geert_2

When giving lessons or presentations in front of a public, or when doing electronic exams, or research projects or production runs, you definitely don't want automatic updates interfering. Stability and predictability are key. So, for me only manual updates, no automatic, even not optional via settings. Very often the default is "yes", and if you forget to reset this after a manual update, next time you're in trouble. I remember the Firefox 29(?) disaster, the Windows 7 to 10 auto-update disaster, Adobe- and Windows-updates in the middle of lessons, updates in the middle of exams of our students (even though updates were switched off), and so on. The only auto-updates that are funny, are those that fail on billboards in the street, creating a new form of street-art. I also dislike software that splatters its program files, settings and temporary files in hundreds of different inaccessible directories all over the harddisk. In Windows typically: the register, program files, program files (x86), appdata, application data, userdata, local settings, my documents, temp, and tons of subdirectories of these. This is creepy and disgusting. Just have a look at where your browser and office-programs put all their files. So I prefer manual installations from a zip-file: unzip and manually copy to the desired directory. And manual test-runs from standalone versions (sort of portable) where the settings are stored in the same directory as the program. I think all programs should be self-contained, independent units (directories), that you can carry around from one system to another, or move from one directory to another, without any negative side-effects, without losing settings, or without influencing other programs or older versions.

If you purge more, for example by printing ten lines of skirt, doesn't it come through? Or if you manually purge a bit of material before the print? Or is that blob sitting in the way and somehow blocking the movement of the filament?

I usually use "Fine", which suits 99.5% of my models. Only for big balls or smooth curves this gives a bit of visible segments. If this is a problem, I would take "Custom" and select lower values. In DesignSpark Mechanical, after exporting to STL, you can import that STL again, see the segments, and compare it to your original model. And so finetune your export-settings before beginning the slicing and printing.

Bubbles are caused by gasses, could be water, could be others. Ethylcellulose, does that contain alcohol, or chemically bonded alcohol, ether, or something similar? If so, could it be that the alcohol/ether evaporates, or the compound is broken down into its basic substances (alcohol + cellulose, or whatever the composition is)? That might explain both the bubbles and the poor extrusion / clogging. Maybe you could heat a bit of filament to the same temperature with a soldering iron, or on the outside of the nozzle (not inside, so you don't clog it), and see if and how it decomposes and degrades? Or google for how this material breaks down and burns? PS: what is the purpose of printing in this material?

Yes, the last two ones remind me of the time we lived in Jurassic Park. :-) They could as well be real fossilized dino skulls in a swamp.

I haven't printed with ABS and CPE, but I did print several models with PET, which is similar to CPE, I think? If the model allows it and there are not too much overhangs, I always print PET *without any fan*, and slow (25...30mm/s). This gives best bonding to the glass, no warping, and best clarity. Never had any layers splitting. For small models, I print PET on bare glass, or glass wiped with salt water. But for big models, I would rather use dilluted wood glue for PET. I don't know for ABS. If the standard brim does not provide enough force to keep the edges of the model down on the glass, you could design a bit thicker brim in your CAD file for more stability. Sometimes I make a brim of 0.5mm.

This reminds me somewhat of my cooling-tests (colorFabb PLA/PHA): When printing rather hot and on small areas, the surface does not get enough time to cool down, and the nozzle above it keeps radiating heat. So the model stays molten and it gets this weird overextrusion effect. Adding a dummy cooling tower (right models) reduces the effect, but does not eliminate it. The blobs tend to accumulate more on one side. Another effect that I see on overhangs, is that the overhang curls up, and then the nozzle bangs into it and pushes it down again. This also gives an ugly sidewall with lots of defects and ridges. I am not sure if these are the effect you experience, but it could be? Printing as cool as possible, slow, and with larger layer heights for overhangs, seems to help for me. Layers of 0.2mm give a much better look than my usual 0.1mm. Also use more than one sidewall, for added stability.

