geert_2

I would also suggest installing big diameter silent fans. Preferably with air canals running towards the ceiling, similar to the fumes extraction system in a kitchen. But then you would have to drill big holes in the cabinets above it, and install tubing in it. Just like in the kitchen. Maybe also increase opening in the back for a good equal flow at low speed? This would also help extracting fumes. Anyway, it looks really nice and clean, professional.

When printing overhangs, the material tends to partially sag (since half of the trace is printed in the air), and it tends to curl up when cooling (since there is nothing below to hold it down). Then the nozzle bumps into these curled up edges. All these effects mess up the sides. Possible solutions you could experiment with: print as cool as possible, so it melts less. And print in thicker layers, for example 0.3mm instead of 0.1mm, if the model allows it: this may have a big effect. Place a desktop fan at low settings in front of the printer for additional cooling. Try this on a little test print, so you don't waste too much material and time, until you get the best results.

Or maybe a different sort of blue tape (for example with wax on it), or different environment characteristics (moisture, heat)?

And what about the heat? Don't they overheat? Or is there extraction or cooling?

For printing gears, which are usually low profile, maybe you could print them one by one, instead of all at the same time? Then the printer will complete the first gear without jumping anywhere else, before starting the next? Not ideal, but maybe it is a temporary work-around? However, if the gears would be very small, insufficient cooling might become an issue. For diameters like 20mm, it should be okay. Another disadvantage is of course that you don't get as much models on one plate: the head neads enough room so it doesn't crash into already printed items.

@kmanstudios: congratulations!

Your *environment* could also play a big role: air moisture and temperature. In the beginning a couple of years ago, when printing on bare glass (=no bonding method), I had reasonable bonding in cold freezing weather (=very dry air). However, in moist, rainy weather, I got no bonding at all. Nothing else had changed. It was very frustrating. This is when I started searching for a suitable bonding method. Now, for bonding PLA my "salt method" seems to level out these differences, even though I still don't know why it works, from a chemical viewpoint. Salt method = gently wipe the glass plate with a tissue moistened with salt water. Keep wiping while it dries, so it leaves a thin, almost invisible mist of salt stuck to the glass. When hot, models stick like glued. But after cooling they come off by themself. It is very well suited for low, long 100% filled models like I usually print: I have done maybe 1000 models since then. But it may be less suited for high narrow models.

Well, that are truly dynamic poses! I like them. Mine look stiff in comparison. :-) Yeah, maybe that is the difference between an artist and a technician? Can they still be printed well in scale 1:100?

Feel free to recycle and re-use these ideas, I would welcome it. At this scale it have to be expressive games with lots of body motion anyway. Kids playing Monopoly wouldn't print too well... :-) Also, this helps overcoming the too flat "cardboardness" of this concept. I had thought about doing a couple of yoga-poses too, or kids playing with big water guns, or playing "cowboy and indian" (this was a favorite in the old days; not sure if it is still politically correct, probably not...). But their use in architectural models might be somewhat limited, unless someone is going to design a school or so, or a scouts camp. Thus that will be for later. Maybe.

Probably like other PLA? Maybe a little bit longer if the bronzefill would be less transparent to UV light than plain light coloured PLA? So, expect it to get a bit duller, harder, and brittle in a year time. But it shouldn't crumble apart (or maybe after several years?). The worst might be the heat: the models could deform severely in summer, especially if it is a summer like we have this year in Europe, really nice and warm. I would suggest that you just try it, and give feedback. This would be a very interesting experiment.

What was it supposed to be? A phone case?

