Jump to content
Ultimaker Community of 3D Printing Experts


  • Content Count

  • Joined

  • Last visited

  • Days Won


Posts posted by geert_2

  1. I have manual leveling on my UM2 printers. I did that according to the official procedure in the beginning. Since then I only adjust it occasionally on the fly. I start a print with a thick skirt (5 lines or so), or when printing the brim if I use brim (rarely), and then I watch closely how the first layer is, and adjust on the fly by moving the screw a 1/8 of a turn, and then watch again, etc. I think it is a little bit closer now than originally, indeed. But I only adjust that maybe once a year... My bottom layer is 0.2mm thick: I found that 0.1mm is too thin and a bit uneven, and 0.3 is too thick and gives lesser bonding than 0.2, especially for objects with very small holes like in my photo above. Everyone has his own taste: I like the simplicity of manual leveling, like manual shifting in a car, and a manual handbrake. :-)


  2. Wow, these are really impressive paintings. They look like the antique art you would find in museums.


    When priming your paint spray cans, maybe you could do that on a plain sheet of white paper, or sheet of wood, and keep this sheet with all color patches, and make photos of it? This would give an impression of each of the colors and effects on its own.


    Question: before painting, do you chemically or mechanically prepare the surface of 3D-prints, for a better bonding? And do you use primers that chemically dissolve and penetrate the plastic, or do they just coat it?


  3. I guess there are defects in the model. SketchUp is going to cause you endless headaches and problems: it does not close its vectors, so the models are not watertight solids, but sort of "glued together paper models". And it often has duplicate walls and other defects. It was only made for visual representations, not for 3D-printing.


    It will probably take you far less time to find and learn another CAD-editor, than continue in SketchUp.


    I am using DesignSpark Mechanical from RS-components, free but requires registration. However, this is only for geometric technical shapes, not for smooth organic shapes. But there are other free CAD editors. Take a look at demos and tutorials on Youtube, and find one that appeals to you.


  4. Are you sure this is the only thing that is broken? And there is not some other fault (maybe electrical or electronic)?


    I would say: if you do have standard PLA, try that and do a *cold* wash (=max 40°C) to see if it works now. Only then go for a more suitable material (but I can't give recommendations, lack of experience myself of heat-resistant materials).


    Do the parts have to be so thin, to fit well, or is there room for thicker parts?


    If I had to do this, for a first prototype to test the fit and function, I would cut it into pieces, so that I could print each one flat on its back. And then assemble and glue it like Faller HO model railroad houses. Probably...


    And what about glueing the original, and maybe reinforcing weak areas with some glued-on metal plates, or metal-containing self-curing putty? Or glass-fiber tissue and epoxy, like in car body repairs?


  5. Yes, irregular temperature could be a factor: I once measured it, and at the edges it could be quite a bit lower. Maybe due to the rising hot air pulling cold air in from below. Also, under the nozzle-fans the bed can be much cooler, especially on small objects where that fan is continuously blowing on a small spot.


    Also check the bottom surface of your prints: is it squeezed equally flat in all areas? Maybe if one corner is too far off, the filament is not squeezed well, and that could cause it to bond less. See the pic how my prints typically look.


    Grease or oil also comes to mind, but if you already cleaned that area multiple times, it shouldn't be...


    Photo of typical bottoms of my prints (this is PET):




    • Like 1
  6. My bad, I had missed that printer-info.


    Yes, trying a higher bed temp might also be a good idea: mine is around 80-90°C for PET (I don't know by head, but I definitely increased it from the original 70°C). The bed temp has te be close to the glass transition temperature, where the material starts to become soft. If lower, bonding reduces. If higher, the part may become too soft and sag, warp or peel off.


    What you could also try if you suspect time reduces bonding: *immediately* after a small test print completes, manually heat the bed to the same temperature as during the print, e.g. 80°C. And then let it sit for the night at that temp. In the morning, see if corners have lifted, and try pulling it off. But of course without brute force, always gently. If you print a tiny cube of 10mm, if the bonding is good, you can't pull it off. If the bonding goes bad, you can. But again, always be gentle.


    Forgot to mention: also consider using a brim: maybe in Cura, or as a custom designed brim (=designed in CAD, as part of the model). This also helps reduce corner-lift.


    I guess you are going to print the pale lid on its back, thus with the huge tabs facing the glass? In that case, consider free hanging supports for the tiny overhangs for the screws. But again: first test this concept on a small test model.



