geert_2

In my experience, if you print too fast, filament has too little time to melt (causing underextrusion), and the printer head's movements overshoot at corners (mechanical vibration), causing visible vertical banding ("ringing"). If you print too slow, the filament has too much time to melt (causing stringing), and the print head stays too long on the same spot and will melt the rest of that area too, causing visible deformation. You often see this when printing very small items (e.g. fine text). So you need to find the optimum inbetween, for each material and each model. Normally I use 50mm/s for standard models, and 25mm/s for very fine models, always 0.1mm layer height, PLA.

Just make sure you never leave it sitting in the sun, in a hot car, or in hot studio spot-lights. The model will not time-warp, but it may dimension-warp. (Don't ask me how I know...)

@jamado: if you want to print with very fine details, smooth surfaces, and hard-to-see layer lines, I would suggest that you read the thread by Cloakfind on smoothing PLA with aceton. This thread contains a wealth of good info on temperatures, speed, layer heights, etc. for finest quality. Plus it has a lot of photographs showing each test. It is a lot of reading, but worth it. I think it is this URL: https://ultimaker.com/en/community/10412-acetone-finishing-on-pla

That's what this 15 page topic is all about. There's even videos if you look through it. I think video is better than pdf in this case because people kind of have to see you actually do it and the number of seconds submerged and all that. Yes, a forum like this is perfect indeed to discuss and develop things together, and to show progress and experiences. Or to get help or instructions in case of problems. But after a couple of years, these good posts tend to get hard to find back, as they sort of get "snowed under" all the new stuff. So the same questions keep popping-up over and over again. That is why, in a separate section, good PDF-manuals and instruction videos also have their place too, in addition to the forum, I think. This allows us to easily point others to these references. Or to download them and store them in a convenient and easy to find place on our own harddisk, and use them as reference, even on places without internet connection. Or when you are in a hurry and don't have time to search all day in forums. It is easier to scroll through one known directory on your disk than through thousands of posts, in tens of different forum categories, on several different forums, where you don't remember where you once read that post that might now be usefull. Especially with a poor memory like me. So, often I print interesting forum topics to a PDF file for future reference. Not just on this forum, but also for example on 3D-design, or on flight simulators, etc. But most forums are not set up for easy printing (they lack a good printing style sheet), so these PDF-prints often get messy: things falling off the page, too much white space left or right, huge gaps, etc. That is why I do appreciate dedicated PDF manuals too, provided that they are designed to be easily readable on a computer. That is the idea.

It took some time and observation to find this method, there was nothing random about it. I often have difficult models: very long, thick, 100% filled, with chamfers at the bottoms. So they exert a lot of warping forces, even for PLA. Printing on bare glass worked only in freezing cold weather when the air was very dry. But even then corners would lift. In moist weather, printing on bare glass did not work at all, and there would be no bonding at all. So, surface tension or surface charge seemed to play a role. Glue stick worked reasonably, but some models got stuck too hard (those that had a very big contact surface) and were very hard to remove, so they got damaged. Other models wouldn't stick or would come off (high models with a small base and huge chamfers at the bottom, 100% filled). And the glue was messy. At that time I did not know the cleaner "gr5 method" of dilluting wood glue 10% in water. Or smoothing the glue-stick glue with water. And 3DLAC didn't exist yet, or was not yet known. If I had known these methods, I would have tried them, and maybe never felt a need to search further. Also, I could not use blue tape, since the bottom of my prints had to be absolutely flat, without the ridges and roughness of the tape. So I needed to find something else. I had also noticed that after cleaning the glass plate with soap-water or window cleaner, bonding was far worse than before. Cleaning with pure tap water worked best. I reasoned that soap reduces bonding by reducing surface tension. To glue something, the substrate (thing to glue) needs a higher surface tension than the glue itself. So, for 3D-printing, the glass plate would need to have a higher surface tension than the filament, I thought, for it to stick. Otherwise the filament would just curl up into a ball and only stick to itself, not to the glass. Thus I needed something that would *increase* surface tension, and then maybe that would increase bonding too. One of the few commonly available chemicals to increase surface tension is salt, NaCl. That is how I got the idea of wiping the glass plate with salt water. I don't know if my reasoning is correct, it could even be totally wrong. But at least the method works very well for PLA (not for ABS). For PLA the bonding is very strong when the glass plate is hot at 60°C. But there is no bonding at all after cooling down to room temp. So, for me this is the perfect solution. I used several colors of Ultimaker PLA and colorFabb PLA/PHA, and all work equally well. I hoped it would also work for filaments that are chemically related to PLA. But for oil-based filaments such as ABS, it obviously doesn't work, and I had not expected it to (but there is always some faint hope of course). That is the background of it, in short. I haven't tried anything else. So, no random food stuff, syrup, milk, cornflakes, or whatever. (Although trying sugar syrup would make sense, since it is sticky, and is water-soluble. So you can get the print off by dissolving the sugar. Some people seem to use this with good results. But I haven't tried, no need to.) For the full manual, full background, and pictures of the salt method, see: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

