geert_2

Good idea. Such scales should have been moulded-in on each spool, for both 2.85mm and 1.75mm filaments. That could easily have been done.

Heb je alle punten op gr5's checklist al geprobeerd? Die is ergens te vinden op dit forum, maar ik weet niet precies waar. Daarop staan zowat alle mogelijke oorzaken. Ik zou in de eerste plaats denken aan versleten teflon coupler, 3e fan (kleintje achteraan) die niet goed werkt, te dik of slecht filament, te snel of te koud (voor die snelheid) printen. Maar er zijn er nog tientallen.

Indeed, years ago when deciding if and how we were going to step into 3D-modeling and printing, a brand with a good forum was a make-or-break point for us (this in addition of course to other breakpoints like: open filament system, multiple materials, relatively open source, good availability of spare parts, nearby dealer,...). We knew that the learning curve would be high, that we might run into problems, and that we might need expert advice from more experienced users. So before deciding on any printer, I visited the forums first. And I went to see the dealer.

Using two-part moulds will for sure cause the silicone to leak, if it is a slow curing one (>10min). A very fast curing one (<5 minutes) might not leak too much. Unless you fill the seams. Silicone leaks through microscopic pores. A common method to seal the seams, is to use plasticine or wax. But it has to be a plasticine that is compatible with silicone, it should not contain any sulphur (plus a lot of other stuff) which inhibits curing. Do a compatibility test before with a little bit of silicone and the plasticine or wax. You can find good tutorials on Youtube. Search for: silicone mould making and casting. Concerning details: if you use good silicone, *everything* that is visible in the mould, will be visible and mirrored in the casting, just everything. Silicone reproduces up to microns. So: all layer lines, underextrusion gaps, blobs, strings, fingerprints in plasticine parts, nail marks, hairs, dust, any defects,..., all will be horribly visible. And you may have entrapped bubbles if you don't remove them by vacuum after mixing both parts of the silicone. Another way to reduce bubbles, is pouring the silicone in from very high in a very thin stream. And gently (!) blow on bubbles, or prick them with a needle. If you have a shaker while pouring in the silicone, that also reduces bubbles: maybe try an electric tooth brush, or your girl friend's pleasure toys. :-) Do a lot of small test prints, before doing a big one, just to get familiar with the materials.

I haven't had NGEN warping and coming off the bed, but I haven't printed big items either: mine took maximum 3 hours and were low flat models. PET sometimes did warp if printing on bare glass, in the beginning when I used 100% cooling. Now when using no cooling, it does not warp anymore. But here too, my models are low and flat, and take maximum 3 hours. So, try to print without cooling if the model allows it (=no overhangs, no bridges), or try a good bonding method. I used gr5's bonding method a couple of times (=white wood glue 10% dilluted in water) on models that needed cooling, which gives excellent bonding. But that bonding is sometimes too strong and might chip the glass, depending on the model... The problem with no cooling for PET is that it tends not to pull nice strings when bridging, contrary to PLA. PLA gives really nice bridges. But PET-bridges tend to snap and fold back onto the nozzle into a blob: it is too rubbery. So the gap or bridge isn't closed. And that blob gets deposited onto the next wall when the nozzle arrives. The accumulation of blobs will finally close the gap, but it is ugly. At least, that is with the PET I have here... So it may take some trial and error to find the right balance. Before doing long prints, make a small test print that will undergo some load, for example a clamp. And try if that survives well in your car, under load. Or try it in a test-environment at the same temperature.

Before ordering on any website, always check if they have an official address listed on their site (company, street + nr, city,...), phone and email, officially registrated business numbers, etc... And then look it up in an official phone directory, and on Google Maps. If you can see the company logo in the street view on Google Maps, it is probably okay. However, if the company is located on an abandoned garbage pile on Google Maps, better be cautious.