    Custom brim around the left model (the right one is post-processed and polished):



    Custom brims designed in CAD around the pink and orange supports, and around the cube:



    Free hanging supports:



    Other view of the free hanging supports: they are kept in place by the strings from the material, and are thus very easy to remove:



    Cross section. The ribs on top of the supports are 0.5mm x 0.5mm. The inverted staircase at the bottom of the supports is 1mm x 1mm. The gaps and little "connection tabs" are designed and tested for this particular model, you may need to change dimensions for yours. Hence the need for a tiny test model, containing only this concept.





    • Like 1
  7. My brand of PET does not like traveling through the air, not for bridging, and not for crossing gaps to the next part. There are two problems:


    1) It does not pull a nice long string for bridging, contrary to PLA: PET tends to snap and scroll back onto its own, like a rubber band that snaps. This causes a blob hanging on the nozzle, instead of a bridge.


    2) When liquid, PET is more rubbery than PLA. So, when crossing gaps, the pressure does not release as fast in the nozzle, causing the nozzle to leak a bit more while traveling over gaps.


    In both cases, the blob hanging on the nozzle is deposited onto the next wall that it encounters, creating a sort of "insect antennas". See the pic below. But I could imagine that if that blob encounters a thin perpendicular strand, as in your infills, it would hit that, and maybe snap it? I am just wildly guessing, because I have never seen that myself...


    Using a magnifying glass, try having a very close look. Maybe you can see what happens? (And if you find out, let us know.)




    Close up of "insect antennas":



    Next to a ruler (mm and cm):



    • Like 1
  8. Without knowing your model, your printer and your particular CPE, I can only guess...


    The best thing you could do, is cut out the most critical and most representative parts of your model, such as overhangs or bridges. And combine these into a new *small* test model which takes only maybe 30 minutes to complete and doesn't use much filament. Test and optimise on this until you get it good. Overdo in one direction, then overdo in the other, and gradually find the optimal center. This is going to cost time, but you learn quickly. Of course there can still be scaling problems later on, such as bonding that might reduce with time. But then at least you get the overhangs and bridges good.


    • Thanks 1
  9. Did you switch off the cooling fan? Or set that at a very low speed?


    For PET, which I think is very similar to CPE, I always print without fan if the model allows it, thus if it has no overhangs and no bridges. Otherwise I use the absolute minimum amount of fan.


    I usually print PET on bare glass. (Although some say this is not recommended, as it might cause a too good bonding, and might chip the glass.)


  10. On 2/20/2021 at 11:47 PM, Pizzastreichler said:

    Suggestion 1:  Besides above named things, you might want to cut and slice the body lateraly, so the angles are steeper and you might not need so much supports. Probably printing the front and rear as a short segment would make things even easier and some well attached supports could be left in place.


    Suggestion 2: Did you also concider the rigidity of the material and the final print? I would have some doubts about using this on a RC car. I'd rather print a quite solid positive model and vacuum form the body with a more stressable material.

    that way it's also not such a big deal when you flip the thing over  😉


    Yes, vacuum thermoforming might be a good idea. I think most commercial RC car kits are also vacuum formed?


    But I am not sure if you can vacuum thermoform around a PLA model, without it being deformed due to heat? Also, it might not pull vacuum very well, since the air can not escape from deeper areas. However, you can definitely vacuum thermoform around plaster. That is how some dental applyance are (were) made. Plaster is porous and lets enough air pass through. So you could 3D-print a mould in PLA in multiple parts (to be able to release the cast), and then pour plaster in it. Then remove the plaster cast, and carefully sand it and correct minor defects like bubbles. And use that plaster cast as mould to vacuum thermoform your car, using a somewhat flexible plastic sheet. In this way, you can always re-use the mould to quickly produce new body shells, for painting in different color schemes, or when one gets damaged in a crash.


    You can buy a commercial vacuum thermoforming kit, or make one yourself. Take a hard wooden plate, MDF or multiplex or so, for a base plate. Drill a hole in it and attach a powerfull vacuum cleaner to it. Then make a double frame out of wood, of the same size as the base plate. You need to clamp both frames on top of each other, so you can clamp a sheet of plastic between those frames. Once clamped in the frame, heat that plastic sheet with a heat gun untill it really sags. If you poke into the sagged plastic with a tool, the dent should almost immediately straighten out again. (If it doesn't, it is not warm enough yet.) While heating, switch on the vacuum cleaner. And then when hot enough, pull that frame with sagging sheet over the plaster model, so it gets pulled vacuum around the plaster. Let cool, and remove. I had a small professional unit for dental models some years ago, but the principle is the same, and I have seen home-built models. Maybe you can find Youtube videos?