Just for your information: my "salt method" to glue models to the glass plate (=by wiping the glass plate with very salt water, and let it dry into a thin mist of salt) does *not* work with ABS+ filament. Even simple models warped and came off, so I had to abort them. I tried it on a spool of ABS+ that I still had laying around, but never used before. (Note: ABS+ is supposed to be an improved ABS, which should have less warping than standard ABS. But I don't know the precise chemical composition. I guess it may have a bit more butadiene than usual, to make it softer and warp less? At least, it is much more flexible than the usual ABS of Lego-bricks or Playmobil toys, and it has a somewhat "rubbery" feeling.) Very weird was that the first layers seemed to "explode" against the salt. It made a crackling sound and foamed. After two layers that foaming and crackling stopped, and it printed very smooth. Until the bottom warped and the model came off, so I had to abort the print. The foaming and crackling is similar to what we could expect from too much moisture. But the filament was very well vacuum packed with a big desiccant pack, so it doesn't seem to be moisture from the filament. (And otherwise it wouldn't have stopped crackling after the first two layers.) So it seems rather to be moisture caught in the salt, or a chemical reaction to the salt. Weird too is that atomic methods wouldn't work: the filament kept breaking, instead of coming loose. This even though when cold the filament was very flexible, and I could bend it 180° without breaking. It seems it got totally glued to the nozzle walls, which is weird, since molten ABS is supposed to *not* stick to metal, contrary to PLA. Or maybe it got glued to the black residu or the etched surface from previous PLA in the nozzle? Since I could not do atomic pulls, I had to flush the remains through the nozzle by pushing with fresh PLA filament, at 220°C. Fortunately this worked, although the ABS+ and PLA seemed to "fight" a bit: they didn't mix well while flushing. After flushing with 15cm of PLA filament, I could succesfully do atomic methods with PLA again. Anyway, for the "salt method", in summary that gives: - PLA: excellent bonding, even difficult models do not lift corners. Recommended. - ABS+: very poor bonding, even very simple models warp, come off, and fail. ABS reacts to the salt or moisture in the salt. Thus definitely *not* recommended. Try any of the wood glue or glue-stick methods, or 3DLAC, or so, instead. - PET: no personal experience, but some time ago ultiarjan reported poor bonding (worse than with the glue-stick), and warping. Thus not recommended either. - Other materials: no experience. If you would have tried the "salt method" on any other materials, please let us know the results, good or bad.

I had also stumbled upon this guy and this video, I don't remember how. His filter concept may need a bit of refining, but I like the simplicity of the basic idea. And for sure silicone oil does not get ransic like vegetable oil, which might become a tough mess over time. And it doesn't seem to make the filament slip in the feeder either (which is a bit surprising to me). Off-topic: also have a look at his gaming computer in the background: with four huge screens and 24 CPU-cores, I guess it will make any flight sim, or a GTA V race game, a whole new experience. Slightly above an average setup... And if you use Windows 10, also have a look at his videos concerning the creepy privacy issues; it is very well explained.

Regarding the small size of this portrait, only a few fingers, the smoothness is impressive. I think you should write a manual (PDF) with your best techniques. It could become the "standard manual on smoothness" in 3D-printing, or something like that. :-) For this item, let me guess: you have acetoned it, cleaned it a bit, and then polished it with some plastic/acryl polishing paste? But no paint, no varnish, I guess?

