gr5

There's a second good reason for this. The feature is called "horizontal expansion" and for pva support it defaults to 3mm. While PLA sticks on top of PVA nicely it's not so good the other way around. So if you are printing something shaped like a 4 legged table where legs are touching the build plate (a stupid orientation but whatever) you don't need horizontal expansion - there's lots of pva to hold up the flat part of the table. But if you are printing a short tube with the hole being horizontal - the PVA doesn't stick well on top of the PLA inside this tube - you really need the PVA to be all connected together and to the glass plate to hold it all in place - so horizontal expansion has the PVA sticking out the sides of the tube by 3mm and then going all the way down to the glass. This is a very crude feature that could be much smarter but it works for "most" designs.

Cura programmers use Ultimakers which use 3mm filament. Maybe the estimation is wrong because you have 1.75mm filament? I mean cura is written to use ANY size filament - but maybe there is a bug with 1.75 and amount? the crazy thing is cura creates gcodes which tell the extruder EXACTLY how many meters to feed - I mean down to a hundredth of a millimeter. So you can get how many meters are going to be printed by looking at the gcode file and scrolling through it (although the E value goes back to zero occasionally as Marlin can't handle E values larger than a certain amount of meters (10? 100? 16784mm? 32768mm?). When you said the amount was wrong was it the amount in grams? or the amount in meters? And you are talking about when you are still slicing - right? Not when printing?

Selecting AA 0.4 is fine. (note I am the distributor for 3dsolex for usa). I have no idea what's up with autoleveling - it should work fine. The UM3 has a sensor in the head behind the nozzles at the base of the head and it keeps moving the bed up until the capacitance stops changing which is when the nozzle just touches the glass. Then it knows exactly the point where the nozzle touches the glass. I don't know how this could screw up. Grasping at straws here but maybe your head door isn't closed all the way making the capacitive part not parallel with the bed. I had to push up on the wires a bit such that when you close the head the hardcore wires are out of the way. I doubt this is it. Personally I have never used auto calibration. I adjust the z height when printing the first layer - I turn the 3 screws until it's perfect (I do this during skirt or brim). After a few prints I can then go months without touching the screws or doing any kind of calibration. 1) If you push down on the bed while it's doing the bottom layer - does it suddenly all squirt out? If so you are indeed too close to the glass. Just turn the 3 screws tighter to lower. If pushing down on the glass results in nothing still coming out of the nozzle then autolevel isn't the problem. 2) It could be that leveling too close has caused the filament to grind back at the feeder. I would cut 5cm off the end of the filament and try again. 3) There could be something wrong with the hardcore - I have no idea what as I have inspected hundreds. but you could just send it to 3dsolex and ask for a replacement. Carl/swordriff will replace it regardless of who your reseller was: attention @swordriff

This. This is a good idea - but maybe Y=180? Y=100 is near the center. I like it best in the rear center. In either of the rear corners it won't squish well and it will fall over.

The board will be fine. It's your power supply that may trip (turn off for 1/4 second and cause your printer to reboot). The power supply will be fine also. It's your print that will get ruined.

Another cool trick would be to print peek as support material except for the interface layer. Then put a layer of PLA way up high in the print where the support meets the main material. It would be a thin layer interface. Then when done you could heat the whole print to 100C and the supports will just fall off. I don't think cura supports this feature but people have asked for it - they want 90% of their support printed with PLA and only the interface layer printed with PVA. This doesn't work well with PVA as PLA sticks well on top of PVA but PVA doesn't stick so well on top of PLA. With nylon and PVA I think this would work great. But again - not sure that cura supports this idea.

You could try it if you have two print heads working. Go to the lower graph on this page: http://gr5.org/mat/ Look at the materials in orange - the right most ones are the potential materials - those that have a softening temp around 120C are probably ideal - so PC, PP, Nylon 645 or greentec are candidates - materials worth testing out as support for peek/ultem etc.

If you can somehow attach the power cord to the table that should be enough to keep the printer from wandering far. Or some thin rubber pads under all 4 "legs". Any individual part you need replacing you can get - including side panel or - well - anything. Everything has a part number so that is the hard part - once you have the part number contact your reseller and ask them for a price. I'm guessing maybe you used igo3d. They can get any part once you have a part number. If you are having trouble getting the part number you can ask here on the forum. igo3d might be able to get you the part numbers or you could potentially contact UM in Netherlands or contact fbrc8 (fbrc8 is in usa but they are very nice and can probably get you a part number - they assemble UM printers for the USA market so they are very knowledgeable but of course UM Netherlands is even more knowledgeable).

