geert_2

I think it might be a good idea to google the BC817 specs, and test things out with a separate transistor and a test board. If it would go wrong, that is a lot cheaper than a new controller. The very limited specs I found (but I only did a quick search), do list a VCEsat of 700mV at 500mA, which is surprisingly high. I would have expected 300...400mV. And a Ptot of 250mW at 25°C ambient temp. So, at a load of 300mA you could still go over the max power rating in a closed environment with limited airflow and cooling (thus probably 50 to 60°C in summer). So you need to find specs that show all the power-rating curves and VCE-curves too, for given loads and temps. My practical experience with discrete transistors (long ago in the stone age, before SMD...) is that they don't like it when you get close to their maximums. Some even didn't like it when you only got half way.

ABS warps a lot and does not easily stick to a glass build plate. And its smell is toxic. There are materials that are as heat-resistant and strong, but that do warp far less, stick better, and are not toxic: PET, more and more polyesters, high-temp PLA, nylon, PC,... And there will be more developed in the next years. So more and more people seem to move away from ABS. An UM2 has the advantage that you can print all sorts of materials from all a lot of different manufacturers, as long as it is 2.85mm diameter, and not too flexible (as it is hard to push an elastic rubber band through a bowden tube). But every material (and every printer) has a learning curve, thus expect that.

In addition to what Labern said: I found that blobs, ringing effects around corners, calibration of the first layer ("elephant feet") and similar deformations also affect accuracy on an UM2. So they may exist also on an UM3 to some degree. Speed and temp have a visible big influence: print a 10mm x 10mm x 10mm test block and manually change speed or temp half way. You will clearly see the difference with the naked eye. If the layer height is 0.1mm, then there are already layer lines, thus ripples of about 0.05mm, in the Z-direction. Blobs and these layer lines generally cause a part to be 0.1mm wider than designed in my models, or holes to be 0.1mm narrower (for big holes) to 0.5mm narrower (for very small holes) than designed. So I guess post processing (sanding, grinding, smoothing by dissolving, painting, polishing, drilling out holes) will be required anyway, for good accuracy.

Another method that just crosses my mind: you could 3D-print the model, make a silicon mould of it, and then fill the mould with a fiber reinforced composite. Whatever fibers you like. Then you might get something that has the correct shape, but with physical properties closer to the real thing. Or first stuff the mould with fibers, and then pour a very thin liquid composite or polyurethane in it. These methods overcome the layer- and infill-problems of 3D-printing. (Note: concerning my idea of "steel wire" above: if you ever do that, make sure it can not cause amputations if some body part gets stuck in it...)

Wow, this is a brilliant idea, both the dropping in of supports, and of the pre-printed plate of peggyb. Concerning the use of release sprays: most dental release sprays are based on silicon oil. Some other release methods are water soluble, I think (but I am not sure) they are based on alginate dissolved in water? The problem is that the release method should not bond when trying to remove the support, but it should bond very well when printing, otherwise the next layer won't stick. Contradictory things. So, if you would want to drop in a support block, maybe covering the support with a layer of dried wood glue might work? Or something similar, which is water soluble? Then at least you could soak the model in water until that layer is dissolved? But of course, if there is enough room for the "peggyb method", that would be preferable. It is just so much cleaner, and allows for modeling that underside too.

For the bonding, have you tried the "salt method"? First, thoroughly clean the glass plate with warm water (without soap, this reduces bonding). Then wipe the plate with a tissue moistened with salt water. Gently keep wiping while it dries, so it leaves a very thin, almost invisible mist of salt stuck to the glass plate. For me this greatly increases bonding, and it works even for difficult models (e.g. thick 100% filled). I never use brim or raft for my models, the salt method works well enough. For a full description and photos, see the manual: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

