Jump to content

geert_2

Ambassador
  • Posts

    2,084
  • Joined

  • Last visited

  • Days Won

    34

Posts posted by geert_2

  1. On 7/20/2021 at 9:28 AM, knutselsmurf said:


    just a warning, “dichloromethane” is carcinogenic !!

    I think Geert is working at a Chemical Laboratory like me, but to you all: don’t use it!

     

    Yes indeed, I have a lab, and a professional fume extraction cabinet for chemical and biochemical fumes. So I use it in this cabinet. In addition I use extra safety glasses (swimwear style googles that fully cover the eyes). And I don't touch the liquid, I use brushes or dipping instead, and chemically resistant gloves when there is risk of touching.

     

    However, I do think you can use it at home, but outdoors only, upwards of the wind, with good safety glasses of course, and gloves. No spectators, unless well educated on the subject and well protected. Definitely no babies, no toddlers, and no animals around.

     

    Indoor use can only if you have good fume-extraction. Some hobbyists do have an extraction cabinet, as they use it for soldering, welding, spray painting, sanding, etc... Or use a gas mask that covers the eyes too (like in horror films), with a *fresh and correct filter cartridge for this chemical*, analog to what you should do for spray painting. Make sure you don't breathe the fumes, and can't get spats in your eyes.

     

    I believe the health risks are similar to those of acetone, but less than chloroforme (at least, it requires no special permissions here, contrary to chloroforme). But don't believe me, everyone should study the risks and safety measures for himself.

     

    But that said, dichloromethane can really improve the quality of PET-prints (and PLA, and maybe ABS too), especially when making moulds for casting silicones. Then smoothing is really required, otherwise you can't get the casts out, as each layer-line acts as an undercut, firmly gripping onto the mould. Chemical smoothing is way easier and faster than endless sanding and polishing in areas that you can't reach. See the photos above.

     

    @ Sander: do you already have feedback from the engineering department on my questions above?

     

  2. If this is on the bottom of the print (=touching the glass), or a very thin print of only 1 or 2 layers, I would suspect grease or oil on the glass, so it does not stick and lifts off. But if this is on the top of a thicker print (hard to see on photo), it is something else: maybe too little infill, or a too thin top surface?

     

    For this sort of things, you could best watch closely while printing: often you can see what is happening, and why. For example, if there would be oil on the glass, you can see how the filament does not stick in certain areas and is dragged around. Or you can see material accumulating on the nozzle, and being deposited on the print, and all sorts of other irregularities.

     

  3. I do understand what you mean, but I don't know if there is a setting for that. What could also help is making the walls and top shell thicker.

     

    But if you are going to cast concrete anyway, what about using the model as a base for making a mould, and cast the whole model in concrete? So you have a fully concrete model without plastic?

     

    In a CAD-editor, subtract the model from a block, so you have the inverse hollow shape of the model. Cut that block into two parts (or more if required), so the cast can be released from it later on. Add alignment features so both halves snap in place. Then add clamping features, and add pouring and venting holes. Think these over carefully. Smooth the inside, spray mould-release spray, and you have a mould that you can use for multiple concrete casts. This combines the advantages of 3D-printing with casting, and you have the stone-like texture of cement in the casts. You could add marble or other powder to the cement mix for additional effects.

     

    On the internet you can find a lot of good tutorial videos on mould making and casting.

     

  4. If you have this only with a 0.25mm nozzle, but not with a 0.4mm nozzle, maybe you are printing too hot for it, and the heat still creeps up into the filament due to not enough filament flow? Or the filament begins to decompose?

     

    If you do a cold pull (atomic pull), the tip should look like the orange one at the very bottom here (minus the color-change, this was when changing colors, to clean the nozzle). If it looks like the white one, with a thick ring or blob where the teflon meets the metal part of the nozzle, then the teflon is definitely worn-out.