Okay - we need to change the topic of this thread to "suicidal Ultimaker printers". lol. Sorry! We are all engineers here and love puzzles and this is an interesting puzzle and beyond most of our experiences. As far as getting parts to stick you should really watch my video but cleaning the glass, using glue, running leveling is not enough. You have to squish it real well - squish the first layer. The way I do that is to never do auto level and to also turn the 3 screws a bit extra to really squish down the first layer. There are other ways to do this but that is my method and my parts never come loose. If you watch the video you can see me picking up the entire printer by just a little UM robot on the glass plate. I have been studying this issue (getting parts to stick) with the scientific method doing lots of experiments and 4 years of experience. Watch the video please. Now that your printer has fallen the gantry may not be coplanar. It may be sadlle shaped. Another way to think of this - one corner of your printer may be higher than the other 3 (or lower). Another way to think of this - think of the gantry as suspended by 4 points in the 4 corners (an approximation). One of those 4 points may not be in the same plane as the other 3. This causes the print head to move in a saddle shape. This may cause some of your glass bed to seem very level but some of the corners to be too high or too low. Just warning you of what to expect. One solution might be to slip a coin or something under one of the corners of the glass plate as glass bends 1mm or so without very much force needed and maybe you can compensate that way. One solution might be to bend your entire frame back to planar. Or maybe your frame is just fine as it is - I don't know - but I'm warning you of something to look out for now that you have a suicidal wandering printer.

I'm sure it did not trigger one of the end stops. It makes this crazy loud noise like your printer will self destruct - but then it's fine, right? You probably need to move one of your end stop limit switches a tiny bit. You can either bend the metal tab on the switch or you can loosen the 2 screws holding it and slide it closer to where it will trigger sooner. First figure out if it's X or Y. Push the head around and listen for where the two switches "click". One of them probably needs more force than the other - that's the problem axis. If it's the X axis also check to see if your fan shroud is maybe bent and hitting the left wall of the printer - if so just bend the fan shroud back.

You could maybe use ABS as support material and melt it away later? Except with a bed of 130C the ABS will slump and not work. Hmm. PLA and PVA will be even more liquid than ABS at 130C. Maybe ABS will work anyway? Not sure. Ultimaker PP has a softening temp around 115C so maybe that would work - not sure.

This is - well - unprecedented I think. Could you try running the model through this free service first just to see what happens?? https://service.netfabb.com/login.php Also could you post your STL somewhere on the internet (actually ultimaker might let you upload it - not sure on this new forum). What cad software are you using?

Build plate at 70C is dangerous. having your print stick to the glass is more important than a little "cave in". I think. Unless you are just printing a small 20 minute print.

Maybe sell your 2+ and get two um2gos? (just kidding - mostly)

Why a VM? I run Cura in debian linux and it works fine. Are you running some form of linux?

I'm not sure but I think this means the first line in the gcode file is corrupt or isn't gcode. I'd look at the gcode file in a text editor and see if the first line is messed up. Maybe reformat the usb stick or try a different usb stick. For networking it really helped me to have my router assign a fixed ip address to my printer - there is a DHCP section in the options for most routers and for me it was trivial as it already had an IP address and it said "ultimaker" next to it and I clicked one button to make it static. I also wrote down the ip address and told cura connect about that ip address. This seemed to help things work better. Wait - wireless? Are you wireless? If so you have to use your cell phone to connect to the printers wifi service and tell the printer your wifi password so it can get onto your network - that might be the problem also (the network problem - not this usb issue).

Which version of Cura? Did you select machine to be Ultimaker 3? Does it show Ultimaker 3 in cura as the machine towards the top right corner - black area?

And +1 for having a different minimum layer thickness for overhangs versus top surfaces.

If you have tinker marlin installed it has a feature called "power management". You tell it how much power the bed and the nozzles use and what your total budget is and it will keep your printer from going over the budget - the printer already turns these 3 heaters on and off 20 times per second but with power management it makes sure the average "on" time in a given cycle (about 1/20th of a second) never exceeds the budget. The bed gets lowest priority. It's true that when all 3 are on you may go over budget for a few milliseconds but the power supply has a large capacitor (all power supplies do) which acts as a kind of battery that barely notices if you draw too much current for 50ms. Of course if you draw too much for an entire second it will shut down. Power budget works great. Dual printing on a um2 with both nozzles in the same head - well that does not work so great for PLA but should be fine for high temp materials like ultem and peek.

