geert_2

In my experience, *random* crashes usually indicate a hardware problem. Often memory modules that got defect, or contact pins that got dirty or oxidated. If the driver update would not help, and if you can access the memory modules, try gently (!) wiggling them a couple of times. This has often worked well in my PCs (I don't know for a Mac). Maybe try a night-long RAM memory test? Memory-problems often show up only on memory-intensive software that occupies nearly all of the RAM: 3D-software, browsers with a ton of open tabs all playing videos, etc.

To me your samples look pretty decent. This is what I would expect. Infill is not meant to be visually perfect: its purpose is to add strength in *invisible* areas. And to reduce material consumption and printing time in invisible insides, where a solid body is not required. So its purpose is a good balance between strength, printing time, and material use. Although today there are a couple of really beautifull infills indeed. If you want nice looking spokes, and you want more control over how they look, I would suggest you design them in CAD. See the model below (its holes are 4mm if I remember well). But even then retractions and corners can introduce little defects, typical for this sort of production: strings, blobs, thicker areas,... Compare this to injection moulding which always leaves marks from the injector nozzle, from the ejector pins, from the flow, and from the seam lines. Each production method has its limitations. Some customers have unrealistic expectations, due to the hype in the press that "you can 3D-print anything", which isn't true. What I do when someone asks me to print a model, and he is unfamiliar with 3D-printing, is first show them the limitations. What it *cannot* do: overhangs without support, smooth curved surfaces, waterclear transparant objects, tiny pilars, small screw threads, tiny holes, etc. I have a couple of demo-models sitting around for that. And only then I show them what it *can* do, with a couple of nice models. This takes out most of their unrealistic wishes, and they will be much happier with the results, if they continue anyway.

On my UM2, the tube is also worn-out, but it keeps clamping well. If the nozzle is not blocked, and there are no temperature problems (printing too cold, so it does not melt), the tubing is the correct diameter, and the filament is not too thick, or any other things that block feeding, then I keep thinking that you might not be inserting the tube well enough? It is not sufficient to just push-in the tube all the way down. At the same time, you also need to lift the white ring as high as possible by pulling hard: this lifting is what will cause the locking afterwards. Could you verify this again? If you already did this correctly, then I am out of ideas, concerning physics. If the materialistic ways of handling do not work, and there is *nothing wrong* physically, then maybe you need to handle it spiritually? Remember the first moment when you concluded that "it was pushing out the bowden tube"? Or whatever wording you had in mind? Then remove that early decision, cancel it, and remove all negative emotions and upsets associated with that past event. So the past is clean. Because the printer might still be operating on that old decision, even though the cause might no longer be there. Replace this old decision with the new firm decision that it will now work well. The reason is that people tend to keep living on old decisions and conclusions, even though their environments are changing, and their old ideas might no longer be valid in the new circumstances. Then their routines and old conclusions get in the way, instead of helping. Before making new decisions or conclusions, people should first remember and remove the old conclusions. Otherwise these tend to stay active and interfere with the new ones. Some people have very strong spiritual powers that can influence objects in the material world. Maybe you are one of them?

For the future, get your staff and students to use other software than SketchUp for 3D-printing, or you will keep running into problems and keep repairing errors. SketchUp was designed for visual models only, not for 3D-printing. It produces sort of "cardboard" models where the edges do not fit together and are not watertight, not solid. It is excellent if you want a quick idea on-screen for an architectural concept, as long as you are not going to print it. Students and educators can often get free or cheap educational versions of professional 3D-CAD software. Otherwise, you could use DesignSpark Mechanical (this is the one I use), or any other freeware offline or online program, designed for watertight *solid* modeling.

I wonder how this would reduce stress? Wouldn't most of the stress be caused by adding hot new layers to a colder base, causing an *upwards* pulling force due to shrinking? Further, I think this might cause more ugly corners: now the ringing effects are on only one side of a corner, and they are the same for each layer. When alternating, these effects would alternate too. This might cause the same combing effects as on interlaced videos?

I just printed a filter for my vacuum pump, to prevent it from sucking up big particles when suddenly applying vacuum. The filter was printed on an UM2, thus in a single material, no supports. I printed the first model at 0.3mm layer height, 25mm/s. This was *not* airtight: when I put tap water on it, maybe 10 tiny jets spurted out of the shell, through tiny holes. The holes were where the layer-changes and take-offs and landings from traveling had occured. The second model was printed at 0.06mm layer height, and 50mm/s. This model is *absolutely watertight*. Yes it took a lot longer, but it works now. This filter is quite small, 30mm diameter, with 8mm, 10mm and 12mm bulbs for attaching tubing. So it has a thin shell. Bigger objects might need a *much thicker* shell to withstand pressure. Major slicer settings: - 0.06mm layer height, - 50mm/s (don't go too slow at thin layers, otherwiste the material sits in the nozzle for too long and might decompose), - shell 0.8mm (=2 lines, 0.4mm nozzle), - infill 50...60%, - top and bottom: 0 (=no solid top and bottom layers, so the filter is porous); I could have printed it in one shot, vertically. But this would require custom supports for stability. I did not want it to wobble or fall over. So I preferred to split and print it in two parts, and glue them together with cyanoacrylate glue. And then I sanded it lightly and acetoned it, for a good fitting of the tubing. Both parts need to be glued together, bottom to bottom, carefully applying cyanoacrylate around the edge, without gaps, but also without flooding and blocking the filter. Anyway, before doing a large print that takes days, try different settings very carefully on a smaller testmodel. Go from where it sure leaks, and then gradually to where it is surely tight. Then keep a safe margin. Also consider post-processing, acetoning, coating, glueing, whatever,... to make sure it is airtight. A thick varnish, coating or paint might also solve the problems. Print without dissolvable support if you can. Otherwise strings of support through the print would cause holes after dissolving.

I think the basic idea is excellent, and worth trying out. But maybe an optical sensor or mechanical switch might be more reliable? I would fear that a thermal switch could cause false alarms, e.g. when printing hot and slow without cooling fan, so there is no airflow. Then it might get too hot and trip. Or it might miss real leaks when printing fast, cool and with full cooling fan, if the air blows directly on the sensor, and it cools the outer shell of the leaking plastic too fast. Such a sensor should never give false alarms and abort a good print, nor miss real events: that will be the most difficult point, I think.

I never used sprays for 3D-print adhesion. But years ago I did use a lot of silicone anti-stick sprays for moulding and casting models. I had built a special cardboard box with fume extraction at the back, to take away the spray mist. And I only sprayed deep into that box. But even then the whole area outside of that cardboard box got slippery with silicone oil too, after some time. I have seen similar setups where people used spray paints, and there too the whole environment got colored. Even though they had fume extraction in their cardboard box, just like me! If you are very adventurous, try using bright red paint instead of glue spray. And see where it goes. (No, don't do it. :-)

If it is a lab with normal ventilation, and you did not sit with your nose directly above the printer all day, and if you only printed PET or PLA, it probably will not have done any damage. Seems very unlikely to me. I don't know where you live, but here in Belgium the laws require that the air in a lab is renewed several times per hour (was it 6x or 10x? I don't remember). So you should get plenty of fresh air in the lab anyway. In a research lab, most other products you use will probably be more dangerous: solvents, biochemicals, composites,... Things would be different if you would be printing all day in an unventilated small room, for years and years. Then fumes could accumulate. If you would start projects that require *intensive* 3D-printing, all day long for months, then I would recommend attaching a tube from the ventilation to nearby the printer. Maybe not if you only print once a week for an hour or so, and you are sitting far enough away. If in doubt, try using visible fumes or smoke to see how the air is extracted. Maybe you can place the printers in such a way that the natural flow in the room takes away the fumes? But this is my personal viewpoint. I don't know if it is in accordance with your local laws, and I don't know your exact situation. Always use common sense.

