gr5

This is an extremely common issue with PLA. It happens only when the walls are not vertical - when they overhang. Just like in the photo. This is very mild overhang yet pretty bad raised edges. I tried dozens of variables and the only one that makes raised edges better or worse is fan speed. You want the fan speed on 100% (which is default). So the first thing that comes to mind is to make sure your two side blowers are at 100% when printing. Cura by default turns them off on the bottom layer and then gradually turns them on and by about 1mm off the build plate they should be at 100%. You may have a loose cable. If you remove the rear two screws on the print head - the very long screws - everything is fine but now the top rear panel comes off (which holds the electric cable in place). Reseat that cable connector and then put it back together but add a shim - add some blue tape around the cable so it doesn't pull out again in the future. The other thing is bed temp. The sensors aren't always identical so I'd put a little blue tape on the bed of a good and this bad UM3 and set the bed to 60C and wait 5 minuts and use an IR thermometer to check that the temp is the same. It should be within 2C of 60C in the center of the bed. If one is hotter then just adjust the temp in cura by the error (if bed is 65C then set bed temp to 55C in cura going forwards).

Look not just at the label on the core but also on the tip of the nozzle. There should be a tiny AA or BB engraved in the brass. Here are instructions on how to reprogram what the core thinks it is: http://gr5.org/cores/

oh wow. Picture helps a lot. How big is this part in diameter? I'm thinking it's pretty small and the problem is it doesn't have enough time to cool. If it's smaller than 20mm in diameter then definitely try putting a rectangular "tower" on either side that is the same height or taller. While it prints the towers the fan will blow on this and it will have time to cool a little more. Putting the towers to the left and right is good as the cooling fans are on either side of the nozzle in the print head.

What kind of printer is this? Is this when the printer gets to the second or third layer? Or before the print starts? I'm going to assume before the print starts. On the UM3 and the S5 you can go to the material settings (for either left or right feeder) and tell the printer what material it is. I don't use the sensor - I just select a material, for example PLA.

Great post! Or... you could just lower the temperature if you keep getting ABS converting to that gummy material. And if you aren't getting good layer adhesion then you need to either increase nozzle temp or air temp or lower an speed.

Well it's basic underextrusion on the higher layers. So what's different starting around 2mm above the bed? I don't know. Typically Cura cranks up the fan and also the speed slowly over several layers. It could be that you are printing too fast on the upper layers but cura is printing slower on the lower few layers (which is default). What kind of printer do you have?

I'm guessing a lot of the end of the gcode file got chopped off and that's why it didn't work. Could it be that your disk is almost full? Or maybe the STL file is so large (millions of triangels) that it is using all your memory? how big (in bytes) is your STL file?

Okay well this is interesting - I'm guessing the end of the gcode file is missing - it should have a big block of comments at the bottom - each line should start with semicolon like this: ;TIME_ELAPSED:1623.651941 G1 F2400 E646.13004 G91 ;relative positioning G1 E-1 F300 ;retract the filament a bit before lifting the nozzle, to release some of the pressure G1 Z+0.5 E-5 X-20 Y-20 F9000 ;move Z up a bit and retract filament even more G28 X0 Y0 ;move X/Y to min endstops, so the head is out of the way M84 ;steppers off G90 ;absolute positioning M83 ;relative extrusion mode ;End of Gcode ;SETTING_3 {"global_quality": "[general]\\nversion = 4\\nname = Draft #2\\ndefin ;SETTING_3 ition = ultimaker2\\n\\n[metadata]\\nquality_type = draft\\nsetting_v ;SETTING_3 ersion = 5\\ntype = quality_changes\\n\\n[values]\\n\\n", "extruder_q ;SETTING_3 uality": ["[general]\\nversion = 4\\nname = Draft #2\\ndefinition = u ;SETTING_3 ltimaker2\\n\\n[metadata]\\nquality_type = draft\\nsetting_version = ;SETTING_3 5\\nposition = 0\\ntype = quality_changes\\n\\n[values]\\nbrim_line_c ;SETTING_3 ount = 4\\nbrim_width = 2\\nretraction_amount = 2\\nretraction_speed ;SETTING_3 = 40\\nspeed_print = 150\\nspeed_travel = 200\\n\\n"]}

So the M221 is telling the printer to multiply all the extruder moves by 104%. The 80% in cura affects the actual E values (the machine doesn't know about the 80% as the 80% part is baked right into the extruder positions in the gcode). And cura doesn't know what M221 does.

Again, no. It's not overriding. They are multiplying. 104% times 80% gives you 83% which is close enough to 80% that you don't notice the difference.

If you have more than around 10 moves in 1mm of distance most marlin printers (including Ultimaker printers) can't handle that and it stops while it loads up more moves into the buffer. Look at the stl file or the gcode file carefully to see if you have too many line segments in one spot. Some CAD programs are better at positioning the STL triangles than other programs. Or there might be some kind of "resolution" feature when outputting to STL. Maybe you have to lower the resolution so there aren't so many triangles. Or I could be wrong and it's a different issue.