     

    Also, it could be that the bearings of the little fan are worn-out, so that it is still running, but way too slow to generate enough airflow? If it makes rough sounds like "rheu-rheu-rheu" then it is for sure worn-out. Or maybe there are hairs stuck in it, reducing speed and flow? Airflow goes with a power of RPM, not linearly, so a small reduction in RPM gives a huge reduction in airflow. When looking at your pictures, I still have the idea that the third fan's cooling is not sufficient, and heat is creeping up, melting the filament before it reaches the nozzle, as gr5 said.

     

    DSCN5237.thumb.JPG.8f29c03aad2ce9dd0d9490ed2ece9d98.JPG

     

    DSCN5238.thumb.JPG.c0b3c52c2588d4c4a95ec6ac24bd3e99.JPG

     

  5. Plastic folders crumbling apart after sitting in the daylight for a couple of years. And this was even behind sun-shielding green glass, which should catch most of the UV-light.

     

    These folders are usually made of PE or PP, but I have seen similar problems with other plastics. Last month I had a dustbin falling apart. A couple of years ago I have seen this in "weatherproof" garden chairs...

     

    That is why I would not trust plastics too much for load-bearing outdoor use, unless they are specifically tested and approved for it. 3D-printing materials are relatively new, and we don't have a lot of experience with long-term durability in this regard.

     

    plastic_folders_crumbling.thumb.jpg.599e36e56df41c193851806d30733021.jpg

     

  6. A 0.20mm line as designed in CAD might vary between 0.19mm and 0.21mm after exporting to STL, because the STL-file consists of straight line-segments, instead of smooth curves. The 0.19mm areas are likely to cause problems.

     

    Idem for lines of 0.40mm in CAD, for printing on a 0.40mm nozzle. I prefer to optimise the design in CAD, rather than rely on doing tricks in the slicer. So I design the lines a bit thicker in CAD: usually 0.5mm for my 0.4mm nozzles: this prints well, and aligns well on a 0.5mm grid while drawing.

     

    Maybe try 0.25mm lines for your 0.20mm nozzle? But I think a 0.2mm thick part is going to be way too fragile. If it was for myself, I would make it thicker. And I would then print it slow and in thin layers, to get good layer-bonding.

     

  7. If you have a single-nozzle printer,  and you can do CAD, then consider designing your own supports in CAD. Test various concepts and dimensions on a small testmodel, before doing large models.

     

    Standard supports are good for standard situations. But special cases might be better off with custom designs.

     

    A few examples:

     

    Pink and orange supports with custom brim to prevent them from getting knocked over. These are stable and solid supports, so I can grab them with a plier and wiggle them out. This model is way too small to get in there with a knife.

    ostrcp_key_v20_zoom.thumb.jpg.c85991865979ff09557a37d9ca6ad20f.jpg

     

    Concepts of supports, to minimise damage to the model. These require good testing: if the gap is too big, the model will fall off. If too small, the model will glue to the support, and separation will be difficult.

    anemometer1b.thumb.jpg.c006a1ccc8e1465c2c9a18fe6d91bf0f.jpg

     

    Free hanging supports, not going down all the way to the bottom, and thus not damaging the lower parts.

    supporttest12b.thumb.png.504e11c4d360abd18960d889c271f2a4.png

     

    Front view of these free hanging supports: they stay in place by the stringing of the material. This makes them very easy to remove (way easier than standard supports). The ribs on top are 0.5mm wide and high; the inverted stairs at the bottom are 1mm wide and high. This model could be printed without supports, but I want more accuracy, since another part has to slide accurately through the opening. The tabs on the side of the supports do almost no damage to the side walls of the model.

    supporttest12c.thumb.png.a4bbb089d43486b97df974e68e645196.png

     

     

    Printing tiny vertical models often causes this effect: the filament can not solidify due to the hot nozzle continuously sitting on top of it. So it deforms. Layer-adhesion and stress-concentrations could also be a factor. So you might want to design supports in such a way that they move the nozzle away for some time, and allow the model to cool (or print multiple models at once).