I usually design custom supports, so that I can make sure that I can get in there with a knife, or get in with pliers and hooks to pull out the support. In custom support I can also increase or decrease clearances as desired. Some of my usual examples, in case you need ideas: Various support concepts. The ribs (0.5mm wide, 1mm aparat) on top allow tighter clearances. Orange and pink are supports: they have extensions so I can grab them with a plier, as the model is too tiny to get in with a knife. The brim is for added bonding, due to the large overhangs (they float above the yellow). A totally free hanging support: easy to remove, consumes very few material, and does not damage the underlying parts. The inverted staircase at the underside of the support reduces its tendency to curl up, compared to a triangle shape.

Maybe you could redesign it so that it somehow latches after attaching, with a sort of snap-fit lock? So it would become a closed circle. That would largely prevent deformations. Almost all plastics will deform under continuous unbalanced load, and especially PLA and light-cured resins.

Usually I also start with a clearance of 0.2mm, and then adjust as trial and error indicates. However this is on a 0.4mm nozzle. As Smithy says, small holes tend to get closed down a bit, and small rods tend to get too big (slightly overextruding). Also layer height has a big influence, as well as the amount of blogs, strings, ringing and other defects. Probably you will need to post-process tiny holes anyway. I always go through them with a drill (see pic): manually, not electrically, because that melts the plastic. The tiny purple things probably won't print correctly at all, so you would be better off using standard nylon screws or pins, or whatever you need.

Yes, sometimes designing the supports takes time. But if that makes the difference between a succesfull or a failed print, or between almost no post-processing or a post-processing nightmare, I think it is worth the effort. I like designing more than post-processing... :-) Concerning the exact dimensions and gaps I would say: design a test piece in which you incorporate several variations, and try which works best for you. For me, the ribs on top work best if ca. 0.5mm wide, with horizontal gaps between 0.5mm and 1mm. Vertical gaps of 0.2...0.4mm usually work well for me. Smaller gaps give better accuracy, but make removal more difficult and vice-versa. This is all for single nozzle printers like my UM2; I have no experience with dual nozzle (I guess then you could make the gaps far smaller?). It also depends on materials, models, printing temperature (cooler is better), layer height, speed. So you need to find a balance that works for your models and materials. Takes some trial and error. Below a new one from yesterday which also works well. Here the supports are totally free hanging, with very small gaps. It works due to the stringing and the sagging first layers of the support just sticking to the side walls. This support causes very little damage to the rest of the model and consumes only a very small amount of material. Edit: to reduce overhangs curling up: thicker layers tend to curl up less than thin layers: a 0.2mm layer is *much better* than a 0.1mm. Also printing slow and with lots of cooling helps. Instead of an inverted triangle, an inverted staircase also curls up less (as in the support shown above). All these things help, but do not eliminate the problem.

Even if you find a material, it is not going to pop up automatically, if printed flat. On the contrary: even if you would pull it up manually, it will tend to snap back into the flat shape as when printed, because that is how the molecules were solidified. You would have to print it in the upright position, and then manually push it down to make it pop up again from itself (if the material is flexible enough to survive this). You could do this by designing a custom support structure that does not consume too much support material.

Maybe like in the old days? Put a paper logbook next to the printer, and for each print, write down username, date, begin and end time, estimated print time (from printer display), meters of used filament, type of filament, etc. And a separate column for total times. This method has worked well for hundreds of years, and is likely to be more stable than electronic means. The data would have to come from the printers anyway, not from Cura. Often a sliced model is printed numerous times. And some others are not printed at all, when changing our mind or improving the model. Cura doesn't know this.

