Jump to content
Ultimaker Community of 3D Printing Experts

geert_2

Ambassador
  • Content Count

    1,688
  • Joined

  • Last visited

  • Days Won

    21

Posts posted by geert_2

  1. In my experience, there are a lot of things that influence accuracy: temperature, speed, flowrate, material, even color (some colors melt better than others, not sure why, maybe due to fillers?), etc. Further there are things like layer-lines (curved edges), blobs, ringing, thickening of corners due to nozzle slowing down,... All these have an effect.

     

    The only way to know, is test. Make your typical features in real size in a small test model, and try that. For me, I usually take 0.1 to 0.5mm extra, depending on how tight a fit I want, and whether it are inner openings or outer sides.

     

    Be aware that PLA will deform even under moderate warmth. You can not leave it in your car, even not in mild sunny weather in spring or autumn (internal stress relaxation). And it will deform quite fast under continuous load (creep deformation). Don't ask how I know.  :-)  This might not be suitable for every technical application.

     

    If the application allows it, I usually go for plenty of tolerance. Otherwise I calculate some post-processing in from the beginning: some filing, sanding, cutting,...

     

  2. I think CPE is very similar to PET? Anyway, PET feels totally different from waxy and floppy materials like PP and PE. It can flex a little bit, but definitely not like PP or PE, only rather like ABS. If overloaded it will very suddenly break without warning, almost like plexiglass.

     

    I have no experience with printing PP and PE, but I guess also a 3D-printed PP-model would probably feel different from an injection moulded one, due to the layer lines and poorer layer bonding?

     

    If I had to do it, I would probably go for what prints the best, the most accurate. And then post-process and smooth that, so that the shape is as close as possible to what you want. Even if it was in PLA. And then let the client imagine how that would feel and work in PP or PE. If he is used to these materials, he will be able to guestimate that, I think. Or at least I woud do that as a first intermediate step, to further discuss things and refine the concept?

     

    DSCN6055.thumb.JPG.c9f2aa4f551f913408727e004905b944.JPG

    PLA carabiner hooks start to crack after some time of use (left, cream). PET carabiner hooks don't crack (right, green), but if overloaded break suddenly (fracture surface below). They don't ply like PP or PE.

     

    image.png.9be6fee3ee3b5b4656e38c84d2416d5e.png

     

  3. You can always design your own supports in CAD, thus as part of the drawing itself. Then you have total freedom of design. When printing in PLA, you could for example design all supports in PLA too, except for a small interface layer of PVA, so you can wash away that and get nice clean undersides in your model. Automatic tools are great for general use, but for very specific cases, it is best to do it yourself I think. The automatics can't know what you have in mind.

     

    See a few examples here:

     

    support_sandwich.thumb.jpg.e42b127a279f4a760c444239c73a086c.jpg

    This is how you could make custom supports in PLA, and have only a thin PVA-layer inbetween them and the real model, using a dove-tail for good mechanical bonding.

     

    ostrcp_key_v20_zoom.thumb.jpg.c85991865979ff09557a37d9ca6ad20f.jpg

    The pink and orange are custom supports. So I can wiggle them out by grabbing them with pliers, as the model is way too small to get in with a knife. Text caps height is 3.5mm. The added custom brim is to prevent them from falling over. This is for a single-nozzle printer.

     

  4. 23 minutes ago, JohnInOttawa said:

    WRT minerals, we get quite a range here in Canada, likely in other countries the same, calcium, but also conductive stuff like iron. (which of course conveniently adds rust to things as it dries)

     

    My concern is really over what gets wiped down and where water goes undesired and unseen.  Like George, I'm careful with water around soldering too and use alcohol around boards to clean when safe to do so (and that's not always).

     

    Not sure if this helps...

     

    J

    Ah, okay, yes I can see the iron/rust concern, if that stuff would get under or into tiny circuits, solderings, and connectors or so. Might also change impedances or resonances at very high frequencies or very low-current circuits. Fortunately iron is not an issue in our tap-water. (Though it could be for people who pump up their own ground-water, which is often brown because of high iron-content.)

