Jump to content
Ultimaker Community of 3D Printing Experts


  • Content Count

  • Joined

  • Last visited

  • Days Won


Posts posted by geert_2

  1. It could be the glass: glass is never totally flat, it has bumps and pits. If you look through a car window, especially on older cars, and in curved areas of the window, you will see the lens-deformations caused by this unflatness.


    But it could also be the rods not being totally straight. If they would only be 0.025mm curved, almost nothing, that would show up as a 0.05mm difference in layer height, in a sort of sine-wave. Thus a full layer in fine settings. If both rods have it, you could get a pattern like yours. Maybe you can check if the repeating frequency is one full turn of the rods?


    Bent rods could be caused by lifting the printer by the rods, instead of by the frame. Or by doing hard atomic pulls (which is why I use a much smoother method without pulling hard). Or it could have been in the rods from the beginning, if their extrusion was not perfect, or if they have been handled roughly during transport (e.g. if it were 3m or 5m long rods loaded on trucks by a crane).


    • Like 1
  2. I just think about it: if you want to make moulds, it is a good idea to have a look at manuals of injection moulding too. These manuals explain very well how plastic behaves when casted. And which guidelines to follow in mould-design. Although intended for injection moulding, a lot of principles apply to any kind of moulding. Search for: injection moulding manual bayer basf. And maybe other companies: most manufacturers of plastic pellets do have such manuals, available for free, because they want customers to be satisfied with their plastics.


  3. I don't think you can easily cast in HDPE or PET, that normally requires injection moulding. Most casting products are two-component epoxies, polyurethanes, polymethylmetacrylates (=PMMA, as in plexiglass and dental models), silicones, and similar. So they are often quite thin liquids, and then chemically cure into hard plastic or silicone. PMMA should be food-safe when fully cured, otherwise they couldn't use it for dental applications in the mouth. I don't know about fully cured PU. Most platinum-cured silicones are also food-safe, like in commercial baking moulds.


    The basic procedure is:

    - Design the mould in CAD, and print at highest quality. Make sure you can still open the mould after a solid cast is sitting in it,  thus no undercuts! Usually this requires a two-part mould. Or more parts if it is a complex model. Also design alignment-features into both halves, so they mate well without shifting. Provide holes or flats for screws or clamps to keep both halves together.

    - Carefully post-process and polish or chemically smooth the mould until it is totally smooth. No layer-lines.

    - Upon closing the mould, seal all seams, otherwise the liquid will leak away.

    - Carefully and generously apply release-coating to the mould. Otherwise the cast will be glued permanently to the mould.

    - Mix the components very well, and then degas them to remove bubbles. Depends on the products and their viscosity.

    - Use low-exotherm products, so they don't melt the mould. Some PMMA can get extremely hot, and even melt metals like tin, or even catch fire or explode, if in too big quantities.

    - Cast your mixed liquids, and let cure.

    - Carefully demould, and then post-process: remove seam lines, etc.


    Search on Youtube for moulding and casting: there are a lot of good tutorial videos. And they also describe the products they use. You will learn a lot, and see a lot of different techniques.


    The main disadvantage of moulding and casting is that you have to work with chemicals: in uncured state they often smell badly, and may be irritating or poisonous like solvents. But once fully cured, they are almost inert plastics. And they can be messy and sticky when still uncured.


    Note that silicone is watertight, because it repells water. But it is not oil- and solvent-tight: these leak through slowly. So you can't keep oils in a silicone cup or pot. Don't ask how I know...  🙂


    • Like 3
  4. Maybe you could put them in a well controlled oven, at the glass-transition temperature, and put a heavy weight on them, so they get flat again? Let it sit for several hours, and then very gradually let it cool. Or heat-up the bed of your printer, and let it sit there for overnight, with a weight on top? And then very slowly cool down. They might get straight again?


    Your models are thin, so I don't think they would warp too much if you use the glue stick, 3Dlac, hairspray or something similar for bonding? Worth trying. If it fails, then at least you know. Otherwise, you have an additional method as tool.


  5. Also, shells are 100% filled, and the shell in the curved part is longer than in the straight, which could give a bit difference. But so much...?


    I think you would best print them at the same time, on the same bed, and 100% filled. So there are no differences in infill-pattern, and no differences in shell vs. infill, and no differences in speed, temp, extrusion-rate (or underextrusion), etc...


    Maybe scale it down to 50%, for a quick test, so you don't waste too much material?