    DSCN5605b.thumb.jpg.2a696904daa58d988117c2f266bd4594.jpg

     

  8. In my experience the "insect antennas" are caused by the nozzle leaking while traveling through air. Upon reaching the next wall, that drop is deposited on the side of the wall. The next layer, the drop is deposited on the already existing drop, and so on, creating the "insect antenna" effect. Watch closely, then you see it happening. In PLA this is rare, but in more rubbery materials when molten, like PET, it is common.

     

    You can easily worsen the effect by switching off retraction, or by printing faster (=more pressure in the nozzle, thus more leaking). Reducing it can be done by printing slower, in thinner layers, cooler, and with all speeds the same (walls and infill). But I can not totally avoid it, unfortunately.

     

    The strings often come from material accumulating on the outside of the nozzle, and then slowly sagging onto the print. Sometimes this is deposited on the print and creates big blobs, often brownish color, and sometimes it just causes strings when moving from one to another part. Here too: it is rare in PLA, but common in PET, and printing slow, cool and in thin layers reduces it, but does not eliminate it.

     

    I haven't tried experimenting with retraction settings, so I don't know if that could help.

     

    So my solution is post-processing: cutting them off, and then light sanding and polishing, or chemical smoothing.

     

  9. This looks like a bit underextrusion. There is a video on this forum about possible causes and solutions for underextrusion, I think from user gr5. See if you can find that (I don't know its exact name, nor link to it).

     

    Further, if you could post more details such as material and settings, and a project-file of your settings, some people on this forum might be able to analyse it and give more advise.

     

  10. I have been printing with older spools of PLA without problems. But I do store them in a sealed box with dessiccant. It is true that old PLA gets harder, stiffer and more brittle, so it may be more difficult to find and to unwind from the spool.

     

    But don't throw the old spools away yet: first try drying them in an oven at 45°C for several hours, but well below its glass transition temp where it gets soft. Store in sealed boxes with dessicant.

     

    And use a bonding method for improved sticking to the glass: I use my "salt method": wiping the bed with a tissue moistened with salt water prior to starting a print, greatly improves bonding of PLA. Other people use the glue stick (a thin layer, and wipe with a wet tissue afterwards to spread it), dilluted white wood glue (ca 10% in water), hairspray (spray it outside the printer, never in), 3D-LAC, and similar. Try various methods and find one that suits you best.

     

    You can still find my old manual on the salt method here:

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

     

  11. I just took a look at the portable ladder in my lab: it is both riveted and welded. This makes it feel rock solid. If only riveted, each connection could still pivot a bit, causing a wobbly feeling. If only welded, the welds could be superficially glued instead of really melted together. I can't weld myself, but I heard from a professional welder that welding aluminum is difficult due to the very high temperatures required to melt the oxide layer. If not high enough, the connection may slightly stick, but is not welded and will separate under load.

     

    If yours will be mounted to a wall, pivoting will not be a problem. But if portable, make sure you provide multiple connections per step.

     

  12. Technically, it will probably work well if you design the parts thick and massive enough, if load is distributed well (=no stress points, correct orientation of layers), and if you print them correctly (temp, fill, flow), and regularly test for degradation and replace them.

     

    But personally, I would not really trust it. I wouldn't trust any plastic parts for such purposes. I have seen too many plastic things crumbling apart or breaking after a couple of years: dustbins, bottles, gardening tools, plastic toys, food boxes, garden tables and chairs, car bumpers, all sorts of composite casts,... And these were injection moulded parts, or cast parts, so they even didn't have the 3D-printing problems like layer-adhesion, stress-inducing entrapped air, uneven cooling stresses, etc.

     

    The biggest risk will probably UV-light degradation, plus to a lesser degree ozone, chlorine (if in contact with tap water or swimming pools), hydrolysis, drying-out (evaporation of plasticizers), fatigue,...

     

    I am not sure, but I do believe that here in Belgium weight-carrying plastic connections on ladders and stairs are even forbidden. At least, they are forbidden in our university: all connections have to be welded or rivetted.

     

    Without photos or drawings it is difficult to give advice, but what about using rivets? Or nuts and self-locking bolts? Maybe with use of copper-grease to prevent corrosion?