     

  5. 2 hours ago, JohnInOttawa said:

    Just a suggestion on the water.  Make the investment in distilled or de-ionized water.  Depending where you live, there can be high enough mineral content (and other additives) in the tap water that you'll give yourself issues later.

     

    J

    Hoi John,

    Out of curiosity, did you have any specific mineral-related issues in mind? We here in Belgium (near Antwerp, it differs elsewhere) we have very hard water, thus very calcium-rich. It really feels "hard" to the touch, totally different from de-ionized water that feels much softer. If you let a glass full of water dry out, there is a thick layer of calcium-deposit. Also in boilers and kettles. But when cleaning with this water, then rinsing, and wiping things dry, or compressed-air blowing them dry, I have never seen any side-effects.

    But this could of course be different in other places with other chemicals.

     

  6. And check that the rods are totally straight, not bent. Even if they are only 0.1mm bent out of shape, thus almost invisible, that is a full layer-height during printing. Maybe put them on a totally flat table, roll them, and see if light shines through below them.

     

    I would also go for window-cleaner, or Mr. Proper or similar, spray bottles to get dirt off. Whatever works in a dirty kitchen, without destroying things, should work here too, I think. And then clean with pure warm tap water, to get the soapy-remains off.

    • Like 1
  7. Or leave a gap of 0.01mm between the plate, and the studs, in the same CAD-model. So the plate would print as one block including top layers. And then there would be a 0.01mm vertical gap, after which the studs are printed. Such a tiny gap will obviously be filled, and the studs will fuse to the baseplate as if there was no gap.

     

    I don't think this is the optimal way of doing it; too much CAD-work. Beter would be to find the exact right slicer-setting to handle this (if present). But it might be a way around if nothing else works...

     

    • Like 1
  8. Yes, I understood that. But your question has come up here regularly, and I have never seen a good solution yet, except for staying around the printer. That doesn't mean there isn't one, just that I haven't seen it.  :-)

     

    I am afraid that you have to go to post-processing methods: inserting the nuts or special threaded inserts after printing. Or maybe print it in two parts, insert the nut, and glue them together? Preferably with a sort of dovetail mechanism or similar, to get mechanical retention in addition to the glue? And in such a way that the part does not see too much load in that direction? For example a sort of drawer that you slide in from the side, and that contains the nut. And then the whole drawer is glued or acetoned in place, so it becomes one solid block?

     

    Or relocate the fasteners, or add a flange, so you don't have to insert them while printing, but you can do it afterwards? It all depends on the design and requirements of course.

     

    Like with all great inventions, maybe someone just needs to see the light? Often the inventions were simple, but no one saw them up till that point...

     

    Anyway, if you would find a solution, let us know. It could help a lot of people.

     

  9. 7 hours ago, stu_le_brew said:

    Some thing I learned of a long time ago was to pull up the white part of tube clip while pushing down the tube, I use a pair of tweezers to do this - it ensures the tube is securely fixed to the tube, then I put the blue clip under the white part

     

    I have not had the tube come out on my UM2 for years

     

    good luck

     

    Yes, I do it similarly: I push in the tube, then while I keep pushing on the tube, I lift the ring of the collet, and slide the horseshoe clip onto it. So it sits all the way down without any play. Never had tubes coming off.

     

    To make this easier, I made a custom horseshoe clip, so it doesn't keep falling when I try to handle it:

     

    horseshoeclip.jpg.f0b28c4ee645faf0935adc8b67b35887.jpg

     

    It is old but still on my page (and then scroll down):

    https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

     

     

    • Like 1
  10. If I understood it correctly, you mean this: buying a gcode-file from unknown source, with unknown settings, tested on unknown printers with unknown materials all differing from yours, and then hoping it is not going to ruin your printer, crash the nozzle into the glass or the walls, overheat, produce spaghetti and clogs,...? No, I am not going to risk that, no way.

     

    Probably not too many other people either.

     

    There are just way too many settings that have to be customised to produce the exact results you want, for your purpose, your printer, your filament: speed, layer-height, nozzle temp, bed temp, cooling fan, first layer-settings, and so on...