Most likely this is a case of two directories having the same name, but sitting in a different location? So Windows Explorer sees one directory, and Cura the other? For example a file downloaded to the folder "downloads" of user Jeff, and Cura searching in the "downloads" folder of user John? Try using old "known-good" files you made yourself. Then try moving the unwilling files to a known-good location, like the desktop, or your own "downloads" or "documents" folder. But occasionally, Windows relocates files behind your back (without telling you) to another directory if you tried to save them in a protected directory. This could happen if you save them to the "c:\" or to "c:\program files\" directory. Windows explorer remembers that it has done this, so it "finds" these files when you search in the original location where you thought you had saved them in. But other programs might not know this. This has to do with sytem protection I believe, although the exact "reasoning" of Windows is not clear to me. But I have run into it a couple of times on Windows 7 in the beginning. Unfortunately, I don't remember the name of the directory it moved them to (it was very weird), so you will have to google for it. Then there can also be "redirection policies" active, which also move things around behind your back. This is especially in business environments. But I don't know the details of it. Also, make sure the files have no "hidden", "system" or "read-only" attributes set. And make sure the filenames only contain standard ASCII-characters: letters, numbers, underscores. No exotic characters. And no dots, dashes or other special characters as first character, because this might make a file invisible or uneditable. If you are not sure, manually rename the files and type a whole new name.

I have printed quite a lot with colorFabb white (and other colors). Last time I ordered, they had two versions: standard white and blueish white: the standard white was a little bit creamy and warm white, definitely not snow white; while the blueish white had a faint shade of blue-grey. I guess this warm shade in the standard white is caused by their base material PLA/PHA ("naturel") also being creamy, like uncooked spaghetti. Both whites print okay, although not as smooth as colorFabb traffic red, dutch orange and naturel. But better than their signal yellow and intense green (spring green). Although the differences are minor. I guess this has to do with filler particles? Also, the white causes a little bit more wear on the nozzle than dutch orange, also due to filler particles, I guess? This is visible after 10 spools. (I have one printer in which I printed mainly orange, and another in which I printed mainly white.)

What you can always do is design all brim (and all other support features) in your CAD design. And switch off any automatically generated brim or support in Cura. Then you have full control over the shape and you can optimise it to your specific needs. For example, sometimes more than one layer of brim might be desirable to prevent the object from coming off the glass, especially if the brim is too flexible and when printing at low layer heights and 100% infill. Or you might want to connect all brim parts into one big plate covering the whole glass for more stability. Or you might want to add reinforcements into the brim, so it is a sort of combined brim-and-support, for example when printing lantern poles (similar to injection moulded bosses). This is what I do here: in some areas I need no brim at all, and no support. In some areas I do, and it needs to have special shapes to make later removal possible. So I designed them into the CAD file. The pink and orange things are supports. The cube is a dummy "cooling tower", otherwise the yellow model cannot be printed correctly (its top melts; this is a tiny keychain). I have no experience with PC though, so I can't advise on that.

If the model is designed in SketchUp, probably its walls are no solid walls, but rather a sort of "folded paper" models with gaps where the walls are glued together. Just like any paper model we glued as kids. Instead of a solid model, it is a bunch of surfaces that don't fit properly, so it is not a solid, and not watertight. You also see this problem when drawing text in SketchUp: some characters do not close. These gaps may be very hard to see, and you may need to zoom in quite a lot. One solutions is to manually and carefully select each vector and each edge, and carefully align and close the gaps. Another solution is to use a good 3D-editor that was designed for modeling solids. I use DesignSpark Mechanical, distributed by RS-components. This is a free but limited version of the commercial SpaceClaim. I have made many hundreds of models with it, and never had any problems. See these SketchUp text-samples (you may need to zoom in):