I don't think I've done any PC prints yet - just PC + ABS blend. Ultimaker recommendations for PC are here - they seem to think PC doesn't stick well and recommend adhesion sheets - at least that would keep your glass from shattering. They sell adhesion sheets. They are a pain as they tear easily. When I use these sheets I print small parts in different locations tearing a small piece at a time and then after 10 or so tears I remove the sheet and start over. https://ultimaker.com/en/resources/22231-how-to-print-with-ultimaker-pc

Yes. It works great for most people. So you may have simply had a cable disconnect from the antenna during shipping or something like that. dev mode is fine and doesn't violate warranty. You seem to correctly understand that you have a "chick and egg" issue. Most of the network fixes involve connman but you can't run conman without ssh and ssh needs network connectivity. Alternatively you could get an "olimex serial cable f", remove the bottom cover, and connect to the olimex linux computer directly using a terminal program and run connman that way. But at least for me, it's so much easier to set up a temporary network connection using a discarded or used wifi access point (I have 3 in my attic collecting dust) or an ethernet crossover cable.

It's much better to let the part cool before removing the part. Sometimes even put the entire glass bed and part in the freezer but that's not necessary. Contact fbrc8.com for a new glass plate.

Probably not. Ultimaker is spending more time I suspect on the newer products and not the UM2 series. The bug you speak of sounds familiar but I forget what it is. The bug is probably fixed in tinkerMarlin. @tinkergnome - do you remember this marlin bug? Is this the one that is fixed by doing one of 3 things and one of the things is to add more comments at the bottom of the gcode file? If so, wasn't that fixed well before cura 3.3.0?

In fact I recommend tinker firmware which has lots of amazing features including "continue failed print" and other features even more valuable. https://github.com/TinkerGnome/Ultimaker2Marlin/releases

So I'm pretty sure 3.3.1 and 3.4.0 come with the same version of firmware for UM2+. That's why. It's confusing. I know. So I think it is working just fine. Not every new version of cura comes with a firmware update.

It sounds like you are pretty good with computers. My printer works with DHCP fine. There's a small possibility that your printer has a defective ethernet connection. I would strongly consider taking it home or something and see if it works there. Anyway if you can somehow get network access to it (for example using a crossover ethernet cable or by setting up a micro/private wifi access point only used by the um3 and one of your computers) then you can ssh into the printer (put it into developer mode first from the menu system) and you can setup the network anyway you want using conman. You can google about conman. I know you said "fixed ip is bad" but you should be able to tell the DHCP server to not send out a small range of ip addresses and you can assign one of those permanently to the UM3. Other than that - it should just work. You can get the MAC address off the network screen on the UM3 (it's kind of hidden in the name but if you look for 12 hex digits you will find it) and then maybe the dhcp server has notes about the "computer" at that MAC address in it's log. Another option is to set up the printer so it uses wifi and your "email address connection" solution using conman.

Is it brand new or used? I would get another card. Is the card shipped from UM or did you buy your own? Try to buy the exact same type that UM ships with. I recently bought these on amazon and they arrived but I'm not sure if I tested yet. No - I'm sure - I stuck it in the machine but haven't tested yet. My original UM printers came with 4GB SD cards. https://www.amazon.com/gp/product/B00FM5E1EY/ref=ppx_yo_dt_b_asin_title_o07_s00?ie=UTF8&psc=1 I would also try formatting the SD card using a windows computer or if you only have a Mac there's probably a way to format it in some windows format. I'm not sure what format is best but I assume possibly FAT32.

With the bed far from the nozzle do move material and move the filament slowly and make sure it comes out okay. You should be able to extrude some filament and it should be able to travel from the nozzle to the bottom of the printer in about 10 seconds (while slowly spinning the knob in move material). Your photo isn't very helpful as I can't see where it stops extruding. If it gets so thin and flat that the bottom layer is transparent then that usually means the head is too close to the glass. Just push down on the glass while it's printing and see if it suddenly all squirts out. If so you had bad leveling - nozzle too close to glass. If the filament isn't sticking to the glass at all that usually means the head is too high off the glass. That's my guess from the picture but it's hard to tell because you are only showing the good extrusion it seems.

You only need to attache the first 50 lines of gcode so you can just cut and paste it into here.

I'm surprised they are brittle. I would lower the fan. Listen to it and make sure it sounds like 1/4 as loud. It also helps to enclose the machine but lowering the fan should be enough. Usually when I use .25 or .15 nozzle I am printing very small things so I also print multiple objects to give one time to cool down (while the other is printing). But with nGen you don't need so much cooling so I suggest you just lower the fan quite a bit.

There are too many thousands of parts to list them all. Just email the reseller. If they refuse to order parts, tell UM you are having a problem with a reseller. All resellers are required to sell all parts.

higher temperature. Dont' try a lower temperature. 190C is a bit low for PLA usually. I found the quality of overhangs to be just as good at 210C and you are less likely to have underextrusion problems. Please provide a better photo (closer to the part so I can see the infill). Did it just suddenly stop extruding before finishing or did you abort it early or did it rip apart easily?

Ultimaker tried for years to get two heads to work in the UM2 and it worked - sort of. It would work sometimes. But they did a lot of testing and one of the problems was excess heat with two hot ends in there (pla wouldn't cool completely and would jam up in the upper end of the print head) and the other problem was that the inactive nozzle would leak filament into the print which is particularly ugly with 2 color prints. The solution worked - but not great. If you want a really good solution then go for the mark2 method where either head can leak a little without messing up the print. [edit below] Oops. I got DXU confused with the simpler dual dozzle um2 solution. DXU looks much better with it's ability to lift a nozzle and better heat control. This looks just as good as um3. Not sure if DXU or mark2 is better but both appear to be good solutions.