In normal view were there red areas there? If so play with the support angle - sometimes that fixes it. Cura definitely gets confused on what areas need support and what don't. I'm not sure why but I think it has to do with the "surface normal" in STL files. Many CAD packages don't set that value properly and this messes up the algorithm that tries to figure out where support is needed. Possibly it might help to run the model through a mesh repairer like the free one at netfabb here: https://service.netfabb.com/login.php

Oh - and love the sophisticated cat.

The slop calibrator is to see how much smaller a peg has to be to fit in a hole. If you make the peg and the whole both exactly 1cm it might not fit so this print has the holes bigger and bigger with a wider margin all around to determine what will fit. There is a feature called "horizontal expansion" that can be set to negative numbers (for either pegs *or* holes) to give you a little more "slop" or "play". For example you could set horizontal expansion to -0.1 and it would make everything smaller in X,Y by 0.1mm -- vertical holes would be larger by .1 and outer dimensions smaller by .1. I'm not sure though how you take there results of the calibrator "help convince some parts..." -- LOL!! I would write to the author of this design and tell him/her the result of your slop test and ask if that means you should try to increase or decrease the parts based on your result or if that's the same result the designer expected and no adjustments necessary. Let's see - the warping. That small part has a bowed in side. I assume you mean that? What happens is that liquid CPE sticks to itself and acts like a liquid rubber band as it comes out of the nozzle. Like snot. The tension is high-ish (not loose) because it shrinks a bit in milliseconds and is already pulling gently against itself. As you go around an outer corner or curving wall (or inner hole) it gets pulled inward. Particularly at corners. Each successive layer is pulled in further and further until you get far enough from the toasty heated bed and the lower layers are more solid and it slowly corrects itself until it is back to nominal dimensions. You can "fix" this by lowering the build plate temp which is fine for tiny parts but for large parts is a hugely bad idea. The main reason for keeping the bed at 80C is because this is slightly above the softening temp of CPE such that it can flex slightly. This helps to prevent corners from lifting off the bed. If you are printing multiple parts you are less likely to get this "warping" anyway because there's more time for the lower layers to cool. If you are just printing one small part then go ahead and lower bed temp to 60C. Or better yet - never print just one small part. Here is a visual guide to this very issue under "cave in": http://support.3dverkstan.se/article/23-a-visual-ultimaker-troubleshooting-guide#wallcavein So... how strong was your part? Did it break on layer lines or randomly? If it broke on layer lines then your fan may be too high and layers might not be bonding (melting the layer below as they go down). One fix is to lower fan to about 1%. This is never a problem for PLA but for higher temp materials like ABS is a common issue. Note that fan at 10% is the same as 100% for UM3 so you need to go down to 3% or less for ABS. Not sure about CPE - I always just left things at default. CPE is much easier to print than ABS but harder than PLA.

I am on 3.1.0 on ubuntu and it works for me. You have to set the execution bit. Did you do that? In terminal window: chmod 774 Cura-3.2.0.AppImage ./Cura-3.2.0.AppImage

This is a common problem and easy to fix (well I'm an electrical engineer). Basically there is an increased resistance somewhere in the path of the temp sensor which starts at the circuit board under the printer and goes to the bed through obvious wires (in that black sheathing) and through a connector to the board itself. The board is extremely unlikely to fail - the most common failure point is at either connector - wires can come loose and the connector itself can disconnect (barely) from the board. I'd start by watching the temp while you poke around with your fingers where the wires connect to the board - it's at the back of the heated bed - 4 individual wires - the 2 smaller ones are the sensor wires. Push and pull and if the temperature changes suddenly then great news - you found the problem area. You can repeat for the circuit board but that requires removing the bottom cover (not that hard). Assuming the problem is at the bed - you can take the bed all apart pretty easily by removing the 3 leveling screws. Start by removing the glass, then undo the 3 leveling screws comlpetely. Watch out for washers and really pause and stare at how it goes back together. Lay the parts out in a safe spot where you won't bump them and lose what order they go. Often a washer sticks to the plate (the grease acts as glue). Once the wire is loose - remove and reinsert the 2 temp wires and power the printer up again to see if fixed. Report back.

You might like these ideas from educators although I think your current ideas are great: https://ultimaker.com/download/73369/Contributor_Cards.pdf