Jump to content

geert_2

Ambassador
  • Posts

    2,084
  • Joined

  • Last visited

  • Days Won

    34

Everything posted by geert_2

  1. If your material is transparent or translucent, you could design the text as hollow characters *inside* the model, at 0.3 to 0.5mm below the surface facing the glass table. This is good for prototyping and 3D-printing, but not for mass production injection moulding of course (unless you would want expensive overmoulding with multiple moulds, like often done on tooth brushes these days). See a photo how this looks with 3.5mm high x 2.0mm wide characters: https://ultimaker.com/en/community/35918-printing-small-letters
  2. Hello all. Here is a tip that can extend the life of small fans, such as those on the UM print head or on old computer CPUs. When these fans start to get noisy, which means that the bearing is worn-out or running dry, try lubricating them instead of throwing them away. The bearings are accessible under the label on the fan. You can partially remove the label from the fan, but don't take the rest of the fan apart. I have used two methods: 1. Partially remove the label from the fan, and add a drop of oil directly on the bearing. Put the label back. Advantage: easy method, and you can see what you are doing. Disadvantages: you have to remove the fan from the print head, oil can leak away easier, and readjusting the label may be difficult or impossible, if contaminated with oil. You can also do without label, but then the bearing can contain less oil, before spilling it around. 2. Do not remove the label at all, but with a sharp point, gently poke a tiny hole in the label near the top of the bearing. You can feel where the bearing is by gently pushing on the label before poking through it. Get an injection syringe and needle (ask in a pharmacy), cut off the sharp needle-tip with a nylon disk cutter on a Dremel, and sand its edges until rounded, so you won't hurt yourself. Fill the syringe with oil, and inject a drop through the hole you just poked in the bearing's label. Advantages: the oil is better contained, and you have no problem readjusting the label. You can do this while the little fan on the back of the UM2-head is still mounted in the system, without disassembling. Just don't inject too much oil. Disadvantages: you can't see inside the label, so you are not sure what you are doing, unless you have opened enough fans in the past, so you know them well. And you can't see how much oil you are injecting. And now the bearings will run smoothly again for some time. How long depends on various factors. If you waited too long before lubricating, and the fan is slowing down, the bearing will be worn out too much, and the oil will not last very long, maybe a few days or weeks. If you do this too soon, and you use incompatible oil, or oil that is inferior to the original oil, you may shorten its life. Incompatible oils might become jelly, and no longer lubricate. I usually lubricate the bearings when the fan or motor starts to get noisy, but long before it begins to slow down. Then this method is usually good for months. I have used various fine lubricating oils in the past, such as those for model railroad trains, or for bike bearings and chains, and they all seem to work. At this moment I am using an oil for hydraulic test machines and industrial equipment (BP Energol HLP-S 46). Why? Simply because I have a lot surplus from my test machine... I don't know if it is the best oil for this, but it works. And it does not dry out easily (contrary to some other fine oils), so it lasts long. At least, it could help in case of emergency, if a fan slows down too much due to wear, and you have no replacement immediately available. Of course, most of you will already have tried this. But some people may not have thought of it yet, so I thought I would just write it down.
  3. Okay, so this is going to be heresy. We all know the rule: to minimise strings and so, you have to increase the nozzle traveling speed. Faster traveling tends to break the strings, and it doesn't give the nozzle enough time to leak. But now while trying a new PET material (new for me), I found out that this does not necessarily work well always. At my normal printing speeds and temps (good quality: 20mm/s, 210°C; standard quality: 50mm/s, 225°C; material PET, ICE brand), I got little hairs, short strings and blobs in the prints. The blobs occured during traveling fast over already printed material, when the nozzle left behind sort of little "morse code" pieces. This is due to the nozzle not retracting when traveling over printed parts. Then, when printing the next layer, the nozzle would hit these morse-code blobs and they would accumulate on the outside of the nozzle. Then they would melt, get brown, sag down towards the print, and cause strings and hairs, and bigger blobs, all left behind on the prints. It seems that if the nozzle travels too fast, its leaks do not have time to melt and fuse with earlier printed material. So they stay on top as separate little pieces, the "morse-code". Then I got the idea to reduce traveling speed to 20mm/s, same as printing speed. Result: far less build-up of material on the nozzle, far less blobs, strings, hairs. And a visibly cleaner print. At least, as long as there are no overhangs or bridges to print. This material does not like bridges: it is rather sticky and rubbery when molten, and tends to glue to the nozzle instead of pulling the bridge-line (contrary to molten PLA which is more creamy). Thus bridges and big overhangs still cause blobs and hairs, although less than before. Disadvantage: while traveling over printed parts of the model, the nozzle leaves behind a line. It did so before too (although it was rather a morse-code line, usually), but now the line is a bit thicker. But now it is equally glossy as the rest of the print, while before it was matte when the rest was glossy. I have only done a few small test prints, and only with PET (ICE brand), sliced with Cura 14.09. So I don't know if it would work for other materials, or other brands of PET, or other Cura-versions. Anyway, those of you who have problems with blobs and hairs in their prints, might want to experiment further. Let us know your results.