For special models and difficult to reach areas, you could model the supports in CAD in the design itself. Then you have full control and you can adjust it to your needs. See these examples for my single nozzle UM2 printers (I have no experience with dual nozzle): A few different support concepts I have used: some with extensions, so I can grab the support with pliers; some with holes to insert pins to pull the support out, some with layers that peel off easily, some with several tiny blocks instead of one big block so I can wiggle each tiny part loose, some with overhangs to improve the first real layer of the bridge, etc... These pink and orange supports extend from the model, so I can grab them with pliers, because the model is too small to get in there with a knife. Also they have a custom brim (footplate) for better sticking to the glass due to their huge overhangs: I do not want them to be knocked off when the overhangs curl up and the nozzle bangs into them. The ribs on top reduce the contact area and make removal easier, and they reduce sagging (the blue tray has to slide in the yellow part). These supports hang to the side-walls, so they don't reach all the way down. This is usefull if you want to reduce the amount of support material, or if you don't want the lower surfaces to be damaged by supports. The concept is based on an idea from another user (was that "smartavionics"?). The bottom layers of the center support bridge will sag a lot, but that does not matter, as they are discarded anyway. The inverted staircase shape reduces curling-up of the overhanging edges of the supports.

I would suggest you look up some tutorials and demos on Youtube of the most popular free CAD programs. And then see if the workflow appeals to you. For my designs where I have to make frequent adjustments to technical models, DesignSpark Mechanical is good. But it is only suitable for geometric forms, not for organic form. And it is very easy to learn. Not in the list above: if you want to make organic forms, and you don't mind a steep learning curve, then Blender (also free) might be an option.

I think this should be a standard for new printers, at least for single nozzle printers. But it might be complex with dual nozzles. Maybe only the coupling halfway close to the head should not be there, it adds complexity.

I don't know if it is technically possible. But mechanically, the print head is so bulky that the nozzle can not come close to any object anyway, without hitting it. So I think it would be best to print the models separate, and then "collage" it together with the rest of the collage?

Yes, I know the experience with PLa demo-models left in the car... Now I use PET and NGEN for parts for my car: dash cam filter mount, clamps,... Up till now these have survived unwarped. These materials are still easy to print.

Thanks for the tip. I once tried to grind down a skaterwheel to make it smaller. But after a full day of grinding on an agressive water-cooled disk, I still had gotten only 1mm off. That was when I realised that skater wheels are supposed to be *very* wear resistant... :-) Something else I just think about: any 3D-printed suction cup is likely to be *not* watertight. So it might not work at all. The indents caused by the layer- and printing lines will allow air to seep in and destroy the vacuum. Even in glass lab vacuum equipment, the tiniest scratch will quickly destroy the vacuum. So, even if you make a 3D-printed mould for casting, that mould would require post-processing to remove or fill the indents. Even a very smooth bottom like this won't hold vacuum at all. Unless you have a strong vacuum pump that keeps running all the time.

You should be able to see this in the layer-view in Cura. But you always have the option to design your own custom supports in CAD, and switch off automatically generated supports. Then you have full freedom, and you can design-in features to make removal of the supports easier, or to increase stability, or whatever you want. See the pink and orange supports here, which extend from the model so I can grab them with pliers. This is a very small model, too small to get in with a knife and cut the supports loose. I also provided an additional baseplate to make the support more stable, so it is not knocked over due to its long overhang. Same in red-cyan 3D, which gives better depth-perception:

Multiple models should of course improve cooling, so it can help. But I think the bow wil always be problematic, even with cooling towers. Each time the nozzle lands on or takes off from a surface, this will leave a mark and risks causing a blob, string or hair. This is an area where filament-based printers are not optimal, and liquid resin printers are better, especially those optimised for juwelry. Another thing you want is a very constant flow through the nozzle. So that the melt always has the same temperature and same viscosity. This is why I switch off features like "minimum layer time", which slow down the printer. But then the melt stays too long in the nozzle, gets hotter, and tends to cause more strings and more cooling problems (in my models). Sometimes changing the printing-orientation can also help, e.g. printing it on its back, depending on model geometry and printer (single/dual nozzle). Placing a desktop fan at lowest speed a meter in front of the printer may also help. This gently evacuates built-up hot air, without cooling the nozzle down too much. Playing with retraction settings can also have an influence, but I have never done that, so I can't recommend anything. I just thought about other pictures I had: very small figurines (>2 cm), some printed upside at various speeds and temps, and some printed on their back. The red one was printed slowest and coolest, but I don't remember settings. The hairy one was printed at normal settings (=too hot and fast for these small models) with multiple models at once. The "mummified" one was printed standalone, at normal settings (=too hot). These were all printed with a 0.4mm nozzle, single color, UM2-printer.