Steep overhangs tend to curl up. Maybe that might cause this deformation? Watch closely while it is printing, then you can see if this is the cause indeed. If you can't get it to print well, another thing you could do is make an undeep thread, so the overhangs are not so steep. And then using a standard thread cutter, cut the final thread. When cutting threads in PLA, go very slow, and with lots of cooling. Otherwise everything will just melt. Don't ask how I know... :-)

I have silica gel with a moisture indicator, which turns from dark blue when dry, into pink when moist. To the Ultimaker-developers: maybe you could try adding this moisture-sensitive pigment to the PVA? Then people can see at a glance if their PVA is still dry enough? I don't need it, since I only have single nozzle UM2-printers. But moist PVA seems to come up again and again here on the forums.

I didn't make a real font-file, since I don't know how to do that. So to set text, you need to copy and pasted each character from the character set, letter by letter. Like in the old days of metal printing. This is good for a short copyright notice, but not very suitable for 3D-printed newspapers, obviously. The character set is in DesignSpark Mechanical's native format, RSDOC; and DesignSpark Mechanical is freeware (requires registration). If you would like to convert these characters to a real font, thus a TTF-file or similar, feel free to do so. Just keep it free for everyone. Also, I can read STEP-files, but can not export to them. So, if anyone has the STEP-converter option in DesignSpark Mechanical, feel free to export this set. Then it is still not a real font, but at least the STEP-format is way more universal than the RSDOC-format. The set is still here (scroll down a bit): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

When mounting the bowden tube, do you push it far enough in, all the way down? Thus: first lift the rentention ring with your finger nails. In my printer this is a white ring on top of the head, where the tube goes in. Then insert the bowden tube all the way down, while you keep the white ring lifted. And then insert the horseshoe clip? If the tube is not inserted deep enough, it may not grip well upon pulling. That could produce the phenomena you see. (But that does not mean there can't be other causes, such as a worn-out tube, or incorrect outer diameter.) For easier removal and replacement of the tube and horseshoe clips, I have designed my own clips a couple of years ago. They were made for my UM2 printer, so I am not sure if they still fit on an UM3? Anyway, feel free to borrow the concept and design your own, if these wouldn't fit well on an UM3. See here (and then scroll down a bit): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

Extruding thin wires of consistent diameter is not easy: try manually extruding a sausage from your printer, by slowly turning the wheel. Through a 0.4mm nozzle, I can get sausages ranging from 0.05mm up to 1.0mm, by varying temperature and pressure alone. The extruded plastic contracts length-wise after coming out of the nozzle, while still molten, as the molecules tend to go back to their usual curled-up shape (instead of being too stretched-out). This makes the diameter very unpredictable, and requires a very constant temperature, pressure, speed, cooling; and constant monitoring and adjusting. If the filament diameter is only 10% off, then the area is 20% off, thus causing over- or underextrusion in your printer (if it still can pass through the bowden tube and nozzle). So I don't think you will get very far with home-equipment, although it might be good enough for a doodle-pen? Maybe you can find second-hand *professional* equipment from bankrupt companies, or from stopped production lines? Or from companies that upgraded to new machinery? If you can get these cheap, it might work well. But even then the setup for each run may consume and waste as much material as a full spool. When I was a kid, in our neighbourhood we had a plastic company that produced straws for lollipops. At that time safety rules were a bit more relaxed, so we were occasionally allowed to walk around between the machinery, as long as we didn't touch anything. Each startup of the plastic extruder produced a huge bag of waste, before the flow was constant and smooth enough, even for lollipop straws. It is not just the extruder: you also need the grinder, feeder, heater, dryer in front of the extruder; and behind the machine the long cooling trajectory with transport band, and the spool-winding machinery. So you should rather do this as a learning exercise, and for the joy of creating new stuff, but not for economy.

Where did the old printers go? Were they sold, given away, or are they sitting somewhere in the basement, collecting dust? If given away and they still worked, I would suggest you just donate the filament too. If they are sitting in the basement, you could revive them in a separate room, and let students mess around with them for their hobby projects, and let the students take care of them, thus out of the official curriculum.

It looks very similar to the Polyalchemy Elixir filaments, which were the first in this class as far as I know. It seems a lot of companies have jumped on that bandwagon of shiny silky filaments recently. Haven't tried them yet, but I might in the future, I like the metalic aspect. In your experience, how is surface quality? Does the shinyness hide defects, or does it rather make them more pronounced? And how are warping, layer bonding, and temperature, compared to standard PLA?