     

    • Thanks 1
  13. Also, check in the gcode if the "long straight lines" are long lines indeed, and not chopped-up in lots of tiny segments.

     

    And watch the printing through a magnifying lens, or through cheap glasses with high diopter (these work as magnifying glasses too) from the supermarket. Then you can see if the blobs gradually accumulate on the nozzle and are then deposited, or if they suddenly erupt.

     

    Maybe wet filament might also cause this, due to steam generation? Although I am not sure, usually it gives little craters and bubbles, like foam. So I don't think this is going to be the (main) cause.

     

    But it still looks overextruded to me, it looks like the nozzle has been wading through the melt.

     

    Some time ago I did a couple of over- and under-extrusion tests.

     

    This is 100% extrusion, seems quite okay:

    extrusion_100pct2.thumb.png.01adb69154c978415b3f3b42f0267907.png

     

    Again 100% extrusion:

    extrusion_100pct.thumb.png.ff43c0e342347a78e2902e72ce17359e.png

     

    Way overextruded: 150%:

    overextrusion_150pct3.thumb.png.424d051378a7a21af52029e29451cbc4.png

     

    Same overextrusion, seen from another viewpoint, 150%:

    overextrusion_150pct.thumb.png.c61778173be97ca083fbbac0ab3b5acf.png

     

    Way underextruded: 50%:

    underextrusion_50pct.thumb.png.a98c4c9891552d4b1aa2875f35776686.png

     

    This brim was printed at 100% extrusion, but it is slightly overextruded due to my nozzle being a bit too close to the bed (on purpose):

    brim_slightly_overextruded.thumb.png.3b035853bb889cdbe6528180bcce0b28.png

     

    It is hard to see for sure on your photo, but your image looks closer to my 150% overextruded tests here. So that would mean that: or you have overextrusion, or your nozzle is way too close (resulting in effective overextrusion too for the first layer), or the temp is too high (melt too liquid), or wrong settings for the filament, or worn-out nozzle, or something along that line, I think. If it was my printer, I would go searching in that direction a bit further.

     

  14. There might be lots of reasons for blobs: overextrusion (for whatever reason); the nozzle being too close to the bed (also resulting in a sort of overextrusion-effect); a too fine STL-file with way too much polygons which slows down the printer; material accumulating on the outside of the nozzle, and then being dropped onto the print; and probably other reasons that I don't think of at this moment.

     

    I would suggest that you watch very closely while printing, maybe use a magnifying glass: most of the time you will see what happens, or you can narrow down the causes.

     

  15. Warping-tendency on the bed *while printing* seems to be lower with thinner layers: their bottom is flatter and has a bigger contact-area to the bed than thicker layers, so a better adhesion and less warping. Also, a very thin layer can excert less forces while cooling than a thick layer. A 0.3mm layer suddenly cooling down and shrinking will be able to pull much harder than a 0.06mm layer, I think.

     

    But for warping afterwards, for example the completed part sitting in the sun or in weather, I have no idea. Prints with thicker layers have more entrapped air (=are less transparent) than prints with thin layers. But I don't know if and how that would affect internal stresses and warping afterwards? Interesting thought though.

  16. The leaks come from when the nozzle jumps to another layer, or another part on the same layer. Just that transition-spot tends to have tiny holes. Printing faster and in thicker layers, worsens the effect, as the beginning and ending of the "sausages" are rounded. Put two sausages next to each other on a table, and you will see the same effect on a big scale.

  17. Layer-thickness also plays a role: generally I found that on my UM2 a first layer of 0.1mm is too thin and gets uneven. A first layer of 0.3mm is too thick and has less contact area and is not so flat at the bottom. A first layer of 0.2mm works best: this gives a nice flat and glossy bottom. The top of that layer tends to overextrude a bit due to it being printed at lower speed and without fan, thus with a lower viscosity and higher flow (less resistance in the nozzle). And I have calibrated my nozzle a bit too close to the bed. But as said above, that first layer is followed by others, so these effects are smoothed out after a couple more layers.

