Jump to content

geert_2

Ambassador
  • Posts

    2,084
  • Joined

  • Last visited

  • Days Won

    34

Everything posted by geert_2

  1. That exFAT is probably going to be the problem? I would suggest you buy a couple of thumbnail 8GB or 16GB sticks, and just leave them with the printer. I guess USB2 would be preferable over USB3? Theoretically it should make no difference, since USB3 is downward compatible, but you never know? And speed is obviously of no concern here. :-) They cost less than 10 euro/piece these days. A tiny little thumbnail stick doesn't stick out too far, so less risk of hitting and damaging it. Make sure they are FAT32. I once tried reformatting a 64GB exFAT stick into FAT32 (for another purpose, not for 3D-printing), but that didn't work well. Had to reformat it back to exFAT.
  2. I never tried it, but I would think the "dot-method" would also work with stacking, on the condition that you merge your model-stack in CAD. And export the whole combination as one STL-file. But obviously you are going to need a lot of support material if the models would be irregular, and I think the risks of something going wrong could increase as you get higher in a stack, e.g. support and model not bonding well, or underextrusion, or running out of filament... So while I do understand the desire to print as much as possible in one batch, especially if you are on tight deadlines or have a high volume to print, it is not something that I would like to do for myself. But it might work well for flat models like coins, or walls of architectural models. So this would be the concept, in perspective and side view (just a quick and dirty sketch, don't mind the non-optimal positioning):
  3. I have two UM2 (non-plus, thus the version before the UM2+) printers and they are still fine for what I need them for, mostly long flat models that need no support. So if you only need one color and no dissolvable support, it should be okay. But I would suggest that you make a testprint first, to see if everything still works fine, and the printer isn't abused or damaged (e.g. by dropping, printing abrasive materials, things burned out, damaging mods,...). This is usually repairable, but then the costs might no longer make it the best choice.
  4. No, I didn't change the flowrate, so it was the standard for PET on my UM2, and I left it at 100% on the machine. I had to look up temperatures for these test blocks (I had written it down somewhere): - all were 215°C, except: - 0.06mm layers at 10mm/s (unpolished block) = 210°C - 0.06mm layers at 10mm/s (polished block) = 200°C - 0.4mm and 0.3mm layers at 50mm/s = 225°C due to the much higher extruded volume The recommended temp range by the manufacturer is: 215...250°C, so I am at the lower edge. I once tried higher temps too, but that gave more bubbles and more stringing, and faster decomposing and browning. By the way: test blocks are 10mm x 20mm x 10mm, and watermarktext is 3.5mm caps height. I don't really know about part-strength vs. layer-height. But from these tests it was clear that thicker layers and faster speeds gave more "frosted glass effect". That means more voids, and probably less strength (but I am guessing now). Printing at 0.4mm layer height was still possible, but then cooling became an issue: the block didn't cool down fast enough and it started deforming. I would suggest you run a similar test, but maybe with a testmodel more suited for your application and more similar to your final models (e.g. with overhangs and bridges if you need these). This gives good understanding of your printer and your filament, and you will win back that time investment very fast.
