gr5

I can't think of anything that would cause the printer to slide off a table. I've put my printers on many different surfaces. I think he's in germany by the way sander. You should be able to recover with a heat gun or blow dryer but it will take an hour or two. Probably nothing will get damaged or maybe one or two wires. Take it one step at a time - see if removing all that gunk is enough. Or see if your reseller will do it for you. Some resellers are better (nicer?) than others. maybe they will charge you. Personally I would do it myself. Every part is for sale. Every. Little. Tiny. Part. You just need the part number. Not all resellers seem to know this or admit this. But once you get the part number they can contact UM and get it. So you could post a picture of a broken part here. But don't get a whole new head. It's probably fine once you get all that plastic out. And watch this video (not now - before you start the next print) so your part never comes loose again:

There's a print core cleaning wizard? I've been using my UM3 multiple times per week for over a year and I didn't notice? Was that there when it was still in beta? I thought I went through all the menus when I first got the machine. Maybe I just forgot about it.

Watch this video. It is long. But is as short as I could make it. I edited it down a lot. it explains why you probably shouldn't do raft but even if you stick to raft, it explains how to get your parts to stick so well that your next post will be "how do I get parts off the print bed? I tried hammers, I tried the freezer..." https://www.youtube.com/watch?v=t58-WTxDy-k

Also the instructions for cold pull tell you to put the head in the corner. I hope you did that because if you pulled hard enough to break PLA that's about 100 to 200 pounds (50kg) of force which is enough to bend your thin rods. hopefully you didn't bend your thin rods (the ones through the head). The idea is that you put the head in the corner so with 10 to 20 pounds force you don't bend the thinner rods through the head. I guess for us people who have been talking about cold pulls for 4 years we just assume you know what we are talking about. There are many videos that show how to do this including Ultimaker has some videos. The point of the cold pull magic temperature is you want the filament softened a bit so it can pull it out but hard enough that it all comes out in one piece.

35C is not the right temperature for a cold pull. For PLA the temp is around 95C. Between 90C and 120C for sure. I don't know what the temp would be for PVA and PC. Probably just above the softening temp of PVA which is 30C hotter than for PLA so maybe around 120C? It would help if you read about cold pulls in detail. So the procedure is to heat the nozzle well above the melting point of the materials involved - so 180C is pretty good for PLA and PVA. Then push in solidly to make a good bond. Then cool to the magic temperature. If you pull it out and it comes out with almost no force then you didn't cool enough. If it takes more than about 10 pounds force then you cooled too far. I usually go "too far" and then crank the heat up to say 130C while pulling upward the whole time until it suddenly gives. I note that temperature where it happened and next time start a little lower than that. For nylon I found 120C is about right.

I hear you but just play with it in cura and in layer height go to the color window and set the colors to show layer height and you can see instantly the range of resulting layer heights. So in other words instead of asking us, try asking cura directly and let us know what you learn. (please let us know).

cura positions the nozzle such that the line width will extrude right up to the edge of the wall. So for example if your nozzle is 1mm in diameter and line width in cura is set to 1mm then cura thinks of all lines as being 1mm in diameter. If you print a cube - or a layer of that cube that is 10mm by 10mm then it positions the X,Y such that it prints 0.5mm (half the nozzle width) INWARD from the wall. Such that the wall ends up being positioned perfectly. In theory. Plastic can shrink and other things can happen. But it's pretty close. So if you make these slots 1mm wide you will really get 1mm wide air gaps. If you make them 0.2mm wide there will indeed be a gap. But if you make them 0.01mm wide that is too small to create an air gap and the plastic will just flow into the "gap" and fill it all up. Does that make sense? Just try it in cad/cura and look at the result in layer view. Ideally I would do this kind of thing in openScad which is very parametric and programmable so you can have a parameter that says how far apart the slots are and subtract these slots from the part automatically and output a "grain controlled" part. But programming (and openScad) are not for everyone! Some hate it.

I'm not sure but I think "layer height" is what you are calling (rightly so) the "max layer height". disclaimer: I haven't played with this feature yet. But I did look over the code. [edit - as shown below - layer height is the middle value and then you can set the "variation" which varies equally above and below the "layer height"]

poke around. The code is not on github. I think. Not sure. But it's in /usr/share/griffin I believe. And subfolders. Maybe tar that up into a tar zip file and scp it back to your main computer? Or, you can just go to this page, download it and unzip it: http://software.ultimaker.com/jedi/releases/

Oh and I don't recommend auto level. The semi-manual process works better I think.

