Jump to content

geert_2

Ambassador
  • Posts

    2,084
  • Joined

  • Last visited

  • Days Won

    34

Posts posted by geert_2

  1. Also important: use a 3D-program of which you can find lots of good training-videos, and that has a relatively low learning curve, feels natural, and still a lot of possibilities. So you can show the kids a video, and let them repeat.

    For toddlers it won't be suitable, but for kids of about 10...12 years with some computer experience, DesignSpark Mechanical might also be a good choice. It is legaly free (requires only registration), is a feature limited version of SpaceClaim, and you find lots of good tutorials on Youtube, or elsewhere.

    Most of the functions use the same "push and pull" style as SketchUp, so it feels rather natural.

    Do not use SketchUp for 3D-printing: it was never designed for that, and has problems creating solid models. It was only designed for creating visual 3D-models of buildings for Google Earth, before Google had a good algorithm to produce those 3D-models from air photos.

    Maybe you could try a few programs yourself, and then find a few kids in your neighbourhood, let them try the software, and aks them what they think.

    What models to design? Well, ask them what they want. Something small (so it doesn't take ages to print), that they can use in their own life.

  2. I think in cars they call it "partial slip", most noticable when braking very hard. On dry surface, you get the best grip of the tires when they slip about 15% (according to an article in an automotive magazine I read years ago). At which point they already start making a grinding noise. A 100% slip is when the wheels are locked up while braking. And (almost) 0% slip is when the car is just rolling down the road, without adding power, without braking. So most ABS systems are designed to prevent locking up, but allow the partial slip that gives the highest friction and best braking performance.

    In 3D-printing filament, under a microscope this slip is visible in that the square pits that the feeder bites into the filament, get stretched out into diamonds, due to the mechanical force and rotation of the feeder wheel against the filament. So if the pits get stretched out by 20%, that would amount to a slip of 20%, and thus 20% underextrusion. So, less friction in the feeding system, should give less deformation of these pits, and thus more accurate extrusion. I guess that is how it works? :)

    • Like 1
  3. On my two UM2 (non-plus) I do atomic pulls more gently and with less risk of damage, as follows:

    - Put print head in a front corner.

    - Remove old filament, remove bowden tube, insert a piece of filament used as cleaner, heat nozzle to appropriate temp,

    - Gently push filament manually through nozzle, until it's a nice stream.

    - Dial nozzle temp to zero.

    - Gently push a bit more, but not too much.

    - Then do a small manual retraction of a few millimeters: this takes the pressure off the half molten filament, at the border between nozzle and teflon coupler, and makes it easier to pull out the filament later. Especially if the coupler is getting worn out. Otherwise a thick blob or lip would form, which makes pulling it out very hard.

    - Cool the nozzle with compressed air (if you have an oil-free compressor!!!), or just let it cool down, or use your lungs as compressor. :)

    - Let it cool to room temp. This deeper cooling makes the filament shrink more, and helps pulling dirt loose from the nozzle walls.

    - Wait a bit longer, so the inner core of the filament too has time to cool down to room temp.

    - When at room temp, gently wiggle and rotate the filament. Very often, with a snapping sound it will start rotating. Always do this wiggling and rotating gently, without brutal force.

    - Then heat up the nozzle again, to about 80°C for PLA (other materials require other temps, try what works). Keep gently wiggling and rotating the filament while warming up.

    - And now you can pull it out, almost without any force.

    In this way, it pulls out as much dirt as with the traditional atomic pulls (=pulling with brute force), but there is no risk of bending any rods, or of displacing the nozzle or telfon couper. So it works much more gently, with for me equal results.

    The main differences with the traditional atomic pull are: first the "small manual retraction" when cooling the filament, then deeper cooling untill room temp, and then the wiggling and rotating before and while heating the nozzle up again, and pulling much more gently. It is the rotating and wiggling action that does most of the cleaning here, not the pulling.

    This works very well with PLA and Ultimaker nylon, but less with flexible filaments like NGEN and ABS, which tend to break at the edge of the hot-cold transition. So I use PLA or nylon to clear out NGEN residu.

    I do not know if this method would also work for the UM3? Maybe one of the UM3 developers could shine a light on wether such rotating action while doing an atomic pull is recommended or allowed? I don't know how fragile or robust the new cores are, and how well they would absorb these forces?

  4. On my UM2s the speed of printing increases, and the speed of extrusion changes accordingly (so you get the same filling of the model); but also the speed of traveling increases. It does not change temp.

    However, if you make brutal speed changes halfway a print, you will see that in the model: it gives different ringing effects around corners, and if the model is small, the difference in cooling time per layer is also visible. Further, if the speed increases brutally, filament has less time to melt in the nozzle, which will give a higher pressure and less extrusion. This is also visible in the print as a faint horizontal line.

    • Like 1
  5. Do not use windows cleaners, white spirit, cheap alcohols for cleaning. They tend to leave soap or oil residues on the glass, which destroy bonding. Only use isopropyl alcohol, which dissolves oils. Then wipe again with pure water. And don't touch with your fingers. The current marks look like fingerprints. :)

    If you live in a moist climate, moist air also reduces bonding. Then I would suggest you try a bonding method: try my "salt method" (PLA only), see: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

    Or try dilluted white wood glue (1 part glue in 10 parts water), or hair spray (spray on a tissue, then wipe the glass; never spray directly in the printer), or the supplied glue stick (you can wipe and spread it with a wet tissue afterwards).

    Whatever method you like most.

    Also, make sure the build plate is warm enough, around 60°C for PLA.

  6. I have very little personal experience with lathes, and that is 30 years ago, thus not relevant. But a now retired collegue of mine is a professional metal worker; used to work on big manual lathes all life. He also tried a small lathe for small models, but that didn't work: it was okay for wood and plastics, but had not enough power for metals. And the clamps for the tools would move if subjected to too much force from the milling. So he dropped that machine, and went back to the big stuff, even for all plastics. I don't know which brand and model, but it was a very small thing, 50cm long or so.

  7. If the filament is near the end of the spool, thus if it has a very narrow bending radius, try stretching it manually (if you haven't done that already). Wind it bit by bit in the opposite direction around a 7cm diameter skater wheel, or something similar. So it has a bending radius of about 50cm, almost equal to that of the bowden tube. This will greatly reduce friction in the bowden tube and in the nozzle.

    Further, if you remove the bowden tube and manually put a piece of filament in the nozzle, you can often feel if the coupler is getting worn out: this often causes a higher friction at some points. So I think Neotko is right that you should have a look at the teflon coupler.

    These two things will probably solve the underextrusion part.

    The bending problem I guess is going to require thicker walls, especially at the bottom, and indeed switching on the fans early, after the first or second layer, so it gets good cooling as soon as the first layers are put down.

  8. If hard filament like PLA is wound too strong on the spool, especially near the end, then this creates a huge amount of friction in the bowden tube, coupler and nozzle. Also, this narrow bending radius acts as a strong spring, that resists unwinding. The UM2 feeder motor has enough power, but seems to have difficulty translating this power into filament movement. I guess that the pits that the knurled wheel bites into the filament get stretched out (=sort of partial slipping). If they get stretched by 20%, then that would amount to 20% underextrusion.

    For PLA, try this: with the PLA spool installed on the printer, unwind a bit of filament. Be careful that the spool does not keep unwinding all by itself, thus have someone or something braking it !!!

    Now manually straighten this piece of filament by winding it in the opposite direction around a skater board wheel or any other cylinder of about 7cm diameter. Then wind the next bit of filament in the opposite direction. And so on, until you have straightened about 2m of filament.

    Then gently roll it up again on the spool. Now it will be sitting very, very loose on the spool, with a radius of about 30 to 50cm. Now it will have almost no friction anymore in the bowden tube and nozzle, and no unwinding resistance.

    Try if this works.

    There could be a lot of other causes for underextrusion: dirty nozzle (partially blocked, narrowed), worn out teflon coupler, bowden tube clips clamping too much and narrowing the tube, too low printing temp, etc... But since you already changed these or verified them, I would guess too much filament friction and unwinding resistance due to a too narrow bending radius is most probable. Especially for hard materials like PLA, that do not bend easily.

    • Like 1
  9. Are you really sure you would want Windows 10 to flash your printer's firmware (if I understood that well in the docs), and to have it provide all the drivers?

    This might not be as bad as let Windows 10 flash your airplane's computer and provide its own drivers, but still...

    This sounds like a nightmare to me, and a recipe for disaster. Especially with Windows' automatic updates that you can not disable anymore (even not the incompatible ones), so they are going to happen halfway a 40 hour print? And with its huge amounts of spying functions and back doors? That means that not only Microsoft but the whole American government can snoop into your system and printer.

    Google for "windows 10 spyware" (without quotation marks), and enjoy. Or not. It's really creepy.

    (Note: I am not involved in Ultimaker development; just another user.)

  10. Cool and inventive concept, and nice achievement, especially for a student job. It must have been a mathematical nightmare, with all those 3D-rotations.

    But I still don't see how this concept is going to work for an upside-down U-shape, a table, or similar bridges, where you would need a 90° upwards rotation. Or for small models, where the head would have problems reaching into the details, while the table is rotating.

  11. What happens if you print the infill at a different density, or 100% filled? (Or just a small test part of it, to not waste too much material.)

    In my experience, any sudden changes in layer area to print give clearly visible markings on the outside, especially on small models, due to changes in cooling time for that layer.

    So, if the infill pattern changes halfway, I guess that could cause this effect too. Further, a change in pattern could also give a change in support, and thus be visible too.

  12. I don't have PVA, but I store all my filaments in a big closed box, and put a big bag of silica gel in it. These can be found in car shops, used to dry car interiors to avoid condensation on the windows. I use silica gel bags with a blue indicator that turns pink when the silica gel is saturated and needs to be heated in the microwave to regenerate. Then there is no need to heat the filament or storage box for drying.

    You could also use a smaller food box (the size of one spool), drill a hole in it for filament exit, fix a spool holder in it, put silica gel in it, place that behind your printer, and thus keep your filament dry while printing. For sensitive materials in moist climates, this might make sense.

  13. What does it do when you print a square plate of let's say 30mm x 30mm x 1mm, 100% infill? Without brim. Does the infill come out well, or also weird?

    And in your photo, is that overextrusion (too much material), or is it the material curling up due to poor bonding to the glass, and thus giving a different reflection? This is hard to see in the photo, but it should be clear if you remove the print and inspect its bottom.

  14. I think for the little fans, you should rather calculate 1000 to 2000h to be on the safe side. As in the old Pentium and Pentium II computers, their bearings might wear out fast, depending on the production batch.

    I have one UM2 fan (the tiny nozzle-cooling fan) that started to make noise after about 1000h. Which means its bearing is getting worn out. So I injected that bearing with oil (*), and now it is running smooth again, but this won't last forever of course.

    (*) I did this by puncturing the silver seal with an injection needle and injecting a drop of oil through that hole. If you puncture the seal near the top, the oil will not leak away. But if you do this too early (before the bearing being worn out), or if you use incompatible oil, it might really reduce the life of the bearing. If you wait too long until the bearing is worn out completely, there is already too much damage and injecting oil won't help much anymore. I believe the optimal point is when the bearing starts to make more noise than usual (so it is near the end of its life anyway), but before it is slowing down due to wear and wobble.

    Anyway, to be on the safe side, order new ones as soon as the current ones start to make noise. But these fans are generally available parts; it should never be a problem to find them.

    For us, this open source aspect was very important and a major reason to choose for Ultimaker, instead of a closed source printer of comparable quality. Open source components, availability of component specs, multiple filament manufacturers, a good forum, and standard spool sizes were high on our priority list, as well as very good printing quality, of course. This is a very new and volatile market, and you never know what is going to happen. It is like cars before 1950. But even in stable old markets, you never know: just look at big car brands like Chrysler (now owned by Fiat and the cars branded as Lancia in Europe), Saab, Rolls-Royce, Bentley, Rover, and so many others that got taken over or totally disappeared in the last years. It is a good idea to calculate this in, so you can relax and don't have to worry about it. And buy spare-parts from Ultimaker itself, instead of from cheap low quality sources: this will also increase their motivation to keep those spares in stock. :-)

  15. You could remove the bowden tube at the back end, and then try to feed the material by hand. Heat up the nozzle and see if you can easily push it through. Or remove the bowden tube at the nozzle end, and feed it manually there. At least this will give you an indication if there are any mechanical blockings.

    Could it be that the filament is too thick (>3mm), or that it is squished into an oval or thin plate too much due to incorrect spring pressure in the feeder. If the spring pressure is too high for the material, one diameter might become let's say 3.5mm, and the other only 2.4mm, or so, and then it will get stuck in the bowden tube, teflon coupler or nozzle.

    The PETG that I tried did work between 215 and 250°C, with lower temps around 220...225°C working best, at relatively slow speed.

  16. Hi Geert,

    i think aluminium will wear out / deform the brass nozzle over an averrage time too. I told my dad about my project and he had already the right silikon for it :Pin his workshop. It`s a package of 3 special silikons from which i give the "Dirko HT (+315°C) " the first try (Link to special silicone). I already lay down a layer on a pla print, its sticks to it and its more like a realy soft rubber way more tough than normal silicon. If this doesn't work I'll come back to your idea :).

    Yes you are right. Try that first. When writing my remark, I had mould-making two-component silicones in mind, with an upper temp a bit above 200°C. Mine changed from very soft and a green color, into brittle and whitish. But I hadn't thought of these high temp moisture curing silicones for engine systems.

  17. ...

    And, of course, the other possibility is to reduce polygons.  Chances are that you won't see all that detain in a 3D print in any case.