Would it be an option to hollow-out the model in CAD, instead of in Cura? Let's say we have a solid cube. Then, in DesignSpark Mechanical you can delete one face, and set a wall thickness for the remaining walls (so it is printable). This turns the object from a solid cube into a hollow cup. I don't know if your software has such an option? Also, make sure you have enough contact-area to ensure a good bonding to the glass. This may depend on the bonding method. For my "salt method", a single wall is good for low flat objects, but it is likely to come off for high objects, or objects with overhangs. I don't know for glue-methods, as I haven't used glue in years. This is what I mean: (my statues are not as artistic as yours. :-)

In PLA-based filament it is known that microcracks can grow if the filament is bent or stretched (thus kept under a load) for a longer period of time. So, don't let material sit still in the feeding-traject after the print is finished, but unload the spool immediately. Some materials also get brittle due to changes in crystal structure, becoming more crystaline (often reversible), and due to moisture absorption and damage (not reversible as far as I know). Both happen in PLA. Keeping it dry obviously helps against hydrolysis. But I am not sure what the best solution is for "un-crystalisation"? Melting should help, but then you lose your filament. Heating up to the point just before deformation might change its molecule structure in both ways: making it soft again, or rather making it harder (encouraging crystal growth), similar to post-curing and annealing. I don't know which one wins. And your warm room might speed-up these effects? I don't know your material, so I have no idea if it is affected by these phenomena. Try what the effect is of heat on a few short pieces of this filament. Keep in mind that when heating it too much above its glass transition temperature, it will shrink in length, but get thicker (e.g. from 2.85mm to 3.10mm), and then it may no longer fit in the bowden tube or nozzle. You could also try unwinding it manually, and manually straightening it a bit, after which you release it again (to stop crack growth), so the bending radius is not as tight as before. Then it will get stressed less in the feeder. It may take trial and error. It could also be a bad batch or spool, or a filament that is very brittle by nature, especially if it is a filled filament. Pictures: Microcracks in PLA/PHA filament after straightening it, and then releasing it again, so the stress is off. If the bending stress would be kept on, these cracks would keep growing until the filament would break.

I was referring to the squishing of the filament indeed. I have two older UM2 printers with manual bed-adjustment, and I adjust it closer, so it is squished more. I prefer a nice glossy bottom, even if that causes a little bit of "elephant feet" sometimes. Not only does that stick better, but it makes watermark text in transparent materials (=inside the model, close to the bottom) easier to read. In Cura, I usually set the first layer to 0.2mm. This too helps in giving a flatter bottom than 0.3mm, and gives better bonding. I have no experience with UM3 or newer, so I can't say much about their procedures and their auto-leveling. I think you can switch it off, but I don't know if there is an in-between such as "auto-level to my prefered height"?

If that is the underside, I would level the bed closer to the nozzle. It is hard to see from this angle, but I think it is not flat enough to my taste. :-) I prefer the bottoms of my prints to look more like this, or like the above photos.

Belts are replaceble, and the procedure should be somewhere on the Ultimaker-site. If you are a bit handy, you can do it yourself. Your local dealer should be able to provide the belts. Or you can have him/her do it. While you are at it, I would take the time to have a look at the feeder too: is that clean and in good condition? Also clean the metal rods near the belts, and lubricate them with a thin oil that does not dry out. And also clean the nozzle with cold pulls or "atomic pulls". See the official method on the Ultimaker-site, or use my gentle method here (I am no Ultimaker-staff, just a user): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ And check if the white teflon coupler above the nozzle is still okay, not deformed, not burned.

Just randomly printing Lego pieces doesn't seem like a good idea to me, since: (1) they won't fit, as a 3D-printer doesn't have the required accuracy (microns), and so the pieces would be useless; and (2) you will soon run into copyright claims and huge damage claims, as these are protected models. What you could do however, is go to a local school, and say that you have unused printing capacity. And then let the kids design things themself, so they learn how to do it and they get enthousiastic, and then print those models. Also learn them the possibilities and limitations of this sort of 3D-printing. This would give them a lot of fun and satisfaction, and understanding. Kids are very fast in learning new technologies, even (what we consider) complex software as 3D-design. If you explain and show how the printer works, they will understand.

Most materials should be reasonably airtight if printed *slow*, in *thin layers*, and with good flowrate. So you get good layerbonding, and absolutely no underextrusion (important). A little bit of overextrusion could also help, but might create blobs. Do not use separate support materials (PVA) that dissolve: if they made strings in a print, these will dissolve and become holes. But all models will have tiny "canals" and pores where bacteria can grip and grow. It's just that the water or air won't blow through. For shell thickness, I would use at least 2 lines (=0.8mm for a 0.4mm nozzle), maybe 3. Don't pressurise 3D-prints: they might explode at a much lower pressure than injection moulded models. For huge models (1m), I would rather apply a coating, I think? Maybe even reinforce them with glass fiber mats, and then apply epoxy resin? Do tests if a 3D-print can handle the required loads and temperatures (PLA softens and warps in the sun).

If you look at the bottom, did it lift from the glass? Thus making a dent in the bottom? If yes, it could be caused by dirty, greasy spots on the glass? Clean with isopropyl alcohol, and then a few times with pure hand-warm tap water only. (No soaps.) A thinner bottom layer gives better bonding for me: 0.2mm is much better than 0.3mm. I guess because the material is squeezed more into the glass, and it has less room to escape sideways? Then use a bonding-method for better adhesion. Some people use the glue-stick, some use dilluted wood glue, hairspray, 3D-LAC,... Find a method that works well for you, and that you like. For PLA, I prefer my "salt method": wipe the glass with a tissue moistened with salt water. Gently keep wiping while it dries into a thin, almost invisible mist of salt. For PLA this increases bonding while hot, but gives absolutely no bonding when cold: models come off by themself. Then re-apply before the next print. This works very well for my typical low, flat, long models. But it is not recommended for thin vertical models like lantern poles, or for high models with overhangs. Overhangs tend to curl up, and then the nozzle bangs into them. So these models might get knocked off, as the salt can not absorb shocks. Glue is better in absorbing shocks I guess. For the full text, see here: (it is old, and I should probably update it, but it is still usefull) https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ Bottom layer of part printed with the salt method: glossy, but with tiny pits from the salt. The second pic shows the mesh of a fan reflected in this bottom (I could not get bottom and reflection both in-focus, so they are separate pics). Typical view of the glass: Inverted prisms are the edge of what can be printed: some warp but can be completed. Some come off and produce spaghetti. The tiny bottom area and huge overhangs and warping forces are a hard test. Try such a model with all your bonding-methods while evaluating them. Bottom with pits from the salt:

I haven't printed with nylon yet, so I can't comment on material-specific things. But if you are not happy with the default supports, you could design custom supports in CAD, and switch off support in Cura. In this way you can provide features to make removal easier: holes and gaps where you can insert knifes, hooks, pliers; or tree-like structures to save material and provide additional gaps for access; sideways support for higher models to prevent wobbling; extended supports with extra brim for or tiny features; free-hanging supports to not damage the plate below it; etc... Below a couple of methods I used during the past years. The pink and orange supports (center left) are extended, so I can grab them with pliers and wiggle them out. This model is way too small to insert a knife between support and model. Ribs on top allow tigher gaps, without the model and support fusing too much. Free hanging supports prevent the area below from damage, and can be cut away easily. Idem. They also reduce material-use. The bridge plates are 1mm thick. The tiny support-connection strands are 0.5mm wide, 0.2mm high, and 1mm long if I remember well, and can easily be cut. This is a small keychain, where the blue fork has to slide into the yellow slider. The orange and pink supports got an extra brim, due to their tiny ground-area (only a few mm^2). Without brim they might get knocked off, since the overhangs tend to curl up, and then the nozzle bangs into them brutally (the pink and orange supports do not rest on the yellow slider, they float). For reference: watermark text capitals height is 3.5mm.