They (fbrc8) have records. Every printer has a serial number. They probably know which reseller got that particular UM3 so they could check their records. fbrc8 are great and honest people. With integrity. attention: @fbrc8-erin

AS A LAST RESORT, you can disassemble the core and work on the nozzle by itself. Either burn it all out at 300C (if brass starts glowing red hot then that's much too hot and let it cool as at that temperature it's extremely easy to squish the brass into a new shape). It's very easy to destroy the heat break so watch this video first:

Oh and EVERYTHING can be done in gcode. So there is no "and more". I mean you can home, turn power on and off, control steps/mm, basically you can do much more from gcode than you can do from menus. On um3 you can control color of lights from gcode. On um2 you can change the display message form gcode. For um2 the best gcode reference is here (just look at Marlin gcodes): http://reprap.org/wiki/G-code

Well the Um2 is simpler. You can control it through USB (it's not officially supported but it works pretty well). It has Marlin firmware which is on hundreds of different types of printers (most printers out there). There is software called "octoprint" that allows you to control printers through USB or you can simply send gcodes with any software you want to write. The um3 indeed has a linux stack. You can log in with ssh username and password is root/ultimaker. You have to put the printer into developer mode first (it's a menu option on the printer). Everything is in python so the source code is all visible and commented on the printer itself. Well some of it is in json files (javascript). Once in you can go to the /usr/share/griffin folder and do this: python3 command_util.py Then do "sendgcode" followed by the gcode you want to send. This might be too slow (not real time enough for you) so you might want to create some kind of web service. If you do *please* share. If you prefer - here is a video of someone doing just what I describe. There are more details on this forum but you can ask more questions on this topic if you want. For example the M104 and M109 commands might not work without enabling the hot end. That gcode is um3 specific and is somewhere on this forum. But here it is: M145 T0 -- enables ability to heat that core. Oh and be very very careful when editing json files. If you have a parse error in a json file (missing comma for example), the boot process halts *before* it starts the ssh daemon so you effectively brick the machine but I here are 2 methods for unbricking if that happens - you might want to buy these cables ahead of time if you will be editing json files:

The um2+ dimensions are roughly 210X210X210 - definitely not 300mm in any direction. And the UM2 extended+ is 300mm *high* not wide. You want to print this flat. Not vertical. If you print vertical the quality will get worse as you get higher up because it can wiggle a little. Instead you should get the um2+ and print it flat in 2 pieces and glue it together. Regarding fonts and such. There are many ways to create this - all professional cad programs can create this in any font but if you don't know how to use cad software you could do it 2 dimensionally in say microsoft word and save it as a png image and then you can load png images into cura and define brightness as height where black is thick or white is thick and the other color is height 0. It's very simple - cura takes in jpegs or png files no problem and leads you through the steps (cura is the software that comes with the printer).

Yes. You probably also have to adjust layer height and temperature. The whole point of using a 1mm nozzle is for speed. Half the speed benefits come from taller layer heights. e.g. 0.5mm or even 1mm layer heights. The other thing you might have to play with is temperature. I found for 0.8mm 3dsolex nozzles which are "race" meaning they have dual paths to heat the filament more thoroughly, I have to print 15C cooler. Initially I just played with this in the TUNE menu but I had to print a lot of things with that nozzle so eventually I adjusted all the temps in cura down by 15C (by my 4th print). Also I had 2 UM3's at the time and I was printing 100 of the same object in the same color with the same brand pla and I had an ultimaker AA 0.8 in the left printer and a 3dsolex with 0.8mm in the right printer. This is when I realized how much better the 3dsolex heats the filament. At least with the 0.8mm. I suspect the difference is smaller with the 0.4mm nozzle. BUT THIS DEPENDS ON LAYER HEIGHT and print speed and filament and filament color. So no matter what you do you will have to play with temperature. Even among UM filaments - the white filament I find I have to print cooler. So the profiles only get you so far.

Yes - the infill and wall lines would be .26mm apart but the amount of filament going through the feeder and out the nozzle is the exact right amount for .26mm wide lines. Similarly for "the other way around". That little nozzle is making fat lines - it takes a lot more pressure to do that but it works if you don't print too fast or too cold. I've seen a .4mm nozzle make 0.8mm wide lines. This is the default for the old cura in spiralize mode (used for cups and vases).