Yes, PLA starts deforming slowly around ca. 50°C. I also had that. Never leave it in the sun, behind a window, or in your car (even not in spring, autumn). (I can read Deutsch, but writing is more difficult.)

What I am sometimes missing now, is status info during printing, such as: nozzle temp, bed temp, fan %, speed (actual mm/s, and in flow %), flow %, layer nr. of total layers. This comes in handy when playing around with settings, so that I don't have to open all submenus to find out. Would it be possible to add this, if it would fit on the screen? Probably some parts need to be alternating. For example (I limited each line to 20 characters, like it is now. I don't know if you can put more text per line?): -------------------- file_being_printed time to go 9999 min nozzle 222° bed 100° fan 100% flow 100% layer xxxxxx/yyyyyy material flow 100% material PLA -------------------- (Anything I forgot, or that other people would find useful?) Also during warming up and cooling down, more status info would be nice, such as both nozzle and bed temp. For the buildplate, a function to manually move the plate up and down step by step, by turning the wheel, would also be nice. This when not printing. (I don't know if it would be a good idea while printing, maybe too risky?)

Ik calibreer eerst volgens de officiële methode, met een blad gewoon papier (ik heb twee Ultimaker 2, geen plus). En dan maak ik een testprintje, een dun kader van 200mm x 200mm. Tijdens het printen van die eerste laag regel ik de schroefjes voorzichtig een beetje bij tot het goed lijkt. Maar ik heb de nozzle liever iets dichter bij de plaat. Dus soms geeft dat inderdaad een beetje overextrusie bij die eerste layer, afhankelijk van het model. Maar de onderzijde is dan wel altijd bijzonder mooi glad, zodat mijn onderdelen nadien goed over elkaar schuiven zonder haperen.

This all sounds very interesting. A question: how difficult is it to machine this Celazole PBI in reality? And how resistant is it to breaking or chipping, by your experience? I read in the specs above that it might require special tools? But to me it is not totally clear if that applies to high speed, high volume serial production only, or also to small scale production? And if you would print high temp materials like nylon or ABS-PC blend (270°C), does it still deform, or not at all?

The finish and detail of your models is excellent, almost indistinguisable from real bronze sculptures. Really amazing. If there is anything that could still be improved, I think it would be the modeling of the eyes. But that is rather a question of personal taste. Some people like the "blind eyes" style (=without pupils and iris), because it gives some sort of distance and abstraction (I don't know how to describe it very well in English), and it leaves more to the imagination. Other people, including me, rather prefer more realistic eyes with clearly defined pupils and iris, to give the face a stronger expression, as if it is looking at you directly. To see what I mean, have a look at the video of Philippe Faraut, where he is sculpting in clay. Especially the video named "Crazy" has very expressive eyes, among the best that I have seen in sculpting (at least according to my taste, other people may have different tastes, of course): http://philippefaraut.com/pages/video-clips But I don't know if such eyes would print well in 3d, due to the overhangs? Maybe if the model would lean backwards a bit?

Out of curiosity, a question to the Ultimaker developers: have you ever tried "Rulon" and "Fluorosint" for the PTFE couplers? Rulon is an enhanced PTFE material, with higher mechanical and temperature resistance than normal PTFE (some sources say 260°C, some say 288°C max continuous temp). It is used for high performance seals and bearings. Being a PTFE-based material, low friction and self-lubricating, it might also work well with PLA? Fluorosint is PTFE filled with mica, with one version going up to 315°C max continuous service temp. I stumbled upon this, after reading the recent post by gudo on using PBI (Polybenzlmidazole, Celazole) to replace the teflon coupler. When looking up PBI, I also found these materials which may be a lot easier to machine. See: http://www.boedeker.com/rulon_p.htm (red color) http://www.boedeker.com/fluorosint.htm (grey color) http://www.plasticsintl.com/enhanced-ptfe.htm http://www.plasticsintl.com/rulonlr.htm (red color) http://www.plasticsintl.com/rulonj.htm (brown-yellow color) If you have tried them, how wel or how bad did they perform? If not yet tested, it might be worth asking a few samples? Geert