  4. I don't know the official guidelines, but what I do is: - Clean the nozzle after each print, immediately when it finishes. This avoids build-up of brown goo and contamination of next prints. - Regularly do an atomic pull to clean the nozzle's inside. But I do a more carefull atomic pull than most others: First, I do a manual retract of the filament after flushing some material (by pulling it back a few millimeters, similar to a retract while printing). This makes it much easier to pull out that piece of filament later on. Then I let it cool down much deeper, to room temp. Then gently twist and wiggle the filament while still cold. Then heat up again to 80°C (for PLA), and *gently* pull while still gently twisting and wiggling the filament. I do no brutal pulling. This gives less risk of damaging the rods or teflon coupler. - After an atomic pull, poke with a needle (with cut-off and rounded tip) through the nozzle, to remove any coal in the tiny nozzle opening. Then do another atomic pull. - Regularly oil the rods, and grease the Z-worm. - Blow or clean away any dust that accumulates in the machine and feeder. - Replace the PTFE coupler only when deformed (you see this in the atomic pulls). And that is about all.
  5. For objects like that, I think it would be a good idea to design custom support structures in CAD, while creating the model. It takes extra design time, but they will be easier to remove and give better results if well done. Then you could also model your supports more like a tree, and consume less material. And you can design any necessary gaps, holes or other features into the support to make removal easier, and to increase or decrease stiffness as desired. There has been a thread some time ago with several pictures of support suggestions. Maybe you could have a look at that? (I can't find it now, at this moment I get zero results when doing a search. Not sure if this is a compatibility problem with my browser (Pale Moon), or something in the search database.)
  6. In addition to the above of valcrow, if you have even more time available, way more time, you could drop speed to 20mm/s, and nozzle temp to 180°C or 190°C for PLA. Buildplate temp: max 60°C, but if it still bonds, you might try 55°C too. This reduces sagging. Also, if the filament is wound up too hard (near the end of the spool), straightening a few meters manually might help a bit too. Then you won't get irregular filament movements when the machine has to pull harder at some places to unwind it, and to force it through the bowden tube and nozzle. A more consistent flow = better quality. Rather than one big thing, it might be a lot of smaller things. Try this on a small sample first. Not on a 2 day print...
  7. Or you could design one yourself and print that. You just need to copy the part that fits into the back plate, and for the rest you can choose whatever you like. Or take one of the versions on the internet: there are some with bearings. I have designed one with bearings for my machine too, but it won't fit for yours, as it only fits exactly one standard spool of 750gr.
  8. Hello Krys, I can imagine such a straightener works very well as long as the filament is moving, but doesn't it cause bends in the opposite direction, if the filament is left too long sitting still, for example overnight?
  9. I think a problem might be the size, since you say: "at least...". So, how much more is needed? Or can you cut your designs in parts, and glue or screw these parts together after printing if needed? This is something you need to verify. Also, if you want to print lamp shades ("lampekappen" in Dutch, I am not sure if I have that correct in English) with huge openings, the required support structures to print these might be an issue. Supports do leave ugly marks after removing, so that requires postprocessing (sanding or so). Thus, in case you need lots of supports, an UM3 printer might also be an option? (I don't know about other brands, didn't follow that.) I would suggest that you download and install Cura (slicer program), select an UM2+ or UM3 printer, and try if your models can be loaded and do fit in it. Also you can see how they would be sliced, and how the supports would look. This would give you a good idea if things can be printed. You can do this without printer. Further, I would suggest that you find a 3D-print hub in your environment that has an UM2, UM2+ or UM3 (whatever you want most), and have one of your typical models printed. Costs a few euros/dollars, but it may save you a wrong buy. Or find a dealer willing to print a model (of course, offer to pay for it). Personally, I am very happy with the two UM2 (non-plus) printers we have, although I did make a few modifications (e.g. spool holder with bearings), but no essential changes. This is one of the best investments we have done in our lab to develop custom stuff for use in the hospital. While far from perfect, we can now develop lots of things that we wouldn't be able to do otherwise. Important considerations for us were: - relatively open source, or "open plans", so we can find part specs if needed, - good supply of spare parts, both from the Ultimaker dealer, but also from other manufacturers (in case Ultimaker would go out of business. I