  11. With DesignSpark Mechanical I also never had any problems: always correct solid and water-tight models, always nice on the build-plate, always correct size. But it is not suitable for artwork, only for technical models based on geometric shapes.


    I had occasional problems, but that was when I had made modeling errors, thus operator-errors, not software-errors.


  12. It is not clear from your photos, but it looks like the part came off the glass bed, and started sliding around? If yes, the problem could be your bonding method (or the lack of, or you forgot to use one), or a bed height leveling problem?


    If the model would still be stuck to the glass at this point, then maybe you have loose pulleys somewhere, maybe on one of the stepper motors? Or maybe something hit the printhead while printing (cat, dog, kid,...?), causing it to skip steps? Or printing too fast for your stepper motors, so it could not follow and skipped steps?


    I guess it must be something along these lines...


  13. The two-shell concept of gr5 might be worth considering, at least for those areas that need most detail. In injection moulding this is also used: the main mould gets a cruder finish, while areas that need lots of detail, like fine text or optical lenses, are done with a smaller insert with high quality finish. It do not need to be complete sleeves covering the whole model, maybe a smaller area is enough, depending on the model.


    I have made several PLA moulds for casting silicones. The most important aspect is to remove layer lines as much as possible. They act like a zillion of tiny undercuts, and make it very hard to remove the cast. Any layer lines are also terribly visible in the cast, and they make the cast sensitive for dirt.


    (You are probably well familiar with the other normal moulding aspects like: no undercuts, venting openings, alignment features, clamping features, etc...)


    Chemical smoothing might also be an option: I sometimes use dichloromethane. See my recent post here on PLA and PET smoothing with dichloromethane. This greatly reduces layer lines.


    One of these models is chemically smoothed, reducing the layer lines; the other is as printed.





  14. What is good enough depends on the model and application, of course. But 0.4mm from a 0.4mm nozzle as in the pics below, is moving away from the optimum. I think coloring it orange is good, meaning: "possible but not optimal". But maybe make that color a bit more gold-orange instead of deep orange (e.g.: RGB=255,180,0), so it visually moves further away from the red?


    Both left blocks in the overview-picture are 0.4mm layer height, from a 0.4mm nozzle. Top row printed at 50mm/s, bottom row at 10mm/s. It prints well, but you have very crude layer-lines, and you start to get deformation. And cooling is difficult and slow, as the heat from such thick sausages takes time to radiate out, so it tends to sag while cooling. It has a nice sparkle in it though for transparent materials.


    From left to right: layer height = 0.4mm, 0.3mm, 0.2mm, 0.1mm, 0.06mm. Top row printed at 50mm/s, bottom row at 10mm/s. This is clear PET. The brown discoloration is because it is staying long in the hot nozzle at low speeds and thin layers.







  15. In addition to the above:

    - Glue suitable for the filament you use, e.g. cyanoacrylate for PLA.

    - A drill for *manually* cleaning small holes (see picture).

    - A very thin needle (0.39mm) in case the 0.4mm nozzle would clog. Be sure to round-off the needle tip, so it does not damage the fragile and soft brass nozzle.

    - And indeed, all tools you would use for model airplanes, cars, trains,...

    - Tissues for cleaning the nozzle immediately after printing, to keep it clean.

    - A sealed box, with a bag of desiccant to store filament dry.


    And then try a lot of simple but usefull models at various settings. Keep watching while printing, so you see and learn what happens.










  16. For cleaning, don't use soap: that reduces bonding. That could well be the reason for your warping. Or do it the other way round: first soap water (like window cleaner or Mr. Proper), then isopropyl alcohol, and then with pure handwarm tap water only (but no soap). After using alcohol, don't touch the glass anymore on the areas where you are going to print. Finger-grease also reduces bonding.


    Also very moist and rainy weather might reduce bonding, when printing on bare glass. That is why I started searching for a new method, the salt method.


  17. On 2/8/2021 at 7:59 PM, Mari said:

     Keep them dumb, then you have to do the work... 🤪 very inefficient...They may be smarter then you think...


    Indeed. I am a big fan of educating people (although in my case the "clients" are PhD-students and collegues), so that they *understand* what they are doing. Not just verbatim duplication like a parrot, but real understanding, so that they "see it". That sometimes takes hours. And then it saves me days and weeks of time, because next time they can solve it themself, and they can often do it faster than I could (which is very good). Additional benefits are that they become way more happy. And you become more influential, because with each problem-solution you teach them, you also get a better understanding and wider viewpoint yourself, and your viewpoints will be more appreciated by others, because what you told earlier, did work. And if they later find new solutions based on your help, they will often come back and tell you, so you learn from that too.


