geert_2

I have both Ultimaker2 printers sitting in a fume extraction cabinet, since I have that anyway in my laboratory, so I can as well use it... However, most of the time, I leave the extraction switched off, so there is only very little air flow through the narrow gaps in the shut-off valve (it's one central vacuum pump for all laboratories in the building, and shut-off valves at each cabinet). This does not cause any smell problems for printing PLA, PET and NGEN. But for ABS I really need to switch the extraction on and close the glass lid of the cabinet. So, don't let this concern stop you from printing. Just don't print ABS. Probably the best thing you could do, is find someone in your environment who has a 3D-printer (school, 3D-hub, reseller, hobby club, fab lab,...), and go have a look while they are printing.

Sometimes an "unflat glass plate" might actually be an unflat aluminum plate. Recently when removing the glass I noticed that at the underside the glass had a deep circular indentation of 3mm diameter in one corner, scratched into it. It looked a bit similar to serial numbers stamped into a metal plate. So I wondered what caused that, and I inspected the aluminum plate. And indeed, one of the 3mm holes had a little brim. So I used a drill to deburr it, and now leveling the glass is easier in that area. This does not mean that there can not be unflat glass plates, it only means that there can be other causes too. You might want to have a look at that. Then rotate your glass 180° (back to front), and see what happens: if it was the glass, the problem should now be in the front? But instead of 3 screws to adjust the build plate, I would prefer 4 screws: one in every corner. This would make it easier to level out such problems, by slightly distorting the aluminum and glass plates, to correct any existing distortions. This can not be done with 3 screws. Maybe the Ultimaker developers could give this a try?

Or you could use a fume-extraction system similar to the one in a kitchen, and make a pipe to the outside. Or connect extra piping to existing air ventilation canals (make sure they go to the outside, not recirculating). ABS really stinks and is toxic, but most other materials like PLA, PET, NGEN, etc... only have a light smell, non-toxic as far as I know. However, in a small room, even that smell might spread everywhere and it might become very annoying to other people in the room, who often have less tolerance for what they consider weird smells, unrelated to their own activities. So, for your own peace of mind, probably some ventilation is desirable. Most illnesses are stress-related, or triggered or worsened by stress (especially astma, allergies, eczema, lung problems, nose problems, stomach and intestine problems, heart problems,...). So if someone would be worrying about something, that alone might trigger those health problems he or she fears... There has been some research on particle emissions from 3D-printing (I have read it, but I don't remember where, so you will need to google). But according to that study, these emissions were far less than emissions by smoking, pet animals, candles, and cooking. Your body (dead cells) and clothing (fibres) are going to give off a lot more particles.

In some UPS you can easily replace the batteries, just like in your car. Might be worth checking out if you can borrow some for testing. But in my UPS 15 years ago, that didn't work well: they would not fully recharge, probably due to a little bit different voltages, even if only a few tenths of a volt? There could of course be other problems too in the UPS, such as damaged capacitors and worn-out electronics. I do not know the official formula to calculate required UPS-capacity, but based on common sense, I would follow this logic, less or more: count the total power consumption of equipment to protect, in Watt. Define the time it has to survive a power-outage in hours. Multiply this by a factor 3 to account for inductive loads (power factor), battery wear after a couple of years, occasional power surges in the load (like when switching on something), and some safety margin. That should give you a rough idea. And then you might get shocked by the required capacity, size and cost, and lower your standards a bit...

I have wondered if washing off the PVA would go faster if you use an aquarium with good temperature control (~35...40°C), and a powerfull circulation pump? That should remove PVA-saturated water from the model. The only risk is that the pump would get clogged by chunks of PVA, so the entry should be located carefully, and maybe protected with a big sift, like a big kitchen sift? Such a setup uses standard components and doesn't cost too much, compared to 3D-printing equipment. Beautiful ship by the way.