     

    The bottom of the bottom-layer is more important than its top, in my opinion.

     

    This is what the bottom usually looks like (above = camera-focus on model, below = camera-focus on reflection in the bottom layer; I couldn't get both in-focus at the same time):

    underside_mirror.thumb.jpg.d9e8c12251778b0a33338a0eac202c6f.jpg

     

    Another one:

    DSCN4938.thumb.JPG.90124a14e04953b171581afa5e8f9e9a.JPG

     

    Small item printed very slow:

    DSCN6083.thumb.JPG.6fa2f0776aca10a340718c2065decdbf.JPG

     

  18. I made a filter some time ago, and indeed, when using standard settings, a lot of tiny water jets came out all over the surface.

     

    Solution: print very slow, in thin layers, and at the lowest edge of the temperature range (because otherwise, if sitting too long in the nozzle at high temp, the material may decompose or burn). But it will take a long time to print.

     

    Alternate solution: post-process the print (if PLA or PET) with dichloromethane, which also helps sealing tiny openings. Search for my post about "smoothing PLA and PET with dichloromethane" for photos and more info. Or seal by painting or varnishing.

     

  19. (I can read German, but not correctly write it, so I will write in English.)

     

    Yes, PLA deforms under continuous load, creep deformation: clamps lose their clamping force, straight items bend, etc. In fact, a lot of plastics do, but PLA is worse than most others.

     

    Indeed, PET or CPE might be a better choice: these are still easy to print, and since they are more flexible, you stay in their elastic range longer before they permanently deform. I use it for carabiner hooks, etc., that require some flexibility. Chemical resistance is reasonable.

     

    PE and PP are more chemically resistant, but they are much softer, and difficult to print. Might be good for pipes, probably not for clamps.

     

    If things need to be water-tight, such as nozzles for cooling-lubricating fluid, print them slow and in thin layers. Fast printing, or in thick layers, will cause lots of tiny leaks, as I experienced in printing a filter element.

     

  20. 1 hour ago, ahoeben said:

    ...

    If you want to print the bolt sideways, experiment with sinking it into the buildplate a bit. A bolt does not have to be 100% circular to work.

    ...

    Yes, this is a good point.

     

    For designing parts, a lot of the rules for injection moulding apply here too, although not all. You can find online manuals on how to design plastic parts for injection moulding, from big companies like BASF, Bayer, DuPont, etc... How to design plastic threads is also explained in some of them. But this is more suitable for bigger sizes anyway, like plastic bottle lids.

     

    The company protolabs also has good manuals on designing for 3D-printing, I think.

     

    With a standard nozzle of 0.4mm, you can not print lines and corners narrower than this 0.4mm. So an M3 bolt sideways won't work. I consider 0.5mm my minimum linewidth in design.

     

    An M3 (or any similar thread) upwards will also not work: such very steep overhangs of 60° tend to sag and to curl up at he same time, severely distorting the shape. Also, when printing too small items, the hot nozzle stays on top of that area, radiating heat, so the print can not cool and not solidify: it becomes a mess. If you need to print very small items, you need a dummy "cooling tower" next to it, to move the nozzle away for some time.

     

    See these effects, with and without dummy cooling tower (the square blocks):

    DSCN5603b.thumb.jpg.83c20560cfab90d56590243bc6015f12.jpg

     

    dummy_inverse_block6.thumb.jpg.2bdb2396588983363b48127ee12d8174.jpg

     

    dummy_cutout2.thumb.jpg.750722bab5fa1c22a5e38d2a5717ab5b.jpg

     

    Minimum practical dimensions: text caps height = 3.5mm, text leg width = 0.5mm. Note how much more crude the text is than the standard M4 screw thread:

    DSCN5645.thumb.JPG.8c0a292e655cd93fd53baded45b4a915.JPG

     

    I would recommend that you design several small (and not so small) test items, and experiment on them. While they are printing, stay with the printer and watch closely what happens. Change temperature and speed on the fly, and see how that affects print quality. In this way, you will quickly learn a lot.

     

     

    • Like 2
  21. When I am not printing, I store all my materials in a closed plastic box with a big bag of dessicant. Use one with blue/pink indicator, so you can see when it needs drying (=by putting it in an oven at low temp).

     

    Nylon is known to get too moist in only a couple of hours. It should be dried prior to printing, and then be printed from a dry-box. Idem for PVA. (I don't have a material station, so I don't know if that incorporates a dry-box functionality?)

     

    For PLA and PET: in real life I haven't seen too much degradation when sitting in the open, although theoretically they could suffer a bit from hydrolysis. So I store them in a sealed box with dessicant too, just to be safe.

     

    If a material is moist, just storing it in a box with dessicant is not enough: it needs warmth (but well below the glass transition temperature where it becomes soft) to shake water molecules loose.

     

    Dessicant: available in auto-shops, for drying car- and caravan-interiors in winter. When the blue dot changes to pink, it is too moist and needs drying.

    DSCN5613.thumb.JPG.d8dc14bb4c0ec6e523e94f2e569434f9.JPG

     

    • Like 2
  22. Most of the time small printed bolts and nuts won't fit anyway: they are way out of spec. As ahoeben said: buy standard screws (nylon, PC, inox,...), and cut them to length (if required). I do that all the time.

     

    In Europe there are lots of good online suppliers, like RS-components, Skiffy/Essentra, INDI, Farnell,... Or go to a brico-shop.

     

    The only reason to design a custom threaded part, is when there is no standard available. Let's say you invented a nice water gun for your kids in summer, and you want to fit a standard Coca-Cola PET bottle to it as water-supply. Then you could design a custom threaded cap in your gun, so it screws onto standard bottles.

     

    Standard nylon M4 thumbscrews and nuts in a custom design:

    DSCN5645.thumb.JPG.8c0a292e655cd93fd53baded45b4a915.JPG

  23. This is why you should not leave a PLA-print in your car in the sun.

     

    This was a PLA-mould for casting silicones. I had way overdone chemical smoothing, so in combination with the low-infill, the solvent seeped deeply into the plastic, and slightly warped it. So the mould wouldn't fit properly and wouldn't close anymore. I thought I would straighten-out the warping by clamping the mould, and then putting it in a computer-controlled incubator (=precise low-temp oven) at exactly 60.0°C for a few hours. Just above glass transition temp. Not the best choice, obviously.  🙂

     

    In a car in the sun, it can easily get hotter than 60°C, even up to 80°C. So, never leave PLA prints in the car, or elsewhere in the sun. Don't waste time trying to anneal PLA, because that by itself also deforms and warps the part. For car-parts, use PET or some other higher-temp material instead.

     

    deformed1.thumb.jpg.e1e5b207ad82b03d9b7e7bf1843697a2.jpg

     

  24. I do use surgical blades (scalpels) to remove blobs and strings, so I am a plastic surgeon.  :-)

     

    You can chemically smooth PLA and PET with dichloromethane. This works like acetone on ABS. First remove all big defects with a knife and with filing/sanding, and then chemically smooth. One light application is best, a quick brush-on. If you overdo it, it will penetrate too deep into the plastic, and make it weak, deform and crack after a longer time. It's a solvent, so it dissolves stuff and then dries out...

     

    It smooths into a high-gloss, which feels very smooth, but this also highlights any waves or bigger irregularities like ringing.

     

    Smoothing is highly required for 3D-printed moulds: you can remove the cast much easier than in an unsmoothed mould.

     

    Search for my post "Chemical smoothing PLA and PET with dichloromethane" (or something similar), with more explanation and photos.

     

    One part as-printed, the other smoothed:

    pet06.thumb.png.7940ebb2066a2abf22daab3684f7f35a.png

     

    matrx_adapters_13.thumb.png.e9bd9aa7ddeae3c5fca5fa2a18422740.png

     

     

×
×
  • Create New...