Just a question: could an effect as in the man's hair be created if there are too much line-segments in a too short distance? So the printer has to slow down? For example 100 little segments of 0.01mm, instead of 1 segment of 1mm? I vaguely remember reading a post where too much STL-triangles in a model created a similar problem? The indentations in the white boat rather look like a start-stop effect to me, after the nozzle arrives from somewhere else? Maybe you can see in the slicer if they mark the start of a new layer? Also, I noticed that blobs in PET are quite common: this material is rather thick and rubbery, and it tends to accumulate on the nozzle. Then after some time this goo sags and gets deposited on the print, usually as a light-brown blob. But that gives thick outwards blobs, not indentations. So this might be a cause of some fo the blobs in the man's weapon shield, but not of the indentations. Maybe you could cut the model into little pieces, and only print those where the effects are worst? And then keep watching if and when it happens? So you don't waste too much time and material? Sagging blobs of goo, and layer start/stop effects schould be clearly visible.

Maybe you could try the "salt method" on a small test piece? Or even better, on the same buildplate, print this test model multiple times using different methods. With a marker, draw lines on the glass, and use a different method in each area: - none at all - salt method - glue stick - dilluted wood glue - PVA-layer - 3D-LAC (spray on tissue and wipe with that, so you don't cover the whole glass) - hairspray - other...? The "salt method" works by first cleaning the glass thoroughly, then clean again with warm tap water only. And then wipe with a tissue moistened with *salt water*. Gently, always gently, keep wiping while it dries into an almost invisible mist of salt. It should look like a clean wine-glass that hasn't been used for a year: it is still clean, but you see a very faint haze on it. For next prints, you can just re-wipe the glass with the tissue with salt water. No need to take the glass out of the printer and clean it again. I haven't taken out and cleaned mine in a year. For regular PLA, this gives a very good bonding when hot, but absolutely no bonding when cold. However, it is not optimal for narrow, vertical models like lantern poles: the salt can not absorb shocks and then it tends to come lose. But for long, low, flat and 100% filled models like mine it works very well: I have printed many hundreds of my typical models now without problems. But I don't know if works for tough PLA? So make sure you stay with the printer. And please let us know, I would welcome feedback. Photo of my typical models (regular PLA, 100% filled): the front one is freshly printed. The back one has been sitting in a laboratory oven at ca. 70°C for a couple of hours, to try the effect: this gives an idea of the build-in material stresses that a bonding needs to withstand while printing.

To me this looks like poor adhesion to the glass plate. I don't know your printer, and I haven't really printed with ABS (apart from small test pieces), but in the very beginning I had a similar problem with PLA when printing on bare glass: if the climate was too moist, I would get this effect, and have poor bonding. The solution was to improve bonding (in my case, by wiping the glass with a tissue moistend with salt water, but that only works for PLA, not for ABS). So I think you might want to search for a better bonding method: ABS-slurry (=ABS dissolved in aceton), dilluted wood glue, 3D-LAC, strong hairspray, or something similar. And/or better cleaning of the glass, with isopropyl alcohol, but without any soaps because they reduce bonding. Also without cheap thinners or white spirits: these often leave traces of oils. Maybe also play with the bed temperature: adjust it up and down in steps of 5°C or 10°C. It should be very close to the glass transition temperature of your ABS (=the point where it just begins to soften).

I *do* think it is infill shining through, because the infill pattern in the first pic (of Cura) exactly matches the lines in the second pic (the photo). Maybe too much overlap between infill and walls? Or indeed overextrusion like yellowshark suggests: if the infill overextrudes upon slowing down at the end, causing a sort of blob, then I can imagine that this shows through, since the edge has nowhere to go but outwards. This effect would get worse at higher speeds, since the pressure in the nozzle can not immediately go down to zero when the printer slows down. Printing hotter could also increase this, I think, since the material is more molten and thus more easily flows away? But these are just educated guesses... Usually I print with 100% infill, but when I occasionally print with low infill on my UM2, sometimes I also see the pattern shining through, although not so much as yours. But I print slower and cooler.