Lol - it must have gotten red hot. I use a gas flame to clean out my nozzles - on a gas stove. I always put a drop of water on whatever it is I'm heating up (hardcore block, olson block, block v3, nozzle, whatever). I count how many seconds it takes to heat enough for the water to sizzle away. Then I add another 1/2 to whole amount of seconds again. For example it might take 8 seconds to heat a nozzle so after done sizzling I heat for another 4 to 8 additional seconds. For an olsson block it might be 20 seconds so I add another 10 to 20 seconds. At this point I figure it's probably around 150C and ready for me to clean it out or unthread something or remove some pla or whatever.

I don't know the term "run out" but after watching the video I think you mean "error". Yes your method of measuring the rod bend is great and it shows where to push as well!

You should be able to print that part and have it MUCH better. This is PLA, right? Please verify. Please don't ignore this question and move onto the rest of my topic (this seems to happen a lot to me). Well it has very little to print on each layer so the nozzle is spending more time touching any given part of the print. Your basic issue is you need more cooling. 1) One way to achieve this goal is to get more fan in there. You are probably already at 100% so if you can get 200% that would help - add more fans, add a desk fan, blow on the damn thing. 2) Another way to improve this is to lower the nozzle temp. 10C at a minimum. Maybe 20C. Try printing 15C cooler or 180C whichever is hotter. 3) Another way - maybe the best way - is to print two of these at once - this lets one part cool a bit while the other is printing. Place them side by side along X axis so that fans travel over the "waiting" part.

Warping on the bottom is something I know a lot about. Are you printing on a heated bed with kapton? What temperature? You really need > 100C. I recommend around 105C or 110C (I mean the ACTUAL temperature measured by probe should probably say 105C so 110C is probably your goal temp). You also want to enclose your printer if possible and get the air temp up to around 35C. Also you need to use "ABS glue". Why ABS? Some people think ABS is better than PLA. It isn't. It's a very difficult material with properties that aren't all that impressive. If you need high heat sensitivity (if it has to withstand temps that kill people which many parts do) then there are better options (e.g. nGen). Here is a table with some material properties (strength, flexibility, temperature): http://gr5.org/mat/

This is incredibly common. It's pretty much always caused by a slipping belt. I'd have to see your printer in person I think but basically the problem is that a set screw on one of the pulleys is loose. That's what causes this on most printers. I don't even know if your printer has belts. That this happens in cura and not another slicer is not a major surprise. This kind of thing happens only on extra high acceleration and a combination of factors and features (bugs?) in Marlin can cause higher acceleration (Marlin has a max acceleration and then a MAX acceleration, lol - the lower max acceleration is rarely violated and the higher one is never violated). cura 15.* is a great slicer - I do not recommend you go to version 3.* unless there is a specific feature you really need in there.

I don't think you'll damage the glass any more than if you let it cool. I could be wrong about this. But you *are* more likely to damage the print. For PLA and ABS the bed is kept above the softening temp of the material meaning the bottom mm or so is quite soft and can be easily deformed. For PLA you want it to cool to 50C before removing. For ABS let it cool to 95C before removing.

Method A is to simply increase shell/wall width to an extreme - say 1 meter. Anything larger than the size of the part will do. The problem is you don't get the strength you wanted in the circled area. Maybe other areas. But you get strong walls around the edges - similar to your goal I think. Method B you put in these slots in cad. They are rectangular holes all the way through the part. But make them only 0.01mm wide. Small enough that when cura prints them it will seal them up solid across the gap. You might have to expleriment - they might have to be larger than 0.01mm. I haven't tried this but I've sliced other people's designs (emmet on thingiverse) who taught me this technique and it worked in cura at the time. You can shape these "holes" any way you want. You can even make them tubes instead of rectangular. They can be any shape. Cura will follow the shape as it thinks of them as walls. Note that the width is exagerated in my drawing. Note that I ended the holes before edge but you can even separate these into "separate parts" meaning have the slot reach all the way to the end. Possibly, ideally the holes should be internal only such that the top and the bottom prints a diagonal pattern that helps melt everything together. Like plywood with it's alternating grain. Basically you can get cura to make the grain go into any complicated pattern you want with these "internal holes".

This would be a good thing for cura programmers to look into. the best way to get that to happen is to get a github account and go onto the cura project and add a new issue with a link to your project file. Mention the slicing times for different nozzle widths.