     

    This is the reason why a lot of schools and maker-labs do not accept "print-ready" gcode-files anymore. They learned that the hard way. Now they require STL or similar, and they require it to be sliced and printed by their own qualified tech people, or qualified students with plenty of experience (e.g. student jobs).

     

    This in addition to all the remarks above.

     

    At best, I would consider buying an STL-file, if proven solid without defects, or better the real design-files (STEP-files), so I can examine and adapt the design. But only if it was a really valuable design, such as let's say a beautiful combustion engine demo model, or something like that. But probably I would just design it from scratch myself.

     

  11. No experience with this. But a thought: you also need to consider the maximum temp that the printer can deliver (I think 260°C for the UM2?), and the temp required for your material. Some engineering materials need way hotter. I don't know yours.

     

    If you already have the hardened nozzle installed, remove the bowden tube, heat up the nozzle, and try to feed some filament directly into the nozzle from above. Just to see if you can melt and extrude it.

     

  12. Yes, probably the PLA got already soft while still in the feeder or bowden tube?

     

    If you want to reduce noise, you could try to put absorbing panels around the printer, but not touching each other, nor the printer. Lets say a panel 30cm away from each side, and with *big* openings in-between each panel and all printer parts. Plus one under the printer. So that air still has plenty of room to pass. This should greatly reduce noise-reflections, echo, and standing waves (and all the other audio-stuff I don't know about).

     

    On Youtube, search for "how to build your own noise dampening panels" or so. There are a lot of Youtubers who built them to improve sound-quality of their videos. Even a handfull of panels in a room have a great effect, no need to cover everything. But be sure to keep a good airflow around the printer.

     

    Not sure, but hopefully this strategy might work for you too?

     

  13. What about staying around and keeping and eye/ear on the printer? Or via a webcam? Me being a simple man, I myself would go for simple solutions...

     

    Alternatively, if the design allows it, and depending on the load-direction, you could go for other methods of inserting the nut, for example via side-openings, or by melting-in threads or nuts with a soldering iron.

     

    I have successfully used these methods (the cone-shaped opening is only for inwards-directed loads obviously):

     

    ostroncp_v20170104c.thumb.jpg.6dae46fd9b48b292823335415abefbf4.jpg

     

    image.thumb.png.78c8f2262d3b2d5beaf962fdad1883c7.png

     

    m4_nut_metrinch3c.thumb.jpg.283f9b43968ee6752683682ed29fde3f.jpg

     

    m4_nut_metrinch3b.thumb.jpg.cd76def6dd57935a6c78b91be5294a79.jpg

     

  14. I would also check the part and STL for errors, and always inspect the layer view. Maybe design a new "known-good" test part that only has this problem feature. Design from scratch a cube of 10mm x 10mm x 10mm, with such a hole, in a decent editor (definitely not SketchUp).

     

    For small holes, you need to calculate-in the narrowing caused by the molten material being pulled inwards. Try pulling a string of honey or yoghurt in a small circle, it's like pulling a rubber band into a circle. Drill-out the holes *by hand*, if necessary.

     

    Often I add a rounding at the bottom edges of prints, with radius 0.3 to 0.5mm. This generally takes care of elephant feet and such, without me having to mess around with any settings, and it allows for a good squeeze into the glass.

     

    See these pics with tiny holes (not post-processed, and ruler is in mm and cm):

     

    DSCN6083.thumb.JPG.6fa2f0776aca10a340718c2065decdbf.JPG

     

    DSCN6081.thumb.JPG.b4339d77becd99527d2c1f0e0fbff415.JPG

     

    DSCN5622.thumb.JPG.bcd33809236414534d665e6ac120651f.JPG

     

     

  15. (I can read German, but not write it, so...)

     

    I would say: go to a local distributor, who has a lot of different printer brands and models on display, each with demo-prints made by each printer. Go have a look and inspect all these prints, the printers, and have a chat with that guy. He is most likely to sell you a printer that suits your requirements and budget, and still gives good quality. Because he knows, if he messes up, he will lose you and all your friends forever as customer.

     

    For example, here in Belgium we have Trideus, but there will sure be others too.