I do not have an answer to your question, but there might be other issues with your approach: - Overhanging edges tend to curl up, which causes the nozzle to bang into these curls, and which might knock the model off the glass. - The support material might not stick very well to the steep slopes of the arches. I think it might be worth trying vertical support columns and then a horizontal dummy bridge. Then you only need support material on top of that bridge, so it would consume even less support material. The bottom of that dummy bridge will of course sag, but that doesn't matter, since you throw that away anyway. On top of that support-bridge you could even make some dovetail slots, to make the support material stick well. You could design all this manually, but I think there is an experimental feature from user smartavionics to do this automatically? Some time ago he was working on it, if I understood it well? If Cura wouldn't let you remove the supports, an option would be to disable support totally, and design all the supports in your CAD file, and then assign them to the correct nozzles in Cura. This gives you total control. See these pictures. They are from various different designs and tests, but you get the concept: Edges of overhangs severely curling up, about 1.5mm, so the nozzle banged into these curls, which knocked the model off the glass and created a nice spaghetti. Dummy support bridge, which is removed after printing. Here the support bridge is hanging from the walls of the real object, so it does not have its own legs. But for more stability or if you don't want to scar the real object, you could of course give the bridge its own legs. Dovetail system for adding mechanical grip of the support material to the dummy bridge.

I don't know the UMO, but on an UM2 there is a little fan at the back of the nozzle, cooling the seam between nozzle, teflon coupler and bowden tube. Probably there is one on the UMO too? Did you check if this little fan is running at full speed? If not running at all, or if too slow, heat would indeed creep up into the filament, melt it above the nozzle, and make feeding and printing impossible. The fan could be worn-out, or it could have sucked-up strings or hairs from prints and gotten stuck, or a cable came off?

Nice. And what method did you use for bonding to the glass?

My models are full of small circles and tiny features, and they print fine, both in PLA and PET. This is a keychain-miniature of dental models used in the hospital. I tend to print these things rather slow (25...30mm/s) and cool (195...200°C for PLA; 210...220°C for PET), bed temp 60°C for PLA, and 80...90°C for PET. Bed is leveled rather close so the first layer is squeezed well (manual leveling on old UM2 printers). Everything else is pretty much default (I am using an older version of Cura). For reference: text caps height = 3.5mm; character legs are 0.5mm wide. View of the bottom layers. This is transparent PET. The lighting from the back gives some weird highlights due to reflections and defractions in the model, but the circles are printed almost perfectly. They are ca. 2.5mm diameter. The ruler is in mm and cm. The tiny dark spots which look like corrosion pits are caused by salt, as I printed these using my "salt method" (=wipe the glass with a tissue moistened with salt water). This is the routine quality I get, with close to default settings. So there might be something else going on in your printer or settings? A setting that inadvertentie got changed? Or printing too hot or cold? Or a non-optimal bonding? Bed too far away, or too close?

Although this is not an answer to your question, it might also improve the situation: the cause of warping might also be because of the way you clean your glass plate, the bonding method you use (or lack of), the distance between nozzle and bed, and the environmental situation (air moisture, temp). Since I started using the "salt method" for bonding, I don't have Ultimaker PLA warping anymore. I use their nice "Pearl" color; I don't know about other colors. The "salt method" = first clean the glass with whatever means you want, e.g. isopropyl alcohol, acetone,... (but no soaps or oils). Then clean with pure warm tap water only. Then wipe the glass plate with a tissue moistened with salt water. Or put a few drops of salt water on the glass and wipe them dry with the tissue. Very gently (!!!) keep wiping until this dries in a thin almost invisible mist of salt stuck to the glass. Especially in moist weather this gives me a much improved bonding when the glass is hot (60°C). When cold, there is no bonding at all, so the models come off by themself after finishing and cooling down. The ease of applying and ease of model removal makes this method attractive for me. This works very well for low and flat models. My models are usually 12 to 15 cm long and 100% filled. But I don't recommend it for narrow and high models (e.g. lantern poles) or models with huge overhangs, because the salt can not absorb much shocks when the nozzle bangs into edges of a model (overhangs tend to curl up). An original model (bottom), perfectly flat after printing. And one that sat in a laboratory oven (top) at an elevated temperature of 70°C for a couple of hours, to see how it would react and warp due to its internal molded-in stresses, as a result of the 3D-printing process. Other people are very successfully using other bonding methods: hairspray, dilluted wood glue, glue stick (some wipe it with water afterwards), 3D-LAC, and other means. Each method may have its advantages and disadvantages. So it might be worth sorting out the basic cause of the warping and remedy that, instead of trying to find a workaround.