    So, practically, if the client is not too far away, just go there for a couple of hours, or even an afternoon. Have a nice chat, have a look at their systems and their workflow, and find out and explain them how to to it correctly. Most people are smart people, who just lack some technical knowledge. When someone is lacking the grammar and vocabulary of a language, he can not speak it, but he is not dumb: he can learn it.


    • Like 1
  18. I may have missed it in the posts above, but I didn't see what bonding method you used? Could you describe that, thus method-name and exact procedure? That could be a reason too for the warping? Also describe the cleaning method you use for the glass? And the weather conditions in your environment now, temperature and moisture-level? These all play a role too in bonding, in addition to bed-distance and underextrusion.


    Since it isn't solved, it is best to go over each step again.


    For underextrusion: on my older UM2 (non-plus), there are a couple of things that are likely to cause underextrusion:

    - When the teflon coupler is worn-out: if you do an atomic pull, and there is a thick sort of ring or blob at the edge between coupler and nozzle, then it is worn-out and needs replacement. See the pics below.

    - When the filament spool is nearly empty, especially when printing hard filament like PLA: the hard filament acts like a strong spring, that resists unwinding. If you have a very hard time unwinding it by hand, then the feeder has so too.

    - When printing in the back-left corner: then the filament has to make a very tight curve in the bowden tube, and this causes a lot of friction in the nozzle and bowden tube.

    - When the nozzle is partially clogged by accumulated dirt and coal on the inside of the opening. Regularly do a cold pull to keep it clean.

    - And obviously, when all the above are combined.


    There could be lots of other reasons too, such as incorrect filament diameter, or dirty feeder wheels or damaged feeder, but the above I found the most common for my system. User gr5 has a good list of causes of underextrusion, so try to find that.


    As bonding method, for PLA (not for other materials) I only use my "salt method": after thoroughly cleaning the glass, I put a few drops of salt water on it, and wipe that with a paper tissue until dry, so it leaves a very thin mist of salt on the glass, almost invisible. This greatly improves bonding compared to printing on bare glass. When printing on bare glass, the results are very dependent on circumstances: it may bond reasonable well in very dry weather (like freezing cold), but it may not bond at all in moist weather.


    The salt method works very well for me on long and low models, like rulers, at 100% infill: they stick like rock. But not so good for thin and high models like lantern poles: these might get knocked over.


    Have a look at my salt method, and my gentle atomic pull method here:





    1. If you think friction due to the end of the spool being wound too tight, and too hard to unwind: manually unwind a bit,  and wind it in the opposite direction around a skater wheel, thus bend it in the opposite direction. Then let go and do the next part. So it sits very loose on the spool, has the same bending-radius as the bowden tube, and causes almost no friction and no resistance when unwinding and transporting through the tube and nozzle.



    2. These kind of long low models print very well using the salt method: I have done many hundreds without any problem.



    3. The thick sort of ring or blob in the white filament after an atomic pull, indicates that the teflon coupler is completely worn-out.  When printing, half-molten filament gets stuck in that widened area, and causes underextrusion.



    It should look like the orange one at the bottom here: totally straight. The other two were pulled-out a bit too soon, before the filament had completely solidified. Obviously, these were all done while changing color, to remove remains of the old color from the nozzle. Also check if the cone-shape is good, like here.



  19. I once measured the bed temp with an infrared gun thermometer: the edges were ca. 5°C cooler than the center of the bed, I guess due to convection that pulls-in cold air from the sides. But below the nozzle-fans, the bed could be up to 15°C cooler than elsewhere, especially on small models where the nozzle kept blowing on the same small spot. Cooling fans are there to cool... So that could cause warping in those spots.


    Also, on UM2 printers (which have a less powerfull feeder than newer models), the left-back position is more likely to cause underextrusion, and thus poorer adhesion to the bed, than the front-right position. Because the filament has to take a much tighter curve to get into that spot, and this causes a high friction in the bowden-tube and in the nozzle. High friction means more resistance, and thus less flow. I don't know if and how this affects newer printers.


    And there could also be dirt or grease in some areas, due to uneven cleaning.