Then you can be sure the batteries are dead... Usually these are lead-acid batteries, similar to those in a car. Except that they are usually of a sealed gell-type that does not need refill, instead of the liquid types in the car. At least, as far as I have seen (I had an UPS for a server). In a car the batteries usually live 6 to 8 years. But near the end they lose capacity anyway, like all rechargable batteries. So you should replace them at least about every 5 years. It is a good idea to regularly test them, every six months or so: switch off the power (pull the plug) and look how long they provide power under realistic load conditions. Write down the survival time and load conditions, so you can exactly repeat the test. Near the end of life, time will go down very fast. Then they basically provide only protection against spikes and too high or too low voltages. But even with UPS, spikes could still kill a print. I have had that in winter last year, when the air was so dry that I would cause huge "lightning- sparks" whenever I touched anything, sometimes up to 3 cm. So I had painfull arms all day. When touching the UM2 frame, and causing a spark, it would lock up. Sometimes it would continue printing, but I could no longer use the buttons, sometimes it stopped. Yes, and I agree: a faillure-recovery would be a good thing. With the ability to continue on the same print, if the outage was not too long (so things didn't cool down and the model didn't come off), or the ability to print only the remaining part in a new print. Selectable per your preferences.

In your next photos, could you show what basic shape(s) you started from for each model, next to the final result? Are it simple lines, curves, leafs, circles, or already more complex basic shapes? Or combinations? Could be interesting.

Static could raise or lower voltages on high impedance circuits by induction alone, to below or above switching levels. Without damage. But of course, as soon as there is a discharge spark, damage may occur.

I think, if your printer has a heated glass bed, and you print PLA, you really should consider printing directly on the glass. Heat the bed to 60°C (for a start, adjust in steps of 5°C as required), and use one of the many available bonding aids as required by your models: 10% wood glue dilluted in water (=gr5's method), hairspray (=neotko's method), 3DLac, salt method (=my prefered method), glue stick (wipe with wet cloth afterwards), etc... Give them a honest try and chose the one that works best, and/or that you like most. On the Ultimakers, I think no one prints on rafts anymore, unless for a very good reason (e.g. model has no flat bottom). I believe rafts were a good solution from the past before heated glass beds existed, when only ABS was available, and before the current good bonding methods were developed. But now they are obsolete for most printers, materials and models, apart from a few exceptions.

This is an interesting test. Maybe you could replace the X by a couple of vertical lines of same depth, so they show up in every layer? On my UM2 (non-plus) it generally looks like the areas from 50mm/s and slower in your test. But I only print between 25mm/s (for fine details) and 50mm/s (general prints). Ringing is more visible than the widening of corners.

The design rules for injection moulding are quite strickt, otherwise the part can not be mass produced at all. Startup costs (=mainly design adaptation and mould production) are very high: expect 10000s of euros for high volume production moulds. The rules for silicone moulds and resin casting are way more flexible, but then of course production volume is very low and production cost per item is much higher. So this is suitable for art reproductions, hobby, low scale tests, but not for cheap toys that have to compete with commercial stuff. I just remember that there are companies that do deliver quick injection moulding services, and also other services like 3D-printing and 3D-editing, for relatively low volumes (100...1000...10000 parts). For example "ProtoLabs" in the UK. They have good technical info and guidelines on their website. I would suggest you study these. In Belgium we have the very well known companies "Materialise" (big size 3D-printing, mostly plastics), "Melotte" (also metals), and several big others which I don't remember. Might be a good idea to study their services too. If your design is a technical innovation (=has to be new, innovative, technology, producable) then you could apply for a patent. But expect that to cost 10000s of euros. Copyright does not apply to technological inventions and methods. If your design is a piece of art or design, it cannot be patented, but it can be registered and protected as a model. This in addition to automatically applying copyrights (in Europe, maybe different elsewhere). If you would speak Dutch or French, and live in Europe, have a look at the website of the Belgian government concerning protection of intellectual properties: https://economie.fgov.be/nl/themas/intellectuele-eigendom This gives excellent info in simple language. Or find similar sites from other governments, institutions or companies.

Mass production with injection moulding is quite complex. There are lots of things to consider, such as wall thickness, bending radius, no sudden changes in part thickness, sink-defects, avoiding voids, drafted side walls to be able to release the model from the mould, no undercuts, carefull placement of seam-lines, carefull consideration of mounting options and screw holes, and so on. I would suggest you google for words like: injection moulding manual or: how to design for injection moulding or similar terms. Search for PDF documents. Big plastic manufacturing companies like Bayer, BASF and lots of others have published good manuals online on designing for injection moulding. And then probably you will need to redesign your part in a vectorbased editor, incorporating all these requirements from the very beginning. Think of editors like: DesignSpark Mechanical (freeware, but its STEP- and IGES export is to be payed), Onshape, SolidWorks, SpaceClaim, Rhino, Form-Z, and lots of others that I can't immediately remember. (Don't use SketchUp.) If you only want to mass produce relatively few models, on a hobby-scale or for artwork, you could 3D-print one item, post-process that until really perfect (sand, polish,...), and then make silicone moulds from it. And then cast in resin, gypsum, concrete,... Or you could print multiple models, and use them for casting in metal. Some jewelry designers here do that. Search on Youtube for good videos on mould making and casting, there are lots of really good ones available (and lots of crap too of course, so you need to filter and use common sense). It is all very interesting, but do expect quite a learning curve.