hope not, but this sector is still very much in development, with huge market changes; we have to calculate that in), - good direct support from dealer, - good forum for support, - nearby dealer, so I could go there and have a look, - dealer willing to do a typical test print (and us willing to pay for it, of course), so that we were really sure our models came out well, - ability to print different types of materials: PLA, PET, ABS, nylon, ... - ability to print materials from different manufacturers, - standard filament spools, - adjustable settings, - possibility to modify the printer if necessary, - heated glass print bed, - within budget (thus no 100.000 euro/dollar machine), - print size at least 15 x 15 x 15cm, preferably a bit more for spare, - preferably dual extrusion for support in other materials (this obviously didn't work out), - good printing quality, close to mass production quality, - not too much or too time-consuming post processing required, - no dangerous chemicals required, and no chemical waste, - no messy post production required, - reliable machine, The UM2 at that time fullfilled our wishes best.
  10. Hello neotko: the text on the blue and green plates have different sizes. Could you tell which height they have in mm (capitals height, not raised height)? And with which nozzle these were printed? When printing small text of a few mm capitals height, I always have clearly visible "circles" on each character, where the nozzle stops a moment, retracts and then lifts off. Sort of little donut-deformations. But I don't see these donuts on your characters. So, I was wondering how you did that? Concerning the image rotation: could that be in the image settings recorded by your camera (the so called EXIF or IPTC info in each JPG file)? Some image viewers auto-rotate a picture based on these settings, some don't. So if your viewer does auto-rotation, and you upload it, but the website doesn't (or vice-versa), you may have this effect. Just guessing, but it's an educated guess.
  11. By the way, apart from the bonding, keep the default settings to start experimenting. For PLA: 210°C nozzle temp, 60°C bed temp, 100% flow, 50mm/s speed (20mm/s for the first layer), 0.1mm layer height (0.3mm for first layer), etc... And good bed leveling, not too high, not too low. These defaults are good, so that should work well enough. If it fails, it is not because of these settings, and there is something else wrong. Clean the glass with whatever window cleaner you want, or alcohol. But then clean it once again with pure luke-warm tap water only! Some window cleaners or alcohols do contain chemicals or soaps that do reduce bonding, so you have to remove these too afterwards.
  12. For PLA I only use the "salt method": moisten a paper tissue with salt water and then gently wipe the build plate. Gently keep wiping while it dries. So there is a very thin and equal, but almost invisible mist of salt on it. But not too much. For colorFabb PLA/PHA and Ultimaker PLA this gives a very strong bonding when hot (60°C), and no bonding at all when finished and back at room temp. This makes removing the parts very easy, as it requires no force at all. For ICE PLA the salt method still works, but no longer perfect: corners do slightly lift sometimes. Fans can be full on, as required for PLA. No need to take the glass out of the printer, no need to clean it. Very easy. I don't know why, but a paper tissue works better than a sponge. Maybe because it distributes the salt in finer drops than I could do with the sponge, so it gives a better covering? But this is a guess. For PET I also got the salt method to work now. But I had to increase build plate temp to 70°C and use no fans. Otherwise corners lifted a lot. Now it gives a reasonably strong bonding when hot, and very little when cold. Here, without salt, the PET is more difficult to remove when cold, so the salt somehow helps in releasing. It is not perfect, but works good enough. I have tried the dilluted wood glue method too for PET: this gives a stronger bonding than the salt method, so it could also be used with fans on, but it was way more difficult to remove the print after completion and cooling down. Once I even pulled a big piece of glass out of the build plate. So I will keep this "dilluted wood glue method" for high models or delicate prints that need cooling, and use the salt method for low and sturdy prints that can do without fans. For ABS the salt method did not work, but I did not try very hard. I disliked the horrible smell of ABS too much, and it sort of reacted with the salt (or with the moisture in the salt?) and started to foam terribly the first two layers. I did not investigate this further. Maybe I will try again in the future, maybe not...
  13. For hard filaments like PLA, I usually unwind a bit of filament, then wind it in the opposite direction agains a skate board wheel (7cm diameter). Just half a turn of filament, about 15cm. Then I release it, take the next 15 cm, wind in the opposite direction around the skater wheel, release, next 15cm, etc... This straightens the filament a lot. In two minutes I can straighten about 3 meters like this. Then I let it wind up again on the spool. But now it is sitting very loose, and the unwinding resistance is near zero. Also, the friction in the bowden tube and nozzle is now near zero. Apart from straightening the filament, it also softens it a bit, so it adapts easier to any curves along its path. This avoids underextrusion too. It seems to increase life of the PTFE coupler too. (But be careful with brittle filaments, such as particle filled stuff, as they might break.) For more flexible filaments like PET it hasn't been necessary yet, but I am not yet at the end of my first PET spool.