Improved set of people: I was not really happy with my previous set: it was too monotonous and just not good enough. But at that moment I had run out of time and out of inspiration... So here is an improved set, with more people, more variation, and in which the people are more alive. Now there are around 35 people in total, men, women and kids. And no more duplicates, although some people still have similarities since they are derived from each other, as it goes in families. :-) This set supercedes the old one. I also corrected a few details that did not come out well in the first set, e.g. hats and shoes. The philosophy is of course the same as in my previous post. STL-file: humanoid10.stl ZIP-file (hope it uploads well):cardboardian_people.zip The set is now printing and looks good so far. First layers are already done. Printing parameters: 25mm/s, 195°C, 0.1mm layer height, colorFabb PLA/PHA, my salt method for bonding (you can see the traces of salt and the shadows they cast). Keep in mind that most of these people are only between 15 and 18mm high. They should blend in well with the cars from the previous contest, and with typical vegetation in architectural models. The zip-file contains the same STL-file, a few JPG-images, and also the native DesignSpark Mechanical files (=the program in which I created the set; freeware). It also contains a few intermediate versions: if you want to edit the set, these might be easier than the final version, depending on what you want to do. Still missing from the set: people sitting, watching TV, reading, laying down, sleeping, using their handy,... But these are not the most easy to design and print recognisably at scale 1:100. Maybe I will add them later, maybe not. Feel free to edit and improve the set, and to add people or accessories as you see fit. You can use it for any purpose you want, also commercially. But if you redistribute the set, or your derivation of it, you have to keep it free (Creative Commons CC-BY-SA license); you can not take that freedom away. Also, feel free to borrow the concept, and design your own sets according to this philosophy. I hope the set is useful. If yes, I would welcome a few pictures of architectural models which features these people. PS: due to limited internet-access during the holidays, it may take a few days or weeks before I see any replies and comments. Have patience... :-)

Results before and after "acetoning". These are just old scrap pieces, which I kept around for testing. Material is colorFabb PLA/PHA, Dutch Orange. Printed with 0.4mm nozzle, 0.1mm layer height, 50mm/s, 210°C, 100% filled, on UM2. The visible part of the model is ca. 70mm long, 10mm wide, 3mm high. The green pieces are silicone casts of it, as they show some irregularities better than the original models. While the seam lines are still visible in the silicone-treated part, they are almost "unfeelable": it feels way smoother than the untreated part. All little openings due to minor underextrusion are completely sealed, which is best visible in the silicone casts. Text is gently rounded. Text caps height is ca. 5mm, raised 0.2 or 0.3mm (I don't exactly remember). I did a few treatments of brushing on acetone, and let it dry inbetween. (Next time I will try one thorough treatment, instead of multiple short ones, to see if that gives any difference.) There was whitening indeed, but that was only superficial and could easily be brushed off with a hand brush, while the model was still somewhat soft. It has to be a soft brush, since a hard nylon brush caused scratches. And it should be a soft brush that does not melt in contact with acetone. :) I guess that the whitening is caused by additives that are dislodged or broken down, and that get deposited on the surface? There are no signs of warping. Immediately after treatment I could push my fingernail about 0.5mm deep into the model, while it was soft. But now it feels hard like original. Anyway, I am glad with the results: if I take care to print more carefull (slower, lesser layer height), this will make silicone casts a lot easier and better. Thanks to cloakfiend for the research on acetoning.

Maybe it could also have been Robert's troubleshooting guide? https://support.3dverkstan.se/article/23-a-visual-ultimaker-troubleshooting-guide

Thanks again for the write-up. I am trying the acetone brush-on method now on scrap models, just to get some experience and feel for how it exactly behaves. Doing a few brush strokes, let it dry, a few more strokes, etc... The layerlines and seam lines (on top) are still visible, but far less "feelable". Sharp edges like raised text get rounded quickly. Under a microscope it appears that the indents are filled up with material that is dissolved from nearby areas, and deposited there. Concerning warping and cracking, we'll see next week. :) The real moulds will be for august, I need feedback from the users in the hospital first.

In addition to printing on glass (without tape) as gr5 says, I would also suggest you put the models closer together, a bit further away from the edges. And then enlarge the brim so that is gets so big that it thoroughly connects all models. Otherwise, thin but high models might get knocked over. Or else, design a thin flat plate in your CAD models, which covers the whole bottom and connects all models. What could have happened to you is that one model started to wobble (for example if the tape was a bit weak in one point, maybe due to air under it, or grease), then the print head crashed into it, knocking it over, maybe caused a layer shift due to the impact, unsettled and threw off some more models, and then a sort of chain reaction started? Just guessing of course, but something must have happened...