Try reducing the temp in steps of 5°C while printing the fine details. With some filaments, you may be able to go down to 185°C. Also reduce speed accordingly to improve the shape of the text and lines (less vibration effects of the head), and to give the filament time to melt. For small text, similar to your "2016", I usually get the best results around 190°C and 25mm/s speed, for colorFabb and Ultimaker Pearl filament. (I don't know the yellow.) This greatly reduces the stringing and the blobs, but I could never totally eliminate these "blobs": whenever the head wants to jump to another area, it pauses a fraction of a second, retracts, and then jumps. This leaves a small round mark at the take-off point. I think your design is at about the edge of what is possible with a 0.4mm nozzle.

Bill2 and peggyb: thanks for the info! Geert

I use DesignSpark Mechanical, distributed by RS Components. This is a free but limited version of SpaceClaim. Pro: very handy user interface: can work by pushing-pulling features or by entering dimensions numerically. Easy to learn. Lots of good training videos available (you can also use SpaceClaim's videos, if you take into account the limited features of DesignSpark Mechanical). Ideal for designing mechanical and machine parts. And for parts that you want to keep editing afterwards. Thus ideal for let's say designing parts for a 3D-printer, feeders, etc... Contra: STEP- and IGES-file import/export is a non-free option, and it is quite expensive. There is no photo-realistic rendering. Not suitable for organic life forms. Not so good for complex designs (e.g. a complete motor). So, for simple machine parts: yes. For art and life forms: no.

That machinable wax, is that available in 2.85mm filament? Or do you use some feeder/melting pot you designed by yourself on top of the printer head? If filament, how wel does it transport through the bowden tube and feeder? Thus without breaking, and without grinding? Such wax might also be very usefull for artists who want to make bronze sculptures in the traditional way by moulding and casting. Instead of making a clay original, they could print it in wax. Or for casting small automotive parts in iron or aluminum, such as model airplane engine parts.

For your next prints, I would suggest that you give the "salt method" a try. (=Wiping the glass plate with very salt water, prior to printing. No glue, no hairsprays.) For PLA this gives a very strong bond when the glass is hot, but no bonding at all after cooling down to room temp: the models just pop off. No more need to take the glass out of the printer, no more recalibrating the build plate... I use this since more than a year, and have had no problems since: no bonding problems, and no removing problems. For the full description, see the PDF-manual at: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ If you try it, could you let us know how well (or how bad) it works?

Thanks!

Since more than a year I haven't levelled the bed on both my UM2 printers. No need to. Using the "salt method" to glue PLA to the print bed, my models stick like rock solid when hot, and pop off by themselves when cold. So I never have problems removing the models from the plate, I never need force, and never need to take the build plate out of the printer. So, for me autoleveling would be useless. (Although I can understand that people who daily change nozzles and materials might see things differently, especially if the nozzles are of different length.) Instead of autoleveling, I would rather prefer a thicker and more stable aluminum plate and glass plate, which don't vibrate, are totally flat, and do not warp when heating up. The current "unevenness" of the glass plate may be less than 0.05mm, but when printing the first layer at 0.1mm, it becomes visible.

If you created the model or you can edit it, you could try designing the supports yourself. For very fine details, you might want the supports to sit closer to the model for higher accuracy (e.g. a gap of 0.2mm). And for larger parts the support could be further away, for easier removal (e.g. a gap of 0.4mm) but then with less quality. Also, I often design features into the supports to make removal easier, for example: extensions on which I can get a good grip with tools, or spaces where I can access some areas easier with a knife. Or holes to put in a hook and pull the support out. It takes time to design these things, and think it over, but it is worth the effort. These are things you can do manually, but that would be extremely difficult to generate automatic. (I just design all supports in the same program in which I create my models, DesignSpark Mechanical.)

Wow, that is really amazing detail. Do the parts move correctly too?

This is how the bowden tube should sit in the teflon coupler, and how I would suggest that you cut off the tip of filament for easier inserting.

Hi nallath, Can old and new (>2.1) Cura versions be installed parallel on the same system, without messing each other's settings up? Do they use separate directories and files? If so, people could simply install both, and use the version that works best for a particular task. Or they could just keep the old one as a fall-back. Geert