  5. On my UM2 (=single nozzle) I get watertight PET parts by printing at low speed, low temp, thin layers, no fan. Usually 25...30mm/s, 215°C, 0.1mm layers, bed 80...90°C, 0.4mm nozzle. But 0.06mm layers are even better. At slow speed, print cool to prevent the PET from decomposing and getting brown. Layer bonding is excellent: when overloaded, fractures run diagonally through parts, disregarding layers. Main disadvantage of no fan, is that overhangs and bridges are terrible since the material won't bridge well. But for my long flat models, this usually is no problem. If you would use PVA-supports on an UM3, any PVA strings might cause holes in the PET. Consider using PET for supports (thus a single material). Print a few small test parts before doing a large real part. Pictures: Test blocks in PET, layer height from left to rigth: 0.4mm, 0.3mm, 0.2mm, 0.1mm, 0.06mm. Top row printed at 50mm/s, bottom row at 10mm/s. All 0.4mm nozzle. Slowest and thinnest-printed test blocks are getting brown due to slow material flow in nozzle. Testblock at 10mm/s and 0.06mm layers, left as printed, right polished. Transparency indicates good layer bonding and very few internal voids. Absolutely watertight. The 3D model. Text is a hollow watermark, sitting halfway in the testblock. Keychains in PET. Ruler units are in mm and cm. Fractured carabiner hook: fracture pattern radiates in all directions, mostly not following layer lines, indicating good layer-bonding. Idem, other side. ----- end -----
  6. Ik denk dat niemand nog een raft gebruikt. Maar waarom zou je niet gelijk op de glasplaat printen, dat werkt prima? Als je perse een soort raft-achtig ding wil, kan je misschien een dunne plaat ontwerpen in CAD, en die printen met beperkte infill (ca. 50%? - Probeer op kleine teststukjes), en zonder top layers?
  7. Maybe move your bed closer to the nozzle, or make your first layer thinner, 0.1 or 0.2mm? I get really glossy bottom layers with both PET and PLA. Print on a glass bed, not on tape or some other rough surface. Hard to see in the photos below, because the reflection is out of focus, but these bottom surfaces are high-gloss and do reflect like shiny injection moulded parts. The "copyright design ..." watermark text is hollow and sitting 0.5mm below this glossy bottom surface. For reference: text caps height is 3.5mm, character legs are 0.5mm wide. High-gloss bottom surface of keychains. The ruler is in mm and cm.
  8. I don't know your printer, so this is guessing. To me this looks like a hardware issue. Maybe some connectors that are not seated well or are oxidated? Or electromagnetic noise on the power line or environment? Motors starting or stopping causing pulses, or sparks you get in cold dry winter weather? Or a cell phone going off nearby? This sort of things can easily lock up electronics. (Cell phone often locks up my USB-keyboard and mouse.)
  9. If you want to examine a 3D-print in detail, but you don't have a microscope or good macro lens. Try using your webcam or smartphone, and add a close-up lens in front of it. Then watch the picture in full size on your computer screen. You could use any old camera lens as close-up lens, provided that it is concave and does not cause too much distortion. Or buy a dedicated smartphone adapter with close-up lens. Magnification will differ depending on its focal length. If required, design and 3D-print a lens holder for mounting the lens onto your camera. The results are not perfect, but not too bad either. Sharpness is okay. There is considerable cushion distortion, but that is not too important for this kind of pictures. The photos below are from a Logitech C525 webcam, and an old ocular camera lens as close-up lens. But the concept should work with any similar webcam or phonecam. Test photos: 3D-printed keychain with watermark text inside. Text caps height is 3.5mm, leg width is 0.5mm, and text is sitting 0.5mm below the surface. Material is transparent PET. Underextrusion test. This is at 50% flow rate. PLA. Insect antennas on a print, after I forgot to enable retraction. A standard Bic ballpoint tip. Micro-electronics from an old video camera (good for recovering lenses too). The resistors (little black things with numbers) are ca. 0.6mm wide x 1.7mm long. Fingerprints. Flintstone-text: here caps height is 3.0mm, leg width is 0.42mm (in the design, probably more in print), printed on an UM2 with 0.4mm nozzle, 0.1mm layers, 25mm/s, 200°C, PLA. The equipment: Logitech C525 webcam, with an old camera ocular lens attached. Obviously, I still have to design and print a decent lens holder... But any similar equipment with small cam and lens should work. Test subjects, with ruler in mm and cm for reference. ----- end -----
  10. I don't know your printer, so just guessing... Maybe a blocked nozzle, and then the filament cooked, burned, and spilled all over, due to sitting too long in the hot nozzle? Or something along that line, where the filament got stuck? Some materials burn away rather cleanly or leave black coal dust (like the PLA that I have), but some burn into a sort of glossy varnish that is hard to remove (like the PET that I have).