     

    Also go for a printer that has an open filament system, so you can print materials from different brands. And preferably go for a reasonably open-source printer, or one where you can easily find spare parts, info, plans, manuals, slicers, etc.

     

    I spent quite a lot of time researching the market, and talking with people, and going to collegues from the department of product development, before deciding. It was well-spent time.

     

  16. I have no experience with soft filament myself, but from what I have read here on the forum, very soft materials tend to be very difficult to print. Pushing them through the tube and head is somewhat like pushing a rubber band through a keyhole. Doesn't go very well. Oiling the filament, and printing very slow, and a higher flow-rate, do help a bit.

     

    What about printing a mould in standard PLA, sanding and smoothing it, and then casting silicone in it? Then you get the real stuff: chemically inert, temperature resistant, and flexible silicone, but with your own custom design. And you can re-use the mould, if well designed. Combines the benefits of 3D-printing with casting real silicones.

     

    In the mould design, use good draft (slanted walls), and allow enough room to flex the silicone when removing it. It should not get stuck in narrow areas, as these are hardest to remove. And print slow, cool and in thin layers to minimise layer lines, because they will make it hard to remove the casting. Also sand off and smooth the inside of the mould: this really makes removal easier. In multipart moulds, which I would recommend for complexer designs, be very sure to seal off the seams with wax or plasticine. Otherwise, if you use a slow-curing silicone, it will leak away: it gets into the tiniest pores of only micrometers.

     

    • Like 1
  17. Some of these silk materials are a blend of PLA and other plastics, which causes the silk-effect. Although I don't know for this particular type.

     

    If there ar no nozzle clogs and no overheating, have you tried how hard it is to unwind the spool? PLA is already stiff by itself, and gets even stiffer after time, which may make it very hard to unwind. It acts like a spring, trying to wind up again. Especially near the end of the spool where the bending radius is very thight. It might be just on the edge of what your printer can feed. If it is a bowden-tube printer, a tight radius also gives a lot of friction in the tube and printer nozzle.

     

    In such cases I manually unwind a few meters of filament, straighten it by winding 15cm of it in the opposite direction around a skater wheel (7cm diameter), and then releasing it again. And then the next 15cm, etc..., until I get a few meters done. Then I loosely wind it up again on the spool. In this way there is no unwinding-resistance, and no friction in the bowden tube and nozzle, as the filament now has the same bending radius as the tube.

     

    This straightening goes very fast, and I do it while the printer is warming up, so I don't lose time. Just be very carefull that the spool does not start unwinding from itself like a spring coil, and gets tangled up.

     

    I am not sure if this is your problem, but it might be worth having a look into?

     

    DSCN5776b.thumb.jpg.04c8093e153fa768024c3c36391f8b23.jpg

     

  18. Exporting to STL converts circles - thus curved lines - into polygons with straight segments. So if your wall thickness is exactly 0.4mm in the design, upon converting to STL-segments it may vary between 0.39mm and 0.41mm for example, depending on the chosen accuracy. The 0.41mm walls print fine, but the 0.39mm wide walls are dropped, leaving holes. That is why I often use 0.5mm for thin walls and tiny text legs, which prints well. I never tried changing nozzle-width to 0.39 or 0.38mm, because I am prone to forget that later on, messing up later prints.  :-)

  19. Hello,

     

    I am considering using Breakaway to make casting moulds that can not easily be made from simple two-part moulds.

     

    Normally I would use the traditional method of two mould-halves that are clamped together using alignment-keys. But this may not work for complex shapes with lots of undercuts, or shapes where I don't want parting lines in the model. Then Breakaway might be an option? This would require printing a new mould for every cast, but for special items like machine parts it might be worth it.

     

    Question: can Breakaway be acetone-smoothed, similar to ABS or PLA (cloakfiend-method)? Or by using any other solvents? If yes, with which solvents, which procedures, and how effective is it (strong effect like acetone on ABS, or weak effect like on PLA)?

     

    Anyone any experience?

     

  20. I think if the model does not stick to the bed very well, then it is not going to stick to the PVA-interface as well.