Have you considered designing and 3D-printing a mould, and then cast the cup in silicone or polyurethane? Silicone is *very* grippy, temperature resistant, abrasion resistant, but not tear-resistant (tears apart easily). It is water-tight but not oil-tight: oils and solvents seep through slowly. Silicone for casting is usually non-stick (=does not glue to other parts). Be sure to use platinum-cured (=addition-cured) silicone, no tin-cured, because that is not stable and only good for short-term use. PU is usually less grippy but still good, but far stronger than silicone. Think of skater wheels: it is very hard to grind or cut them (don't ask how I know). PU usually bonds very strongly to other parts: so you can cast it around a base and get a good chemical bonding to it.

Also not seeing the pictures here. You can usually drag and drop, or copy and paste, any pictures into your message from Windows Explorer or from any document or website. Maybe the effect you are getting is this? Not enough cooling due to printing on a too small area, so the nozzle stays on top of that one area, keeps radiating heat, and does not allow it to cool down and solidify? Printing slow, cool, and with a dummy tower next to the real model (or multiple real models at once) helps a lot, but does not eliminate the problem.

What about changing the current link name "Join our community" into "Community / Forum"? This would increase the visibility and discoverability of the forum, I think?

This is an interesting thought indeed, but I doubt it is the case here: it happens way too slow over the course of *hours*, not seconds or minutes like bimetals in temperature gauges, or in old flashing lamps and turn signal lamps in cars. And the oven is fully closed, temperature controlled, and heated from all sides. So the samples warm up pretty fast and equal from all directions, initially without warping: this only starts later. So I think the cause is rather to be sought in the direction of slow relaxation of the stresses in the long polymer chains; stresses which were baked-in due to the uneven cooling during printing, line per line and layer per layer. Plus a gradual change in crystal-structure (similar to differences in crystal-structure in metals when they are cooled quickly or slowly, or reheated). That would explain the shrinking in length and the widening. But it still does not really explain the warping in the *inverse* direction after hours for some samples (not all), after they initially warping in the expected upwards way. At least, I don't see it. :-) So I would welcome further ideas on this. Anyway, when heat-treating plastics, people need to take (the risk of) deformations into account, and maybe iterate their design to get the correct dimensions after treatment. Nylon and other plastics can also have different crystal structures, so the same effects could play there I think.

I think with embedded text you will more likely run into issues with strings, blobs and gaps, ruining the aesthetics, than with raised text. Recessed text (pits) is also likely to give issues: the pits getting closed-down, just like small circular holes. The smallest text that I can print and that is legible is 3.5mm high characters (caps height), with legs of 0.5mm wide. This on an UM2 with 0.4mm nozzle, and of course single nozzle and single color. So it has to be recessed or raised text, or hollowed-out (watermark) in transparent materials. This is *not suitable* for aesthetic purposes, it doesn't come close to the beautifull prints of XYZDesignPro, but it is good enough for simple technical labels, such as on the back or bottom of equipment. The reason why I use 0.5mm wide legs instead of 0.4mm (nozzle size), is that in corners the legs often get a little bit narrower than 0.5mm, e.g. 0.48mm (and sometimes a bit wider too, but that doesn't matter). If this was on a 0.40mm leg and it would get 0.38mm, it might not slice correctly and too thin parts might be dropped. Printing slow and cool also helps improving quality. I have a couple of test-plates with recessed, raised and hollowed-out text. Feel free to try them or to use them as inspiration. See here (scroll down a bit to "character set..."): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ This is what such text looks like: