gr5

Okay so this is very common. And very easy to fix. ABS is a trickier material to print only because of it's higher softening temp which causes about 10 issues and this is one of the 10 issues. Most materials with higher softening temps have this issue (not nylon though). The basic problem is layer adhesion aka layer bonding. While printing a layer, the layer below is too cold and the current layer is not melting the layer below to get a good bond. The fix is typically 3 things: warmer ambient air, less fan, hotter nozzle. The warmer ambient air is typically the easiest to achieve - at least for me - just enclose the printer and if you have a heated bed it will just take care of itself. If you raise the air temp from 20C to 35C that will make a huge difference. softening temp of ABS is around 99C so air at 20C is about 80C cooler, air at 35C is about 65C cooler. Not quite twice as good but a huge improvement. On the original reprap machines that printed ABS they didn't enclose the machine so they relied on the next 2 methods. Fan - in the old days people didn't do much fan with ABS. Fan really helps with overhangs and tiny parts but you want much less fan. In fact the older slicers designed for ABS would turn the fan on for just a few seconds during bridging and overhangs and turn it off for infill and such. You want the minimum fan but where it still rotates. If you have zero fan the part won't look as nice. A little fan is a good thing. On an UM3 you want around 1% or 2% fan. 9% fan is MUCH too much. On a UM2, or UMO, 30% fan is decent. Nozzle temp - this is the hardest. ABS has a very narrow temperature printing range. If you raise the temp too much and let the ABS sit in the nozzle for a few minutes it will turn into a gummy substance that can only be removed with heat and a toothpick or similar tool. It's horrible stuff. If you print too cold you get bad layer bonding. In the old days they printed .2mm to .4mm layer height which meant the ABS didn't stay in the nozzle long so they could print hotter. That's one solution (print fast and hot). But if you want 0.1mm layers you will have to keep raising your nozzle temp by 1C at a time until you clog it some day and then go back down a bit. And be very careful not to preheat that nozzle and let it sit for several minutes not printing.

The TUNE menu on the printer tells the firmware called "Marlin" to adjust all speeds by that amount. So the answer is "yes". It also affects retraction speed, z axis speeds, etc. Each axis also has a max speed (typically 300mm/sec for x,y axes for example) and if the TUNE adjustment results in a speed over the max it will limit to the max speed instead.

I'm not sure what you are printing but it sounds like it's doing a lot of retractions? That can cause grinding issues. You don't want the same spot of filament going back and forth more than 20 times through the feeder. In fact to test out my theory let's limit it to 10 times so set this: "maximum retraction count" to 10 and set "minimum extrusion distance window" to 4.5mm (that's typical retraction distance) Reslice and there will probably be a lot fewer retractions and your filament should be less likely to grind. This is assuming you have a lot of retractions. It should also speed up your predicted print time.

@TedStryker maybe now that you have one approved message it will let you do a direct message. So try again. I don't know what will happen.

I'm glad you are seeing success. Do whatever works and stick with it until it doesn't work. I'm the guy in the video.

Good luck. The odds of you getting two problem printers in a row is quite low. UM is a small company - they aren't like apple. They make thousands of these things per month, not millions. So don't expect the same level of quality. But as far as I can tell Ultimaker is among the best out there when it comes to quality. They do test every printer before they ship. I doubt they test the power supplies - maybe they should - maybe they do. But your issue might not have been the power supply - it might have been something else.