  6. Yes, to me too your part looks big enough not to worry about layer-time. Except if you would have tiny features protruding from the top. I wrote that reply because you asked it. But it is more for very small items. Very often I have models with a tiny top surface of ca. 10mm x 10mm with very fine details (as in the blue thing above). And then it becomes a necessity. And even then I often give the dummy a very crude shape: just hollow below the top area, and filled as soon as the tiny features are reached. Just to get the nozzle away, let that cool, and keep the flow going steadily.


  7. 8 hours ago, stan_bulgaria said:

    Hello folks,


    Today I'll add the support as Geert adviced.


    Yes, but don't do the whole model yet. Do a small test part: let's say only one bulb, attached to a short stem. And that with different gap-variations. See which gap works best. Then you will lose the least amount of time and material.


  8. User MakersMuse on Youtube has designed a testpart with various tolerances, specially for this purpose. After printing, you can see which parts still move, and which not. Then you know what tolerances you need to apply in your own designs.


    Also, printing circumstances have a huge influence: printing hotter, will melt them more together. Thick layers will reduce accuracy, as well as printing fast. So, printing slow, cool and in thin layers should give best results. Theoretically...


  9. Dichloromethane works for PET as well, and it is even more agressive: in one generous brush-on application, it removed all layer lines in my tests. See the photos.


    The outer layer stays soft for quite a while. Sometimes it gets a bit milky, but that goes away with time, and/or can be wiped off the day after.


    Pro: much smoother surface, no layer lines anymore. And internal features such as watermarks in transparent/translucent PET become better visible.


    Contra: the tiny bubbles... And the fact that I have no idea about long-term effects: deformations, changes in material characteristics,...?




    All photos: material is PET (brand ICE, from Trideus in Belgium), and layer thickness is 0.1mm if I remember well. Watermark text is 3.5mm caps height, 0.5mm leg-width, and sitting about 0.5mm inside of the model, thus hollow letters.


    Top one = treated with one generous brush-on application. Bottom one = untreated.



    Left area = untreated, right area = treated



    The bubbles... They start forming some time after the application. Looks like degassing in the still very soft, half-molten plastic. The CH2Cl2 is highly volatile, like ether.



    More bubbles in the treated one... The text is a hollow watermark, sitting about 0.5mm in the model.



    Smoothed part (top) looks better and is better readable from most angles.



    Left = treated, right = untreated



    Another view of the watermarked models.



    Another angle. Even though there are bubbles, the treated one (top) looks way better than the untreated (bottom). One application removed all layer lines. What is still visible, are the tiny gaps deep inside the material, in-between the printed sausages.




    • Like 1
  10. 41 minutes ago, stan_bulgaria said:

    Thank you Geert,


    I didn't know of that trick with the gap. I made my own support that I removed but will put it back again, only with a gap of may be 0,5 mm.





    Not sure, but I think a gap of 0.5mm might be a little bit too big? Maybe make a small test piece with gaps of 0.1, 0.2, 0.3, 0.4, and 0.5mm. Print that and compare what works best for you? It will also depend on the material, cooling, printing speed, printing temp: obviously, hotter will melt things more together. It will also depend on the width of your support: wider will be more difficult to remove. I usually take "one nozzle-width" for such supports (thus 0.4 to 0.5mm wide for a 0.4mm nozzle): that is still easy to cut off. And gaps between 0.2 and 0.4mm, depending on the model. My fixed rule is that there is no fixed rule.  :-)


  11. If I had to print this model, I would consider one of two options:


    1. Split the model in half, top and bottom, print both halves separately, and glue or bolt them together. In case of bolting, provide holes for inserting bolts and nuts. I often use standard nylon M4 for most of my designs that need to be assembled and disassembled.


    2. Design a custom support and brim in CAD, so it fully supports your model where you want. If well done, it should print better, give less deformations. But it will cost time and may take a bit of trial and error. If you leave a tiny gap between your custom support and the model, it will still bond enough, but be easy to remove. Also, you could redesign the bulbs and stems to print better, with less steep overhangs. See these quick concept-drawings of a bulb with custom support, oval bulb with less steep overhangs, and a stem with less steep overhangs and support. The gap between support and model will be trial and error: try one bulb and stem until good, before doing a whole print.





  12. In my experience not. (But other people may see things differently of course.)


    Because then the nozzle is sitting there and just waiting. The flow is stopped, the filament spends time in the nozzle, heating up more, thus getting hotter than before and more liquid. It may ooze, and it will surely have different flow characteristics than before, which will show up as horizontal lines. Also, the heat accumulated in the filament, still has to "evaporate" after printing. Thus if it gets hotter due to waiting, then that extra heat also needs to get out of the print. The only advantage is that the nozzle does not radiate its heat directly onto the model.