I decided to take the challenge, and to design a set of people myself. :) However, when it comes to ultrafine details and lots of variation, there is no way we can beat the Preiser models with our FDM-printers and standard 0.4mm nozzles. So I didn't try that. If you want realistic, finely detailed models, the best option is to buy unpainted Preiser sets: this gives you about 100 little people for about 30 euro. Google for: preiser ho figuren Instead, I decided to take a totally different approach and optimise my set for 3D-printing, so it can be printed easily and fast. I designed it in harmony with the cars already shown in the previous contest. Thus my people are a bit abstract and cartoonesque, but they do fit-in well with the cars. They are sort of cardboard-people, "Cardboardians", living in the country New Cardboardia, which is nearby Tsunamia, Quakeland, and Volcanostan. Scale is ca. 1:100, as requested. But by slightly varying this, you would get more variation. The biggest advantage is that these people can be printed on their back, so they print way cleaner than "upright people", and way faster. After printing, they can easily be broken off the support bar which helps making printing easier. They require very little post-processing. They can be printed without warping, and without damaging while removing them, if printed in PLA using my salt method for bonding, I tried it. Other bonding methods should work too, but you may need to be carefull not to damage them when removing them. If you don't want to glue these people to your architectural model, you could glue them on a transparent piece of plastic (mica or similar) of ca. 10mm x 10mm, like Preiser does with most of its painted models. Then you can move them around and recycle them for the next project. The transparent base does not visually hurt. I recommend printing at 0.1mm layer-height, speed 20...25mm/s, and at a low-medium temperature. All can be printed with a standard 0.4mm nozzle, since all legs and arms are wider than 0.4mm. See the photos of upright printed people compared to my cardboardians below. The upright model is not my design, but downloaded from thingiverse, named "Elf girl", designed by Robyn, daughter of "southoz"; if I have all that info correctly. Actually this Elf Girl is a very nice design, and 3D-printing it in scale 1:10 (17mm) does not do justice to it. The red elf girl is printed with 4 together, plus a dummy cooling block of 10mm x 10mm x 20mm, at very low speed (20mm/s) and low temperature (180°C) in PLA. The green one at the left is also printed with 4 together, slow, with dummy block, but at normal temp (210°C). The blobby one is printed standalone at normal temp (210°C), without dummy cooling block, and slow; but it obviously had not enough cooling and is hardly recognisable. And the last Elf Girl is printed single too, slow, but at low temp (180°C). The cardboardians are printed slow (25mm/s) at medium temp (ca. 200°C). You can see that upright printed people look like zombies, where pieces of rotten flesh are coming off (strings and blobs). Or they look like aliens in space suits (not enough cooling). Not like humans. This weirdness might hurt in an architectural model. My cardboardians however, look way more neutral and unobtrusive. Even though the designs are crude and cartoonesque, they come out cleaner when printed. And they are similar to the cars, and to typical vegetation. So there is more harmony in the scene. The set contains several men, women, and kids. Most are unique, but a few of the neutral models have duplicates. (The additional bonus set was my first try, containing 8 identical people and thus not conform the rules, outside of the competition. But since I have it anyway, I can as well share it.) The models are under a Creative Commons Attribution Share-Alike license: you can use them for free, for whatever purpose, thus also for commercial projects (for example an architectural design for a client, or for publishing on the Ultimaker website). But you can not ask money for the design files: these have to stay free. You may make modifications and derivations too, but if you share them, it also has to be for free under the same license. cardboardians_set.stl cardboardians_bonus.stl Set of cardbordian people, scale ca. 1:100. Bonus set (outside of contest). Beautiful Elf Girl, by Robyn (downloaded from thingiverse), for comparison in print tests. Cardbordians as printed. I removed a couple of thin hairs, but no other post-processing. Elf Girl (printed upright) vs. Cardboardian people (printed on their back). See body text above for more info on printing parameters. Preiser HO-scale people (=1:87, close to 1:100). But they also exist in other scales. In case you need really detailed people, and a huge amound of variation...

Thanks for the info. So I will go the "cooking way", and just experiment. :) The shrinking might be a bit of a problem though, as the silicone castings should keep their shape and fit: they have to fit over some medical equipment. So I will have to be careful to only do the inside of the mould with a brush or wet tissue, and leave the outside untouched and stable. Maybe I will try multiple light applications, rather than one heavy. We'll see what it gives. Concerning health, I have a chemically-safe fumes extraction cabinet in my lab, safety glasses, and various medical gloves, so this is not a problem, fortunately.