  20. I am very much pro recycling and pro clean oceans. So it is good that this plastic garbage is collected.


    But that garbage collected from the oceans is a mix of ABS, PET, PP, HDPE, LDPE, PC, PS, PLA and other stuff. And it is partially broken down by sunlight, eaten by bacteria, contaminated with salt, bacteria, algae, sand, paint, paper labels,etc... There is no way you can make high-quality products from that in an economical way. To sort-out and purify this garbage, you would need to burn maybe 10x more oil than it takes to make new plastic from that oil.


    So I think the best thing to do with oceanic garbage is burn it in a very well controlled way (to prevent and remove smoke and poisonous gasses), in order to use that heat to produce high-quality goods from other raw materials.


    This is similar to electric cars: people think they are very environmentally friendly. But they don't understand that to produce that electricity, transport it, and convert and store it as chemical energy in the batteries, it may take 5x more oil than just driving an efficient car with diesel engine.


    Let's have a look for Belgium: our nuclear plants will be shut-down in a few years, we have almost no natural resources like water, wind, sun. So, electricity has to come from generators fueled with oil, gas or coal. If we would want to run all transportation on electricity, it would require about 600000 wind turbines on 30000 km2 land, or 20 turbines per km2 (everywhere, thus also in cities). And at night there is no sun to load the cars. So this is not possible, and electricity has to come from fossile fuels. Lets use oil in this example, because that is similar to diesel.


    Let's examine the efficiency of each step (all numbers are crude estimations):


    - electric generator running on oil, in a remote industrial plant: 45% efficiency (such a generator is actually a turbine like in an airplane jet engine, almost identical, and depending on its injection nozzle it can run on gas, light oil, kerosine, diesel,...)

    - transformation up to high kilovoltage in several steps, transport over lots of kilometers, and then transformation down in several steps: 70% efficiency

    - battery loader: 90%

    - battery itself: 50% (maybe 70% if brandnew, but it degrades very fast, as we all know from our laptops and phones)

    - motor electronics and auxiliary circuitry: 90% (most people don't realise that electric cars do have a very complex water-cooling system, way more complex than in combustion engines, plus an oil pump and oil system, and a gearbox and differential, all eating power too).

    - motor itself: 90%


    All combined, this gives an efficiency of about 11%. So, you would be much better off driving a diesel car with efficiency of 45%: that eats 4x less oil than your "environmentally friendly electric car". And then we don't talk about the huge amounts of rare-earth materials that the batteries require (which pollutes the third world), and the high recycling costs.


    So, you have to be carefull not to fall into that same trap when recycling plastics.


    • Like 2
  21. Technically you can make 3D-prints from photos: I did portraits in that way.


    However, it won't work well for coins: at best they are going to look like photos from coins, not like real coins. Because one side is getting light and thus highlighted, while the opposite side is in the shadow. So the 3D-print will look very weird, and only resemble a coin-photo if you look through it towards the light.


    Further, it depends on the material: you need a translucent material that lets light shine through.


    So, you can print a 3D coin photo from a 2D coin photo. If you want to print a real coin, you need or a 3D model, or a greyscale drawing in which each grey tint correctly represents the required height.


    • Like 1
  22. I just stumbled upon this old post. A guy Stefan from CNC-kitchen has made several testing machines himself. See his Youtube channel. But to setup such a machine and correctly interprete the results, you will need some engineering education.


  23. Exactly: standard 0.4mm nozzle, but 0.06mm layer-height.


    If you print a model in 5x thinner layers, and print 2x slower, it is going to take 10x longer to complete. And if you then need 10x thicker walls to withstand the pressure...


    My models are very small, so that is still acceptable. But for big models like yours, it is going to take forever. In that case I would consider post-processing, such as chemical smoothing or painting/covering.


    I recently did a post on chemical smoothing with dichloromethane here, with photos.


    Or consider painting with a thick epoxy or other paint for plastics. Car manufacturers often have paints that work well on plastics.


    Another option would be to print a mould, post-process it to get smooth walls, and then cast polyurethane, epoxy, or reinforced epoxies in it. Search on Youtube for mould making and casting, there are lots of good video tutorials. Then it is for sure water-tight, and you for sure have no issues with layer-bonding. The only thing is that you need to design the mould very carefully, so there are no undercuts, and apply huge amounts of release spray, so the cast does not glue to the mould. Also, use low-exotherm epoxies, otherwise the mould will melt. If I had to make such a model, I probably would consider this method, even though it is a bit messy. Or a combination of methods.


    Anyway, test all these options thoroughly on a *small* test model first.


  • Create New...