  11. In the beginning I also had this problem. But now it is extremely rare, even though I have lots of models with small openings. Maybe you can solve it by adjusting nozzle height a bit closer, or changing temp or speed of the first layer? Or improving your bonding method? Also, a thinner first layer gives better bonding and less risk of the first outline being ripped off, on my printers. A 0.2mm layer sticks *much* better than a 0.3mm for PET and PLA: this is just my observation, but I am not sure why.
  12. The whole world has switched to the metric system officially, so I think we should stay with the metric system. And not go back to medieval units that don't work well in high-tech environments, no matter how much sentiment there is in it. I grew up with "horsepower" for car engines, which now has changed to kilowatt. I regret this, because the horsepower-number is higher and thus more impressive than kW, and because I grew up with it, but kW makes calculations sooooo much easier. No more weird random conversion factors. The USA has switched to metric in 1875 (yes, 18...), although some politicians and press are still not aware of this, and the UK has switched more recently. Asia is metric since long (except a few former British colonies), and it is quickly becoming the dominant world power. As soon as they have enough power, I am very sure they will enforce the metric system even in the aviation industry, the only industry that currently is still partially imperial (and then even with a bizarre mix of British and US units, and metric, causing confusion and accidents: some time ago a plane ran out of fuel and crashed, killing all souls on board, after confusing metric and imperial fuel units when refueling). If we don't want to be left behind, I think going fully metric is the only option, even though it might hurt our sentiments.
  13. I don't have resonance problems (2x UM2 non-plus), so I can't test this. But I think this is an interesting approach. The only question is: if people notice resonance, how are they going to find out which frequency it is exactly? If I feel a vibration, I can't tell if it is 5Hz, 7Hz, 10Hz,...? At least not without a known reference next to it. Further, it is very hard to feel/hear whether there are lower or higher harmonics on it, which could also trigger the resonance. Maybe you could find a printing test pattern that "resonates well"? Or a printing pattern that causes vibrations at known frequencies, so that people can try this, and then can compare their own resonance to these standard patterns? I am thinking of a gcode-file with a sort of "frequency stepping scale" similar to the E24 resistor values in electronics? Where the actual resonant frequency is displayed on the front panel? And then, if that frequency resonates in the system, people can just read the value from the front panel, and enter that in Cura? Could be interesting from a theoretical viewpoint, but I am not sure if it is worth the effort though, and if it would really result in better prints...?
  14. A few years ago I lubricated the edges of the belts with hard silicone grease, with the same effect: sound gone. It seems to be the edges running agains the flanges that cause it. Only lubricate the edge, not the belt tracks, otherwise they might jump teeth. And don't use petroleum or plant or animal oils or fats: these might damage the rubber. Hard silicone grease is chemically almost inert, and it does not leak away. It is the same sticky stuff used to lubricate movements of binoculars and microscopes, in which you can't have oil leaking onto the lenses.
  15. While I don't have much experience with Ultimaker PLA colors, I only used a few, I have seen a similar effect in other brands. It seems the pigments and other additives to create the desired color, such as filler particles for white, gold, silver, glitter, pearl,..., have an effect on viscosity, stickyness, optimal temperature and flow, etc. Some colors tend to give more strings and hairs, some tend to stick more to the nozzle leaving residu, some tend to show layer lines much harder, some tend to create more blobs,... Slightly transparent colors hide the layer lines a bit better, while white and shiny colors show them more.
  16. To me it looks like your biggest problem at this moment is basic handling of a Windows computer. Starting with 3D-printing is then like trying to do aerobatic stunts in an airplane, before you can't take fly it safely. I think you would make much faster progress if you would first follow a good course on using computers, just the general principles: how files are handled in Windows, how and where they are stored, zipping and unzipping files, installing programs, etc... Yes, this will take you at least several days, maybe a week. But you will soon win that time back 10 times.