I use silica gel from car shops: these are used to dry car interiors in winter and have a high capacity. Can be re-used after drying in a microwave oven, or conventional oven at low temp (70°C or so). When the blue dot changes to pink, it is saturated with moisture and needs reheating. I store my filament (PLA, PET, NGEN) in big plastic boxes, with such a bag in each, to keep it fresh. But I do not have PVA; so I can't comment on that. I don't know if silica gel is powerful enough to dislodge water from already moist PVA, nylon or other filament, at room temp? Maybe you need to warm the filament (stay below softening temp!) in the presence of silica gel, to shake the moisture loose?

And which is the X-direction: is that front to back, or left to right? And where is the zero point? I can never remember these. Z-direction obviously is up and down.

Congratulations! This is a magnificient add-on: a relatively simple concept, but very versatile. This basic concept could also be useful for engraving (like nameplates), lasercutting, and similar things which require high accuracy but relatively light mechanical loads. A question: when drilling with very small drills (0.5mm), how is stability of the drill head? Does it wobble? When last doing electronics (>10 years ago), I found that this was the biggest problem when drilling holes manually: drills could easily break at the slightest wobble, sometimes due to not holding the board well enough, sometimes due to the slight wobble in the drill head (standard but worn out Dremel).

Why no tape: for me the seam lines where tape strokes meet, would hurt. Most of my models need to slide along each other, sort of custom calipers for medical use. Also, putting tape down without bubbles, is not easy. And we all have had bad experiences with tape or labels that we couldn't get off cleanly (usually not related to 3D-printing, but in general). Maybe that is why a lot of people are a bit reluctant to use tape? Although in itself there may be nothing wrong with the method. If your tape does not have any of these problems, feel free to post the brand and specifications and a few photos. We can always learn from it. Different materials may require different bonding methods. So it is good to have lots of bonding methods in our inventory. Once people have developed a reliable method that they like, they tend to stick to it, because it is predictable and they master it well. This may outweight the possible benefits of trying other methods. I mainly use my "salt method": this works well for my brands of PLA (Ultimaker and colorFabb), but not at all for ABS, and poorly for PET. So, for PLA, I just wipe the glass with a tissue moistened with salt water, and I gently keep wiping while it is drying into an almost invisible mist of salt. That is all. It takes maybe 10...15 seconds. Bonding of PLA is very good as long as the glass plate is hot (60°C). When cooled down, models come off by themself, without any effort at all. No cleaning required, no other messing, no need to take the glass out of the printer. Then I just wipe again with a tissue with salt water, and I can start the next print. If I would ever need to print PE or PP, probably I would go to tape too, since these are very difficult to bond to glass, using glue.

A question: how long do you let your models dry after smoothing with acetone? The goal is to smooth the layer lines of a PLA/PHA mould, for casting silicone models into it. For testing, I dipped a scrap print (colorFabb PLA/PHA, red) into acetone for a few seconds, and after more than an hour drying now, it is still flexible and still emits an acetone smell. So the acetone seems to penetrate far deeper than I expected. Only the thinnest parts are now getting white (just as seen in some of your photos) and seem to be drying.

I found that I cannot drill at all into PLA, nor tap a thread into it, nor use self-tapping screws. It just melts. However, manually filing and sanding is possible. And manually and genlty enlarging holes is also possible, for example to make an M4 bolt pass easily through a too narrow printed 4mm hole, which is actually only 3.5mm. I use a separate drill chuck for this: this gives good grip and allows careful drilling. But even this has to be done very slowly. For higher temp materials, it should be a lot easier, like in commercial plastics. Some materials sand easily into a smooth surface, but some get a sort of fibrous and rough surface. PLA, PET and NGEN tend to break at sharp corners where high stresses are present, like injection moulded plastics. But not along layer lines, in my experience: the fractures cut diagonally through the material, following the stresses. But layer lines may be an issue with some high-temp materials like nylon or ABS, with poorer layer bonding? So I guess lathing should be possible, if done at low speeds and with a small "bite" (I don't know the correct technical term for not taking away too much material at once). But for high accuracy, material creep afterwards due to environmental temperatures, and due to mechanical loads will always be a thing to consider: it may spoil accuracy later on. Just like in most plastics.

Thanks for describing the workflow. Do you use the heat gun only to bend things in shape, or out of shape, as desired? Or do you also use it to smoothen details and layer lines?