    For small models, best is to keep the layer-printing time and flow-rate as constant as possible. So that the printer spends an equal amount of time on each layer. This gives each layer the exact same cooling time. If there are huge differences in time per layer, e.g. when a model suddenly goes from a very huge area to a tiny area, that often creates faint horizontal lines in the print: not much, but you see it.


    What I often do for tiny models, is create a dummy tower that has the inverse shape of my real model. Or better: the complementary shape. And then of course, the tinier the model, the slower and cooler you need to print, and in thinner layers, to get maximum details, and the least amount of heat in the model.


    See these pics for "inverse dummies": the first is the concept, the second is part of a real model, otherwise the top does not solidify well and deforms.








  13. I would rather go for multiple smaller batches, than one big batch. Even though that costs extra warming-up and cooling-down time. But if something goes wrong and you have to abort the print, then you only lose a couple of hours and one model, not days of work and lots of people's work. This works best for my models which usually take 2-3 hours to print. But this is my personal view, and not necessarily suitable for everyone. And I only do all printing myself, no others touch it.


    MakersMuse on Youtube has been doing 3D-printer repair service in the past. His experience was that students tend to mess up printers, because they are too unfamiliar with them: using wrong software, wrong settings (e.g. outside of the print volume), crashing the print head, burning the filament, damaging the bed, etc... He has made a whole video about this, I think. Have a look at that first.


    He recommends that only one or two educated persons handle the printer, and do all the slicing and printing. All the rest should go via these persons, and only watch, not touch. That should give the smoothest workflow, and the least amount of down-time. This is also what schools are moving to now, they are moving away from everyone doing everything.


    However, if you prefer to let them do the printing themself, then absolutely require that they read the manuals, and familiarise themself with the printer first. Let them sign that they have done this, and do a quick exam (can be verbally at the start of their prints). And then let everyone print his own design, and let them take full responsibility for it: in this way they will by far learn the most of course. But you are likely to have the most troubles with the printer...


    Could be difficult to find a good balance.


  14. TPU 95A has a shore hardness of 95, I think. So that is quite stiff, much like hot glue after cooling down. If you print it in a thin layer of 0.1-0.2mm it will be flexible, but not in thick parts.


    You could try playing with less infill, but I have no experience with this, so no recommendations.


    If you need really soft materials, you could consider printing a mould, and then casting silicone in it. This exists in various hardnesses. There have been several posts about casting silicones, maybe you can search them?


    • Thanks 1
  15. What about printing a mould, and then carefully sanding and polishing that mould, and casting your model in a food safe plastic? Then you still have the freedom of designing and 3D-printing, but you don't have the tiny holes in the final models. Use a low-exotherm, slow curing product, so the prints do not melt. And depending on the product, use a good mould release spray or coating so it does not permanently get glued into the mould.


  16. The formulas have been explained somewhere on this forum, but I don't know which keywords to search for (I don't remember in which context or topic it was). Maybe you can find it back? At the nozzle-end it started with line-width x line-height x speed, if I remember well, but the rest for the feeder etc., I forgot.


  17. It looks like you have severe underextrusion, something stopping or blocking the flow.


    There could be a zillion of reasons, blocked nozzle, nozzle too cold, filament too thick or getting stuck, kinks in the filament,...


    On this forum, try searching for causes of underextrusion. User gr5 has made a good and extensive diagnostics list that he posted several times. That should help you going.


    • Like 1
  18. 17 hours ago, UbuntuBirdy said:

    when you vaporize it, there is only a small gap between a smoothed part and a totally melt down...


    Do you have any timing recommendations, to start from? Is it in the range of seconds, minutes, quarter hours, hours,...? My prints are usually 100% filled, and not too big (see dimensions below).


    Edit: and any idea about long-term side-effects? Strength, brittleness, dimensional stability,...?


    I have an old magnetic stirrer: if that still works, it could help distribute vapours evenly, so they do not accumulate on top or on the bottom of the jar, if their weight would be different from air.





  19. I don't know your materials, but in my experience PET tends to ooze out and string more than PLA: molten PET is more rubbery than PLA, which is more like yoghurt.


    I see different types of blobs in my PET-prints:

    - Those that ooze out while traveling through air, and are then deposited onto the next wall, often forming "insect antennas". Printing slower helps, because then less pressure builds up in the nozzle, and thus less molten material leaks out.

    - Material that accumulates on the outside of the nozzle, and then sags and gets dropped on the print, often in big lightbrown blobs. Oiling the hot nozzle with PTFE oil reduced that effect, then it becomes more like a teflon cooking pot. This is not an official method, just my own.