Unless maybe you would guide the tube into a smooth U-turn with an empty spool or something similar? But this would require a very long bowden tube, which is also not good... I was considering PolyBoxes, but now that I see how big they are, almost the size of an UM itself, I don't think I have enough room in my fume extraction cabinet.

I don't know Polymax, so I can't comment on that. But maybe you could also look into the opposite? Could it be that it is electrostatically charged too much, and that this causes malfunctioning? Most plastics charge up easily. This could especially be a problem if the board would contain CMOS or similar stuff, with open pins, or circuits with very high input resistances? If dust and little pieces of plastic debris easily get stuck on the model, it is probably charged up. Apart from other possible reasons of course, like a broken soldering, broken copper trace, or wire?

What I always do before inserting any filament, is (1) manually straighten it, and (2) cut off the end with a sharp tool, in an angle of ca. 60°, so it gets a very sharp point. This feeds well.

Maybe you at Ultimaker could ask a few samples to try them in your test lab? Might benefit both parties?

I think these strings are what I call a sort of "landing strings". I don't know if there is an official term for it. What I mean is that while traveling through air, the nozzle may leak a bit, and when encountering the next wall, that little drop is deposited on the side of that wall. On the next layer, the leaking drop is deposited on the previous drop. And then on that drop. And so on, causing weird upwards and outwards going strings, like insect antennas. It is a guess, but it is the only thing I can think of. I do not have a good solution, but any solution (if this is indeed the problem!!!) should go in the direction of causing less leaks, or less movements through air, I think. Maybe print cooler and slower (=less pressure in nozzle), or travel faster through air, or travel over land instead of through air where possible, or more cooling fan, or changes in retraction settings (or maybe the printer doesn't even do any retractions when printing supports, which is why it leaks),...? Try watching carefully what happens while printing. I do notice something similar when printing PET, which is more rubbery than PLA: so when printing overhangs or when traveling through air, it does not cause thin strings or hairs, but a blob under the nozzle that is deposited on the next wall or next obstruction that the nozzle meets in its travel.

If you are printing PLA, you could try my "salt method". First, thoroughly clean the glass: you can use isopropyl alcolhol or whatever to remove oils and other dirt. And then clean again several times with pure warm water only (no soap, no alcohols, no thinners anymore, because soap reduces bonding and cheap alcohols or thinners might contain traces of oils, which also reduce bonding). Then wipe the glass plate with a tissue moistened with salt water. Gently keep wiping while it dries into a thin, almost invisible mist of salt. This gives a good bonding while hot, but absolutely no bonding when cold. So the models come off automatically. For me this has consistently worked well for 2 years now. No need to take the glass out of the printer, just wipe again with a tissue moistened with salt water, and you are ready for the next print. Glass bed should be around 60°C. If cooler (e.g. 50°C), bonding is not good enough and the models may pop off suddenly in mid print. If hotter (e.g. 70°C) the models stay too soft, and may gradually peel off. So, (for me) the bed has to be 60°C +- 5°C. But this salt method only works for PLA: it works excellent for Ultimaker and colorFabb PLA, and a bit less for ICE PLA. As always, with any new bonding method, stay around and watch what happens during the first prints. For the full manual and pictures of the results, see my page (scroll down a bit): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

I am from Belgium, Flanders, but we also speak Dutch here.