     

    Raft isn't used very much anymore, and I have even never tried it. I would suggest you try brim, which generally works good and is easy to remove. Or if you have a really special model, you could design your own brim-features in CAD. Automated functions are good for general items. But for very specific and unusual things, you as designer are best suited to design custom brims.

     

    For example I used them here to prevent the little red and orange things from falling over. They are very small, only a few mm, and would have an even smaller bonding area without custom brim. I also used it for the dummy tower. For reference: text is 3.5mm caps height.

     

    ostrcp_key_v20_zoom.thumb.jpg.c85991865979ff09557a37d9ca6ad20f.jpg

     

  21. Printing slow, cool and in thin layers reduces the effect for PET (I have no experience with nylon), but does not eliminate it. One cause is the internal pressure of molten material in the nozzle, which then leaks a little bit while traveling. So, less speed = less pressure = less leaking. That droplet is deposited on the next wall the nozzle encounters after traveling through air. On the next pass, the droplet is deposited on the previous droplet, and so on, causing these nice "insect antennas". Watch closely and you see them growing. Another cause can be material accumulating on the outside of the nozzle, sagging, and then being deposited onto the print, but that gives bigger brown blobs usually. Sticky, rubbery materials like PET have this way more than yoghurt-like materials as PLA. And obviously, wet material that starts cooking off water in the nozzle, thus generating steam and increasing internal pressure and forcing material out, is not going to help. So keep nylon and ABS *very* dry, and even keep them in a drybox while printing. Nylon absorbs way too much moisture in only a few hours. There are threads on the forum of people who modified or built dryboxes to use them while printing.

     

  22. Can you make a few close-ups of the bottom of that part? At first glance it looks like the first layer did not stick to the glass, sort of lifted, and created gaps. That would be my first guess. But we need to see the bottom.

     

    If yes, then thoroughly clean your build-plate, first with isopropyl alcohol, then a couple of times with pure warm tap water only. No soaps, as they reduce bonding. Then apply a bonding method suitable for your model and material. If using the glue stick, do a thin layer and spread and wipe it smooth with a wet tissue afterwards.

     

    For the later layers, it looks like it almost doesn't bond to the previous layers too? Seems a bit hard to believe, but it looks a bit like if you were laying plastic PVC-pipes next to each other. So how is layer bonding? Too big changes in extrusion-flow, feeding rate, or nozzle-temp seem unlikely. Or could that still be the effect of the first layer shining through till several layers up?

     

  23. I am not familiar with Fusion 360, but in most other CAD-software there is a Structure-Tree, listing all components. Maybe you can find the superfluous surfaces or parts in that list, and delete or hide them? Then they shouldn't interfere with the model anymore. If you hide them (if possible), they are still available for editing later on if necessary. I sometimes use this in DesignSpark Mechanical, to keep the basic shapes before I merge them into complexer models.

     

    Note: according to that Australian guy Angus (he makes Youtube videos on 3D-printing, but I don't remember his channel name), Fusion 360 is going to severely limit its free functionality. For example STEP-export will no longer be possible. At least in Australia, so probably elsewhere too. So you need to export all your models to STEP (.STP) before October 1st, if you want to be able to open them in other software packages later on. You might want to check if this is correct indeed.

  24. Some materials are very sticky like hotglue. The PET that I have also has this tendency. Hot PLA is more like yoghurt and doesn't accumulate so much. No experience with TPU though.

     

    Immediately after a print completes, I wipe off the nozzle with a strong thick paper tissue, while it is still hot and the goo still soft. Use thick or folded paper tissue, to not burn your fingers.

     

    If the goo is too much burnt-in, I remove it with a long brass M3 bolt or threaded rod. This works as a gentle file. Of course, never use steel or sandpaper or so, as that would destroy the nozzle. Use brass that is softer than the nozzle but harder than the plastic debris.

     

    What also seemed to help me, was adding a droplet of silicone or PTFE oil to a tissue, and then wiping the nozzle with that, prior to starting a print. Seems to have some effect, but not totally prevent it.

     

    All this is for my UM2 printers, and it obviously falls into the category of non-official personal methods. I don't know if they are usable for other models, so all is at your own risk...

     

×
×
  • Create New...