This is a complex subject. 1) Partly because normal cores are one nozzle size and hardcores can be on of at least 7 possible nozzle sizes and can be brass, steel, or ruby so really 7 brass types, 5 ruby types, and one steel type. Or 13 types all which might print at different temperatures, speeds, layer heights, etc. The "AA 0.4" core alone has about 80 profiles (there's one for each quality level multiplied by one for each material plus compatibility files (which materials are compatible) need updating and so on. So there would have to be about 80 X 13 nozzle types or about 1000 files to edit/create. Having said that, there is a utility to create the first 80 or so for the hardcore with the 0.4 nozzle. I can give you that if you want but you will need to understand what it does first. 2) Now there is a problem with this. If your hardcore comes up as "HardCore" or "AA 0.4" (3dsolex used to sell them the first way but now only sells them the second way I think) and in Cura you use a profile called say "HC 0.4" then cura connect won't work. You won't be able to print over the network. Which might really suck for you. But if the core is programmed as "HardCore" and I send you my little utility such that you have a set of "HardCore" profiles (only a single drop down is added and it's designed for 0.4 nozzle only) then, "yay!" because you can use cura connect. But "boo!" because it is no different than the "AA 0.4" and you might have a 0.6mm nozzle on your hardcore. ==== So really, in the end, hopefully your HazrdCore is programmed as "AA 0.4" and you can just change the line width in Cura to match the nozzle width. That is the main difference. In addition I drop the temperature 15C when I use my HardCore with a 0.8mm nozzle because, well, it just prints better there. You can belive that all 80 profiles for the AA 0.4 are prefect and all 80 profiles for the AA 0.8 are perfect but in my experience you are going to have to mess with the temperature a little bit to get really good results no matter what. Every person's printing style is just a bit different. Even changing filament colors (e.g. White is a pain to print and needs to print cooler) can change what the ideal settings are and UM doesn't support that with differences due to color (looking forward to 2000 profiles per nozzle size instead of just 80?). So you are going to have to tweak things a bit no matter what. The most important thing is to set the line width to the nozzle width and then to look over the obvious things: layer height should "match" nozzle width somewhat - so print maybe .3mm or .4mm layer heights with .8mm nozzle. Temperature should be lowered if you are getting lots of stringing or overhangs look bad. By as much as 20C. Shell width should probably be a multiple of line width (otherwise it will do some purposefully underextruded shells to get it to add up). Those are the key differences from nozzle to nozzle: line width, temperature, shell width, layer height. If your hardcore is programmed as "HardCore" and you want to use cura connect, the special profile creator program is here: http://gr5.org/hc/

This is a windows question. Not a cura question. If you format the SD card you will indeed lose everything on there. I get that message all the time and just choose to not format my SD card. But if there is nothing important on the SD card then go for it. This message comes up if windows thinks the card might be corrupted. So it is a good idea to format it. But usually (99% of the time) it's not necessary. But if you are getting strange behavior such as your print ends half way through for no reason and when look at the file on the SD card and the top half looks fine and the bottom half is full of random characters - well then it's time to format it.

So the leveling procedure only gets you so far. It's hard to say if there is a problem in the right edge of the photo - that doesn't look serious - it should be fine on the next layer. But in the bottom left area you definitely have a problem - nozzle too far from glass. As you seem to say. Just ignore the leveling procedure and turn the front left screw 1/2 turn CCW to move the glass up a bit (this will cause the opposite corner to go down a little but that is probably okay for this part - if not then you have to turn the other two screws CCW 1/4 turn to put that spot back where it was). When the UM2 is printing skirt or brim that is your opportunity to correct the leveling in the "corners" or "extremes" of your part. Having said all that also note that the glass is typically not flat and so you want your bottom layer to be nice and thick (0.3mm is a good default and is what all versions of cura default to for bottom layer) to compensate for dips in the glass. "plate" glass is very very flat. But UM uses tempered glass.

The last time I visited Ultimaker (1.5 years ago) some key people unfortunately took vacation that day. Well I look just like my icon and assuming you do also we should have no problem noticing each other. lol.

I could be wrong - Erin's seen thousands of Um2s. I've seen this problem on one (mine). And maybe 2 other people on the forum had what seems like the same problem. But try the non-GST. And yes, look for black carbon charring in the area where the wires connect to the bed and where they connect to the PCB.

Since you have a um3, maybe you should print that first catapult with PVA. That's what's so great about the UM3 - it can print overhangs with dissolvable support.

You have 3 issues that I see: 1) Pointy tops melted. 2) under side of catapult sucks (steep overhang) 3) On of the 4 pyramid ridges is strange People have been addressing #1 as it's pretty easy to fix and common. You CAN completely get rid of it. just the way geert says to do. Print a second thing next to it. Or even better print 3 things and the middle one will have the best quality (two added towers -but add them into an empty cura like this: tower, your part, tower as it usually prints in the order added and you want your part in the middle so cura doesn't do 2 layers in a row on your part). #2 - overhangs This is typical for a steep overhang. I don't know why you didn't see it before. Was it the exact same catapult? Do you still have the old one? I bet it looked almost as bad as this one if it was the exact same angle. Layer height, pla color (yes just the color), print speed, fan speed, and temperature can all affect the quality. Did you do this one the same layer height as before? Was it the same color pla? Same temperature? Cura changes things like temperature from one version to the next so if you changed cura versions... The most important thing to get good overhangs is max fan (100% fan). The next most important thing is to print cool as other's have said. Try dropping the temp by 15C (just do it in the tune menu) and cut the speed in half (also in the TUNE menu). It will take longer but should look a little better. I recommend you print the catapult again and change this half way through the print to see visually the difference. Layer height matters. You don't want too thick or too thin. I personally think around 0.1mm is ideal. But I'm not 100% sure. #3 - that ridge is weird. It looks like you are simply printing too hot. Maybe. I'm going to go with too hot. Try lowering the temp around 10C and reprint the pyramid. Even better, while it's printing, lower the temp by 5C for 10 layers, then another 5C for 10 layers, then another 5C for 10 layers. Maybe mark the part with a sharpie as you go. Let us know what you learn.

Core's come with a tiny eeprom and are labeled typically "AA 0.4". 3dsolex cores usually are also labeled "AA 0.4" for convenience. The older 3dsolex cores were labeled "HardCore". This caused some consternation with the printer firmware and also with cura if cura is reading your printer over the network. It was pretty minor (just hit "ignore" a few times) up until Ultimaker updated how they allow you to print over the network. The latest version of cura refuses to let you print over the network if the core name doesn't match what you sliced. 3dsolex and I created an "installer" that creates about 80 files in the %appdata%/cura folder. I put installer in quotes because it doesn't contain any files like a normal installer would. Instead it's a script that copies the files from your existing cura installation into the %appdata% and then modifies where they say "AA 0.4" to instead say "HardCore". This installer isn't needed for newer customers. Why 80 files? Well there are about 10 different materials (abs, cpe, pla, nylon, pva, PP, ...) and about 4 different quality levels for each one. Each combination (that's 40 so far) requires a different file. There are 3 types of files (e.g. material files). Some files say if things are compatible. Other files say what temperature to print at. It's a big mess. It's pretty cool how UM did it but it ends up being a LOT of files for just one core. Anyway - please don't worry about any of this. When you get a hardcore and use it everything is soooo simple. It's really no big deal. Just set the "line width" to your nozzle size. It's a single parameter in cura - it's the only thing you must change. As you get more advanced you will also probably change the nozzle temperature as everyone finds a temperature they prefer and different nozzles perform slightly differently.

I agree with geert_2 - that's what I always do. Just a small 1cm by 1cm cuboid of the exact same height as the bishop will do. And place the other object in a position such that fans are cooling your bishop while the other object is being printed. Even just 3 seconds away from this part should give it plenty of time to cool. I have a 30cm_cube.stl file I keep handy. Cura lets me adjust the x,y,z scaling in mm (you can do % or mm scaling). I make it the same height as the model and adjust the x/y to be about one fifth the height so that it's a stable part and won't fall over. In other words if the part is 50mm high my base is 10X10. If the part is 10mm high I could go as low as 2x2 (probably would do 5x5).

Well my answer hasn't changed since my first reply here. We need to see the output of both of your slicers - the entire top of the file up until and including the first printing move.

done. moved.

Good to know. Hey I'm going to be visiting UM headquarters February 23. Don't take that day off please I'd like to meet you and @ghostkeeper.

Well that's what G92 does in Marlin. Which is the most common firmware out there - I don't know anything about this guys printer though. According to this document G92 does the same thing in repetier also but they could be wrong: http://reprap.org/wiki/G-code#G92:_Set_Position

This is a common problem with the UM2 and the newer power supplies that ship with the UM2 and the UM3. The UM3 actually has special code to keep the main bed heater from drawing too much power when the nozzles are both drawing power. One fix is to just contact your reseller and have them ship you another power brick. Don't ship back your current one until you get the new one. Another fix is to buy the "GS" model and not the "GST" power brick. Below are the 2 part numbers. This will cost more but you will have a more reliable power supply (GS doesn't meet efficiency requirements for certain countries so UM had to switch to GST). But the GS bricks cost about $90 each. Anyway that is the route I went. It has a much larger "headroom" aka "power margin". GST220A24-R7B GS220A24-R7B A workaround while you are waiting for a replacement supply is to print on blue tape. Make sure to re-level and make sure to clean the blue painters tape with isopropyl alcohol first. Printing at a lower bed temperature is NOT a good workaround because the bed actually can draw more power at the lower temps (the resistance increases with temperature).

salt trick is different from pva glue. Don't combine them. Try one or the other. Although I think your issue may be you aren't squishing the bottom layer enough. Rather than use a .4 nozzle and tell the printer to treat it like a .15, you should instead put a .25mm nozle on there. And definitely your bottom layer is not squishing well. There are ways to do this with glue stick only. You don't need to do hairspray. more info here:

Um... is @ahoeben on the cura team?

I think we need to see what cura added next. Up until it starts printing. New and old cura. Pay particular attention to G28 commands (maybe there is another one in the cura code coming up and you don't want that) and G92 commands (which tells Marlin - hey ignore where you think you are - now you are at this new position even though nothing moved).

First of all the difference is pretty small and I think most people wouldn't care. I mean the final part should look fine, right? Well maybe lower/tighten all 3 screws just 1/3 a rotation to get the bed a little farther from the nozzle. But the final part should look fine I think. Just because the bottom layer is a little thin doesn't mean the whole part is bad. Anyway what is causing this is almost certainly glass that is not flat. Ultimaker uses tempered glass (not certain why they do this) and tempered glass tends to be thicker/higher in the middle just like this. You could just accept it like this or you could ask your reseller for another piece of glass. But they are all like this to some degree. Some glass plates are better than others.

Well maybe you can set the top and bottom thickness to zero and set the infill pattern such that it has about 1mm holes. Maybe around 80% infill - play with it. This only works if it's okay to be porous vertically but not horizontally. This is what people do when they make "molds" for cardboard packaging (start with wet cardboard pulp (or cut up newspapers) and put it on the surface, let it dry and voila. Also what they do for vacuum form molding. Alternatively maybe there is a way to set the flow rate to 50% JUST FOR THE TOP AND BOTTOM. And just do 2 infill layers maybe (top/bottom thickness set to 2 layers). Try those 2 ideas first. If cura doesn't let you set flow for just top/bottom infill then you could do it by modifying the gcodes a bit - just insert the gcode that sets the flow to 50% or 100% in a few places in the code - 100% for shell bottom few layers and 50% for infill. The gcode file is marked up where each layer starts/ends (plus you can look at the Z value in there to see the current printing height) and the gcode has comments showing where shell starts/ends and where infill starts/ends and so on. You might only have to insert these 5 or 10 times. If you end up creating tons of these all the time you could write a plugin that does this automatically. This kind of post-processing plugin would be a perl script that locates all the top/bottom infill parts of the gcodes and inserts the flow gcode around that section. OR you could do it all in CAD. Tell the cad software to put thousands of holes through your part. This is pretty easy to do in openSCAD. Or you could take the STL file from your cad, pull it into openscad (very easy) and then subtract off thousands of thin cylinders and save it back out. Also pretty easy to do in openSCAD. I could write that for you for about $100 if you want. Send me an stl and I'll send it back with holes and if you like it, pay me and I send you the openscad code and help you set up openscad. I'd write it so you can easily change the hole spacing and the hole diameter. But some of the holes would put cylindrical grooves in the edges so that would have to be acceptable to you.