  17. I have had this with PET, when using a bonding of dilluted white wood glue. Bonding was absolutely excellent, too excellent. It chipped while cooling down, even without pulling. Now I wipe the glass with a tissue moistened with salt water, prior to printing. This method slightly reduces bonding for PET, but avoids chipping (contrary to PLA where the salt-method greatly improves bonding). PET models that need lots of cooling now tend to warp slightly, but models that can be printed without any cooling come out fine without warping, and without chipping, such as my typical long flat models. After cooling down, they come off easily. If you try this, stay with the printer to see if it works for your models, and if they don't pop off in mid-print. No issues with PLA so far.
  18. Oil *could* be a bad idea, depending on the oil and silicone combination. Some chemicals inhibit curing of silicone, such as sulphur, mint, some oils,... Try compatibility with a little bit of silicone beforehand, or otherwise you might end up with a half-cured mess that is very hard to clean from the mould. Further, make a small test piece first with the same angles of side-walls as the real model, to see if you need smoothing of the walls. Silicone flows into the tiniest pores, so it gets a very strong mechanical grip, even the non-stick mould-making silicones. For complexer models smoothing will be required, just to be able to get the model out of the mould, unless it is very simple and flat like coins. Also preferable use platinum-cured silicone (additive cured). No tin-cured, as that tends to decompose over time, and is only good for short term use.
  19. Up till now PET has survived undamaged and unwarped in my car, even in our recent very hot summers with Mediteranean weather, quite unusual for Belgium. Contrary to PLA that even did not survive a mild spring or autumn sun. I think PET can have around 80°C? If you would need it for autoclaving and desinfecting at 140°C, that would be a different story.
  20. Excellent info from p-kimberley above. What is said, is absolutely true: any layer lines, blobs or irregularities will show up in the cast, and they will make it *much* more difficult to remove the cast from the mould. I have made silicone moulds from 3D-printed models. It goes as follows: first 3D-print a model, including what will become the filling canals and venting holes. Print in multiple parts and assemble as required. Carefully remove all blobs and irregularities, and smooth the model, as any defects will show up. Stick the model to a plastic base plate, or provide some form of support or hanging it stable, so it does not go swimming away when you pour silicone in later on. Using plasticine, wood, or plastic plates, build a container around the model. Make sure the container is absolutely watertight at the bottom and sides. Make sure it has pouring and venting openings (or leave the top open). And then pour silicone in the container around the model, and let cure. Then cut the model out of the silicone by cutting zig-zag lines, so both mould parts align themself correctly later on. Now you have an empty silicone mould, in which you can pour resin. Spray the silicone with silicone oil as release spray, prior to casting. Silicone is watertight because it repells water, but it is *not* oil-tight and not solvent-tight. So the solvents of resins creep into the silicone and destroy it soon, after a few times of use. Therefore you really need to saturate the silicone with release spray, prior to casting. As with any moulding and casting, keep aware of undercuts, use slanted side-walls, and avoid enclosed features and stuff that could make removal of the cast difficult. I have also done the opposite: cast silicone models in a 3D-printed mould. Thus: 3D-print a mould in PLA, carefully remove defects, smooth it. Then carefully seal all openings in the seams in side-walls and bottom. Pour silicone in and let cure. If you do not smooth the layer lines, and close tiny gaps in the model from the 3D-printing process, it will be almost impossible to get the model out of the mould. Don't ask how I know. :-) You can tap or shake the silicone to remove bubbles, or gently (!) blow on bubbles. Gently pouring from a high distance in a thin stream also helps. Vacuum degassing is best, but only works for slow-curing silicones, not the fast ones I have, and you need professional vacuum equipment. On Youtube there are tons of excellent tutorial videos, showing the techniques. Search for: moulding and casting. You find info on really anything: thick silicone paste (like for body masks), thin liquid silicone, slow and fast curing, epoxy and PU-resins, coloring, info about avoiding bubbles, and so on. Be sure to study a lot of these, even if it takes a few days: you won't regret it, and you will very soon recuperate that invested time. Preferably download good videos and store them locally, so they won't be affected by "linkrot" and error-404. For smoothing PLA-prints, you can use the "cloadfiend method", optimised by user cloakfiend. See his topic on this forum. A few photos (I have them, so I can as well show them): 1. Silicone moulds, with a hard shell for more stability. The original models are not shown here, and were made from plexiglass (right), and very hard wax (left), but you could use a 3D-printed original as well. This was before we had 3D-printers. Notice the zig-zag cuts, so both parts align well. Also notice the notches in the shell, for alignment. And see the pouring and venting openings. The blue model has a base-plate in the shell, so it stays upright. 2. Orange 3D-printed models, and green silicone impressions of them. Top one is as-printed, bottom one is smoothed with acetone (using the "cloadfiend-procedure"). The difference in the silicone impressions is obvious. 3. Same model 3D-printed, but with different post-processing. Left = smoothed with acetone. Center = smoothed with heat-gun: don't do this: it causes craters due to internal bubbles exploding when heating the plastic. Right = as-printed, untreated. 4. Idem, different camera angle.