Yes, I do fully understand that concern. I am feeling the same about post-processing work; I don't like it. :-) But printing vertically with lots of retractions, and the nozzle moving through the air a lot, also takes time, in addition to creating the imperfections you have seen. Depending on your designs and your printer, there might come a point where printing flat goes faster, and requires less post-processing than smoothing out these imperfections. So you could get more done with less effort. Maybe... If you plan on doing lots of prints, I think it might be a good idea to print a small representative test in both ways, and compare the time, quality and required post-processing? Or even not print them, but just slice them in Cura, and compare the predicted printing times?

This looks like a fun project for kids. Did you print each cutter-blade assembly in one piece, or print each blade separately and then mount them on a rod?

Maybe printing the walls flat on their back, and then glueing the pieces together, might give better results? Similar to the HO-model railroad houses we had as kids from Faller, Vollmer, Kibri, etc... Then the material flow is more constant and the print head does not need to travel through air as much. Or maybe you could print the main wall vertically, just like you did now, but print only the windows flat on their back, and then assemble? That might give a better layer line pattern? Also, using sand- and stone-colors (beige, warm grey, light grey, cream) could also help to hide artifacts, and they are very well suited for architectural work. White is often a difficult color to print, and it shows defects quite hard. ColorFabb has a lot of custom sand- and stone-colors. (I am talking about the non-filled, plain PLA colors.)

I hope you speak a bit of English. As far as I know, almost all magnets lose their magnetic characteristics at elevated temperatures, because the iron (or other) magnetic particles become looser and their alignment changes into random. For neodymium this is around 80...90°C if I remember well (but verify it, as I could be wrong). 3D-printing is done above 200°C... So, I think you would need to print a magnetisable material, and then later turn that into a permanent magnet by exposing it to a very strong magnetic field. I don't know how that is done exactly, maybe at somewhat elevated temperature? Probably you can Google how permanent magnets are made? Or maybe you could use a magnetisable material, for example iron-filled filament, and then put some very strong permanent magnets close to the nozzle (but out of the heat), so they align the freshly printed particles while still molten, during the printing? So that when cooled, they keep that orientation? Not sure if this would work, but it might be the easiest method to try. Use an abrasive-resistant nozzle for this.

To me this seems reasonably okay, it looks like what you could expect at 0.3mm layer height. Although there seems to be a bit of "elephant feet" at the first layers (=thick bottom). There may be other reasons, but often this is caused by a bed temperature that is a little bit too high for the material, so the first layers sag a bit. Maybe you could try to lower your bed temperature by 5°C and test if the model still sticks well? Stay with the print, in case the model would come off, so you can abort. When printing on a glass bed (I don't know what bed your printer has), we often have to find a balance between bed temp and good bonding. A too low temperature reduces bonding, and the model may suddenly pop off while printing. A too high temperature also reduces bonding: the model stays too weak, and warping forces tend to peel off the model gradually, in my tests. In both my UM2, and for PLA, the best bed temperature was the default indeed, 60°C. Although 55°C also worked usually. At 45...50°C the models would suddenly pop off. At 65...70°C they would get severe elephant feet and would be peeled off. This will differ from material to material, and printer to printer, so I recommend doing test prints at various bed temps, and definitely stay with the printer!