@skc5741: The above "tiny real model" is a 10mm x 10mm x 10mm block, plus extension of 3mm x 3mm x 5mm on top. I will add these dimensions to the drawing for clarity. You are totally right that the dummy is waste, but otherwise the model can't be printed well (unless I would print 4 at once, which is not always desirable). So it wastes less than discarding too many real but malformed models. That is the idea. Dummies are only required for tiny models. Once the model gets a bit larger, lets say 20mm x 20mm, this is no longer necessary usually, because each layer has enough cooling time anyway. The dummy can have any shape you wish, but subtracting is sometimes easiest to hollow it out in areas where no extra cooling is needed, to reduce waste. Here I used a simplified dummy (model height = ca. 10mm) to provide enough cooling for the top part. It works good enough. This is part of a real design, not just a demo. When printing solids or highly filled models (70...100% filled), like in my case, cooling is more critical than when printing hollow models, since solids require to dissipate much more heat through less surface area. Concerning your bisshop: I don't know how big it is, but if its diameter would be 10mm, then subtracting it from a rod would be the easiest way to create a dummy. If its diameter would be 30mm, it could probably be printed without any dummy, by printing as cool as possible (ca. 185...190°C for PLA), and slow, and with a gentle desktop fan in front of the printer. But I think it looks quite good the way it is now, and I am not sure it would get much better with a dummy next to it. Even with a dummy, you might still have little defects where the nozzle takes off and lands. @peggyb: if printing a draft shield takes enough time to let the rest cool, it could work, of course. Any method that provides enough cooling, enough equalisation of layer printing time (if required for quality), and that keeps the material flow constant enough (again, if required for quality), is a valid solution. You could also add a desktop fan in front of the model, or carefully blow cold compressed air onto tiny areas for extra cooling (I have used these occasionally), although these methods might have other side effects. Just don't let the printer print "spaghetti" somewhere else in mid-air while cooling your model, because the spaghetti would get dragged around and into your model, and it would accumulate on the nozzle (I tried it, out of curiosity). This is another reason why I add a baseplate under hollow areas in the dummy, to prevent spaghetti from the overhangs getting dragged around. I would suggest: add this concept to your inventory of tricks, but don't throw out any other proven methods.

You could design the supports manually, as part of the design process, and disable the automatic generation in the slicer. I often do this, due to the specific requirements of my models. Then you can design-in all sorts of methods to make removal easier, for example extensions so you can use pliers, or holes so you can insert hooks to pull. Or you can add ribs to reduce space and improve quality, etc. See the examples below (these are all very small models, order of magnitude 10mm high). Try this on a couple of small samples first, untill you have it perfect, before wasting a 5 day print.

Maybe it is also an option to print the disks and column separately, and glue them together? This might give a better surface quality and go faster than printing with support. In that case, design some guidance into the parts, so they line-up correctly.

This question comes up often enough to draw a quick demo in DesignSpark Mechanical, just to show the concept. Much easier to explain, and takes less time than typing. So, here it is: This is how I do it. But of course, depending on the models, dimensions, and materials, all sorts of derivations and approximations might work equally well. The dummy cooling towers are waste, or could be re-used as "hotels" in Monopoly or other games.

Obviously, your pyramid has this problem: This is due to not enough layer cooling time, because the hot nozzle is printing on a tiny area only, and it keeps radiating lots of heat. So the model can't cool down and can't solidify. These models were designed just to test that. The left models are printed separately, the right ones are printed cone+tower together. Dimensions (numbers) are centimeters: the cones are 2cm high and wide. The tower is 1cm x 1cm, which obviously is not yet enough: the edges still curl up in the white one, as soon as it has finished printing the cone (so the nozzle is full-time on the tower). You can reduce this problem - but not eliminate it - by printing a dummy tower next to your real model, as shown in the image. Make sure the tower has a big enough base plate or brim, so it does not fall over, and it is big enough to provide sufficient cooling time for your real model. You could partially hollow-out the dummy tower, to waste less material, and to keep the printing time per layer about the same. So that the cooling time per layer is the same for every layer: this gives a better printing quality than wildly varying cooling times per layer. Also, printing cooler and slower reduces the problem, because then it has to dissipate less heat, and heat radiation is less. For your catapult, I guess (but it is a guess!) that the problem is a combination of (1) not enough cooling time and (2) too steep overhangs. Thus I would suggest you try: print slow, print cool, print multiple models at once like kmanstudios said, or print a dummy tower next to it. and/or design a support column under it, which you cut off later. This will require some trial and error, but then you will learn a lot. Keep watching while printing, so you see what happens.

Depending on the amount you need, you might also consider casting them. First design and print an original, or a mould, and then cast as much as you need. Or a combination of both 3D-printing and casting. When casting, you could use other materials too like gypsum or concrete, which can be a bit porous for water and air. Depending on the application, this might be an advantage (allows irrigation, or removal of excess water), or a disadvantage (might leak). Casting is a very educative process too for kids and teenagers, most like it. (Search on Youtube for lots of good mould making and casting videos.) If it is a prototype system anyway, you might try as much different materails as possible, 3D-printed and casted, to see which works best.