  21. Maybe you could do it with a trick? Put four tiny dots outside of each corner of your models. So that they form an imaginary rectangle that encloses your design. Or draw a custom skirt around your design; just one layer height is sufficient. Then that surrounding skirt or border should always be placed in the same position on an empty bed. Keep that skirt, or those four dots, in place, and then correctly place the rest of your components in this space, and save as separate files. When slicing, the skirt (or the four dots) should determine the position on the glass. At least, that is what I would expect. Try this with a small test piece first, before wasting too much material on a real big model.
  22. The cones were printed on my UM2 printers (orange on printer nr.1, white on printer nr.2), several years ago. As far as I remember, they all have 0.1mm layer height, but they may have been printed at different speeds, flows and temperatures (I don't remember). The main goal at that time was to see the effect of using a dummy tower (=the square tower) to increase cooling time of fine details. Otherwise the hot nozzle stays on top of the model, so it can not cool down and solidify. I have the other model printed as well, but I don't have photos. I will see next week (too dark here now). STL-files: (the cone is named "pilaartest" here). Print them alone for the hardest test. But if you would want to test the cone with a dummy tower for comparison, then load the dummy next to the cone, and scale its height so that it is just a few layers higher than the cone. overhangtest3d.stl pilaartest1.stl dummy_tower_10x10x30mm.stl
  23. How do they handle bending loads, and impacts such as hitting stones on the beach?
  24. Now it looks a bit like those chocolate coins wrapped in silver paper, that we used to get for Santa Claus. Isn't there a way to make the plating chemically bond to the print? Somehow etch and chemically activate the outer layer? Or plate a first layer with a mix of glue and metal, or something along these lines? How do they do this for plastic car "chrome" strips, wheels and radiator covers?
  25. Maybe you could consider this: make a test design that includes all your typical problem areas. For example: long flat bars (do they bend?), thin pilars (smooth, stable?), steep overhangs, small holes (size?), smooth curves (layer lines?), fine text, watermarks, sharp corners (ringing, rounded?), thin plates, or whatever else is typical for your designs. But keep it relatively small so it prints fast enough. And then have that test model printed at: (1) the printers best quality, (2) default quality (well balanced between speed and quality), and (3) fastest draft quality. Of course, offer to pay for these models and time. When someone comes to me in the laboratory to have something designed and printed, the first thing I do, is show them the *limitations*: what it can not do. As in the test models below. And only then I show a couple of prints that went very well, and some average prints. So they get a balanced view, they have realistic expectations, and they know what post-processing will probably be required. No surprises. It would be better if sales people also did this (some do, but not all). Concerning demos: I recommend seeing the whole workflow, from slicing, preparation of the machine, printing itself, post-processing (removing supports, cleaning bits), and cleaning the machine (some liquid resin printers take half an hour daily to clean). You don't need the sales guy standing there all day, you could offer to walk along with the operator printing the demo models. That guy might even appreciate the chat. If the sales rep refuses, you should ask: why not, what is there to hide? Edit 2019-12-16: added photo below:
×
×
  • Create New...