    Yes, I think this is a good idea: any details smaller than about 0.4mm in X- and Y-direction, will be hidden by the 0.4mm nozzle lines anyway (if you use a 0.4mm nozzle). And if you print in layers of 0.1mm, any vertical details smaller than 0.1mm will also be lost anyway. So you could as well reduce the details to these resolutions.

    If you plan to sand and paint the model for a smoother surface, you could reduce everything to 0.5mm resolution, I guess. The rest will be sanded off anyway.

    If you can cut the original model in small pieces, you could try this on a small part first, so you don't lose too much time and plastic if results are not satisfying.

  18. Have you tried printing this with the build-plate temp 5°C or 10°C lower? Or have you measured that build-plate temp? If you could get the temp just below where the PLA starts to get flexible, but still has enough bonding, it might work.

    For a stable bottom layer, I think it needs to be at least 0.5mm to 1mm thick. If you only make it 0.2mm or so, it will not have enough rigidity to keep the walls upright, I think.

  19. Mhm, lets call them "nozzle masks".

    Yes, but i prefer a softer material like  high temperature silicon toned in a mold,  to prevent the nozzles for weareout. Trying to do only the standart retraction for the filament (except for the material which dont like to stay a bit longer in the nozzle).

    I think most silicones will get hard and brittle, when exposed to hotter temps (250°C) for a longer time. Maybe you could also try *pure* aluminum, which is very, very soft? You can almost scratch it with a finger nail, and it is too soft to tap threads and screw bolts in it. Maybe try doing a scratch test between such pure aluminum and a brass nozzle (or whatever nozzle material you use), and verify under a microscope if it wears out?

  20. Just thought about this: you can not use other people and patients as guinea pigs. But you can use your own body. So, if you really want to know what these materials do in an aggressieve environment like the mouth: print a bridge for yourself, and put that over a couple of your own teeth. Use a transparant colorless material, so you can see all the little voids inside. For best accuracy, print slow and cool, but that means less flow and thus bigger voids in the model. Keep it in your mouth for a couple of weeks, and then watch it change color as bacteria grow and eat into the model. PLA is likely to decompose due to the aggressive enzymes whose job is to break down food, and PLA being bio-degradable... Other materials will probably get brown soon and start to smell. Part of the game is: you are allowed to clean the outside of the model when you brush your teeth, but not the inside, since you wouldn't be able to do that on a real crown too. So, any accumulated dirt will stay there. This would be a nice little experiment, but I am not going to do it. :)

  21. I don't think it is a good idea to use FDM 3D-printed objects in the mouth for more than a couple of hours. They are not accurate enough, so they are likely to cause irritation. But above all: the layer lines and little holes in-between are going to cause bacteria-growth and infections.

    However, maybe you could use a 3D-print as model to cast a crown from? Then you can sand, polish and seal the model prior to making the mould, so the cast does no longer have any layer lines or imperfectoins? Then of course you could use plain PLA, or whatever you like most, and print slow and at low temp.

  22. I don't really see the problem with the head traveling over the infill? This will be closed and become invisible anyway, so why would it matter? Or am I missing something?

    You are... OP is taking advantage of the infill patterns generated by Cura, to make an interesting outer shell on his print...

    Ah, okay, now I see. Thanks.  :)

    If the design software allows to create "patterns" (=terminology in SpaceClaim / DesignSpark Mechanical; I don't know the name in other packages), then another solution might be to create one hole, and repeat that a number of times in X- and Y-direction, with 0.5mm walls inbetween (for a 0.4mm nozzle). I used that to create a sieve, which has a similar pattern. This requires very little work. Then you might still have some blobs, but at least it is already a bit more beautiful since the nozzle is now traveling along the lines only.

    zeef1.thumb.jpg.4c0f2b050c6d9413ee6215f12145c401.jpg

    • Like 1
  23. I have a spool of Ultimaker nylon, but I haven't printed with it yet.

    Out of curiosity I cut off 10cm of filament and put that in water for a couple of days. After some time it does get more flexible, and it gets a sort of waxy feel: feels more like low-density polyethylene (LDPE) now than like nylon.

    A lot of 3D-printing materials are chemically modified for easier printing, to have less warping and better bed adhesion. Maybe one of the modifiers changes properties in water? Maybe one of the UM chemists could shine a light on this? I guess other brands may show similar changes?

    Since this is just one strand of 2.85mm filament, not printed, I don't have any idea what this would do to the strength of a print, and to layer bonding.

    I am gonna dry it at 60°C and see if it changes again.

    Edit: it has been drying for two days now, and it seems to recover to its original stiffness. At least to my subjective feeling, since I haven't done any scientific strength tests.

×
×
  • Create New...