    - When printing very small items, the hot nozzle stays on top of a tiny area, so the print can not cool down and solidify: thus it stays molten and deforms into blobs. Printing cool and in thin layers helps, and multiple parts (but then you get the oozing while traveling).

    - Without cooling, overhangs are difficult: the molten material does not pull into a string, but snaps and folds back onto the nozzle, like a rubber band snapping. That causes blobs too, looking a bit like tiny grapes. More cooling reduces this, but gives worse layer bonding.


    In general, for PET: printing very cool, at the lowest edge of the temp range for that material, and in thin layers, and slow, gives the best results for me.


    Also see my photos of PET test blocks in my reply of yesterday here:


    This is the effect of printing small items: the nozzle stays on top of it, too hot, so it can not solidify. Printing a dummy tower next to it, or multiple parts at once, reduces the effect.




  20. Still a couple more pics. All photos were taken with a Logitech C525 webcam, with a close-up lens in front of it.


    For some applications, smoothing may be nice, but for others (such as fine text) the untreated prints may look better and crisper.


    I don't know how the glossy surface of the smoothed areas would affect bonding of paint or plating later on.


    Untreated text and smoothed text next to each other. The smoothed looks terribly out of focus due to the rounded edges.






    Smoothed, but a bit overdone: notice the tear (dull area) near the bottom



    Top part smoothed, bottom part not



    Here the lower area is smoothed, and both are bonded by the liquid running into the gap.



    Smoothed and bonded. Notice the bubbles oozing out here too. These plates are ca. 3mm thick.



    And now that is it, as I ran out of photos.


  21. A couple more pics:





    Smoothed, but overdone: see the dull area near the top



    When overdoing, the holes shine through the bottom, indicating that dichloromethane penetrates quite deep into the plastic and keeps melting/dissolving it. These indents take some time to form.



    Untreated text: this text is ca. 7mm high



    Smoothed: it looks out of focus, but it isn't (the dust is sharp)



    Untreated, close-up



    Smoothed, close-up: notice the bubbles oozing out of the edge of the hole, where the dichloromethane pooled and overdid



    Top 1/3rd untreated, bottom 2/3rd smoothed: although much smoother, it shows deformations better due to the high-gloss.






    Bottom 2/3rd smoothed: notice how the liquid filled the gaps, and partially melted the material and bonded the parts together. I haven't tested how strong the bonding is.



    So, that's it for now.





    • Like 1
  22. Here are a few photos of before and after smoothing with dichloromethane.


    Filament under test is only colorFabb PLA/PHA for now, because I have lots of old waste parts available for testing. Nozzle-size is always 0.4mm. Layer-heigts may vary between 0.1mm and 0.3mm (I don't remember for most parts, too old). Most parts are quite small, often ca. 10mm wide.


    All tests were done by brushing-on. I haven't tried vapour-smoothing yet. Smoothing still continues a bit after the dichloromethane has evaporated, while the PLA is drying. So you need to stop a bit early. Similar to aceton smoothing on ABS.


    It looks like multiple gently brush-ons, not too wet, is better than overdoing one big brush-on: too much liquids causes tears, and partial dullness, and more bubbles oozing out.


    Generally, in the photos:

    - untreated, original parts clearly show layer-lines, you are familiar with that

    - mildly smoothed parts still show layerlines, but are far more glossy, and the "valleys" of the layer-ridges are smoothed

    - heavily smoothed parts show very few layer lines, but get dull again if smoothing is overdone. Not only the valleys of the layer-lines are smoothed out, but also most of the tops.

    - if smoothing is overdone, bubbles ooze out. Not sure why: the dichloromethane penetrating too deep, melting/liquifying material, and forming bubbles when evaporating on its way out? Maybe?

    - dichloromethane penetrates quite deep: it shines through on thin plates

    - if brushing-on: use a smooth brush, because a hard brush shows scratches

    - the high-gloss of smoothed surfaces may show defects more than non-smoothed duller surfaces

    - brushed-on liquid penetrates into seams and closes them of, bonding parts together, although I have not tested how strong the bond is

    - also, brushed-on liquid closes gaps due to underextrusion. But don't overdo it, because then you get the effect of bubbles oozing out


    Now the pics:


    Original, untreated



    Smoothed with dichloromethane. Notice the dark vertical lines, which are brush streaks, due to the brush that had hardened.



    White bar: left=untreated, center=mildly smoothed, right=heavily smoothed






    Heavily smoothed










    • Like 1
  • Create New...