  14. Yes, you are right. Bold does indeed work well for some "monoline fonts" where all strokes are the same width, typically road-sign style fonts, and if the characters are not too small. It is a good solution if you need text of ca. 10mm high. I have used that too. But for a lot of average fonts, bold mainly thickens the vertical strokes, not so much the horizontal. That still gives lots of problems in the narrow areas of characters like: a, r, n, b, d, etc., at very small sizes (a few mm caps-height). And it obviously doesn't work at all for serif-fonts. Maybe the reason why it is not said, is that everyone expected that everyone already tried that?
  15. Yes, search seems to work much better now, I got accurate results.
  16. In the beginning, I also had trouble that Cura wouldn't print things smaller than 0.4mm (UM2 with standard 0.4mm nozzle). So for printing small text, I designed my own character set on a 0.5mm grid. Thus all strokes are wider than 0.4mm, with some spare: vertical and horizontal strokes are exactly 0.5mm wide. Diagonal strokes have a width between 0.48 and 0.53mm. Caps-height is 3.5mm, and most characters are 2.0mm wide (except W, M, and a few others). These characters are very easy to design, and they print reasonably well if printed cool and slow enough. I also tried different variants of some characters, to see which are printed best for raised text, recessed text, and hollowed-out text (=completely inside transparant material, as a watermark). See the picture for an idea of the characters. Edit: added pictures of these characters 3D-printed as hollowed-out text, thus text fully enclosed by this transparant PET material. Here too: caps-height is 3.5mm, character width is 2.0mm.
  17. After cleaning, but before assembling it again, it might be a good idea to check tightness with compressed air and a gas pipe leak detector (spray can), or with water? Then you can see if there are still any leaks. I don't know if adding heat resistant copper paste on the threads is a good idea? or some similar anti-leak and anti-lock stuff? Like they do in chemical industry?
  18. I also have this on overhangs. It seems to come from edges of overhangs curling upwards, and then being pushed down irregularly when the next layer is printed. Printing cooler helps, but does not totally cure it. What also helps, is placing a fan in front of the printer, or blowing with a gentle stream of compressed air on the affected parts (especially for very small parts, not practical for large parts). Sometimes you can go down to 190°C at default speed of 50mm/s for PLA, depending on PLA brand, color, state of the rest of your machine (eg. teflon coupler), and the friction or unwinding resistance in the spool. Printing slower generally improves model accuracy, but I do not know if it has an effect on overhangs, I haven't tried that.
  19. Nylon is known for being very difficult to glue and paint. You will definitely need a chemical activator. This makes the surface chemically active and able to chemically bind to other stuff like glue and paint. Otherwise, it won't work. (Nylon can be colored in the mass by soaking it in colored water, but that is probably not what you want here.) I have no idea where to find activators, or which brands. I think you would best contact service-engineers of big nylon suppliers, like BASF or Bayer. Or search their sites for info. These big companies do deliver nylon pellets to all plastic parts manufacturers and injection moulders, so they are most likely to know.
  20. If your design has some hidden areas which can not be seen (or which do not matter), then maybe you could try designing a sort of "zip-fastener" into it? One zip in PLA, the other in PVA? So that the PLA and PVA alternately hook into each other, and they mechanically grab and keep each other in place? I don't have an UM3, so I don't know if this would work. Might be worth trying?
  21. Forgot to mention in my earlier post about silicone socks: another important thing with a silicone sock is that it may have a high friction (like most silicones), and it may rub off the first layer from the glass. Otherwise, using a strong adhesive silicone like Tec-7 might have been an option. Concerning PTFE, I have tried PTFE oil from a spray can. Sprayed it on a tissue, and then wiped the nozzle. But that doesn't get baked on, and it does not stick to the nozzle. It actually gave worse results than wiping the nozzle with a tissue with silicone oil. I have no idea what baked-on PTFE would do; this could be an interesting experiment.
  22. Which version of SketchUp are you using? I should have mentioned it, but I was talking about the freeware version Google SketchUp 8.0 (the latest freeware I think?), with which I had bad experiences for 3D-printing. But of course it may have been improved in the professional versions after it was taken over by another company. Google bought it for designing surface-3D-models of buildings, for its Google Map and Earth projects. Watertight 3D-modeling for 3D-printing was not on their priority list back then. I don't know what the focus is of the new owners?
  23. The curling is greatly reduced - but not eliminated - by wiping the nozzle with silicon oil. The kind of oil that is also used for lubricating car door rubbers in winter, to prevent them from freezing up, and similar. I found it in a car shop. The filament still starts to curl, but since it sticks less to the nozzle due to the silicone oil, it tends to fall back down soon. For wiping I use a sort of quite strong paper tissue, as used in laboratories. I don't know the specs, since it is old, and I don't have references anymore. But any strong and rough paper tissue should do, I think.
×
×
  • Create New...