disrupt electronics, as we have seen with the screen which is why it resets it's synchronization once in a while. This this is more likely in dry environments as moisture causes some random lower resistance pathways for those charges to safely dissipate trough.

 

And a resounding NO on the overflow thing, it just doesn't work that way. It might make the screen update slower but not nothing else as these parts are separate programs.

Edited May 21, 2018 by robinmdh

[chrisgr99 1]

It is pretty dry here.  I'll try printing again at fine resolution without ever touch the machine while printing and see if I get a good print.  It would be nice to print at fine resolution.

[geert_2 557]

Concerning ESD: maybe you could try to discharge yourself before touching the printer, for example by touching a metal plate connected to earth, or by wearing an electrostatic safety wrist band? I had a couple of times that the display and rotating button of my UM2 got locked up when simply touching the frame of the printer. This was in winter when it was very dry, and when I was causing huge sparks when touching anything.

 

And even in calm residential environments, you could have lots of ESD sources. Think of computers, 230V LED lamps, cell phones, wireless phones, wifi, TL-lamps, airco or fridge motors, solar panel converters,... Especially cell phones: try keeping them far away from everything else.

 

[robinmdh 100]

5 minutes ago, geert_2 said:

And even in calm residential environments, you could have lots of ESD sources. Think of computers, 230V LED lamps, cell phones, wireless phones, wifi, TL-lamps, airco or fridge motors, solar panel converters,... Especially cell phones: try keeping them far away from everything else. 

I hope you don't mind if I nitpick a little bit, what you describe here is EMC.

ESD happens when you're wearing your fleece jacket in the winter and you moving and walking around generates a charge via tribocharging . When that buildup charge suddenly discharges, that is a great example of an ESD event.

[geert_2 557]

9 minutes ago, robinmdh said:

I hope you don't mind if I nitpick a little bit, what you describe here is EMC.

ESD happens when you're wearing your fleece jacket in the winter and you moving and walking around generates a charge via tribocharging . When that buildup charge suddenly discharges, that is a great example of an ESD event.

Yes you are right. Thanks for the correction, I appreciate it. Abbreviations are not my strongest point...  :)

That said, both could cause problems: when my cell phone is laying next to the computer and it rings, this often locks up my USB mouse and keyboard (in addition to causing flickering in the screen). Then I have to unplug both and plug them in again. So I can imagine that similar things could happen in a 3D-printer.