Regardless of the material, try to print it as hot as reasonably possible, but without burning the material in the nozzle. Also try to minimise cooling as much as possible, but without the model sagging and deforming. And try to print rather slow, so the new layer has enough time to melt well with the previous layer. This will give the best layer bonding. This in addition to (depending on your printer) increasing the flow a little bit, so you have a little bit of overextrusion, as said above by the others. So that all tiny holes in the print are filled well. When I print in Cura and use 100% infill, my models are always quite solid, without visible gaps in the surface. Only in the corners of the extruded sausages (I don't know how to describe it correctly), there are very small gaps, the so-called "weld lines". Concerning material, how about PET? This is tougher than PLA, but not as flexible as nylon. And it prints reasonably well. If I were in your place, and my life would depend on it, I would add some steel wire somewhere as backup. Maybe keep the steel wire loose enough, so it does not influence the normal function of the part, but still good enough to keep things together if the part would fail. As a sort of safety belt. Obviously, first thoroughly test it on a bench, and hammer on it (simulate hitting a rock), before going downhill at 100km/h.

The problem with Creative Commons licences is that they basically only apply to creative artwork. It is a sort of copyright protection. So it has to be *art*. Copyright does not apply to technical stuff. Although a lot of people do attach a CC-licence to their technical designs. But this has very little legal value: it can only serve as an indication that your invention existed at that time. If I understood things well... The only way to protect technical inventions, is by a patent. And if you don't protect your own invention, then a competitor can. Then the cost is on you to prove that that patent is malicious, and that you were earlier, which can be extremely difficult and costly. So I do understand the need to protect your rights and inventions against abuse from malicious big competitors with a huge budget. But there is still the possibility that a future management might see things different than the current management. For example when the company would be taken over. Who would resist a multibillion dollar offer? I am not sure, but I think it is possible to write into a patent that its use will be free forever for everyone. Or for non-commercial purposes only, or so. Then I guess that clause would have legal value? But I am no lawyer... Anyway, it is something you might want to have a look at? This will always be a very difficult balance.

What I usually do with PLA filament is manually unwind a bit of filament and straighten it: I roll about 10cm of filament in the opposite direction around a 7cm skater wheel, thus bending it in the opposite direction, then the next 10cm, next 10cm, etc., until I have done about 1 or 2 meters, enough for the print to come. This takes only a few minutes. Then I wind it up again on the spool. But now it is sitting very loose. This gives far less friction in the bowden tube and nozzle, improves print quality and eliminates underextrusion. Maybe you could try that? I guess in your case it is the tight bending curve that causes too much friction in the nozzle itself. So with this technique there is very little bending, and the filament softens a bit by straightening it in this way (it gets micro-cracks). Don't do this with brittle filament like wood-filled, or metal filled.

In my experience, white PLA is often easier to sand than other colors. I am not sure why, maybe because it requires a lot of filler particles to get a really white and opaque look? ColorFabb filament is a bit more fibrous (if that is an existing word? - showing more fibres when sanding). On the other hand, ICE filament (distributed by Trideus in Belgium) is a bit more brittle and sands more into fine powder. When printed fresh, both have a satin gloss look, but after sanding of course, it gets as matte as you want. Disadvantage of white is that you see every spec of dirt on it. Especially when grinding it with a file or sandpaper that still has colored particles from previous colored filaments in it.

I haven't tried clear PLA yet. But for PET I found that printing slow helps best: 20mm/s for a somewhat clear result. Temperature plays a lesser role at that slow speed: just don't print too hot (bubbles, fog), and not too cold (doesn't flow well into corners and thus leaves unfilled areas). A little bit of overextrusion also helps to push the melt into all holes, and to avoid voids. Try 105 or 110%. When the model is only a few layers thick, underlaying text shines through nicely. But the diagonal infill lines stay visible. Thicker models always get a "frosted glass" appearance. So you can't use it as a lens.

I just realised that my text above could be interpreted in different ways. So I would like to make clear that up till now we are very happy with our two Ultimaker printers, with the service, and with the openness on the forum and in the documentation (e.g. the technical drawings). According to my collegues, they don't find such openness in some companies with huge +100.000 euro 3D-printers. It's just that I want to make sure that your management is fully aware of the importance of this currently good and open atmosphere, and that they need to make sure it stays so in the future, for all the reasons stated above. While at the same time protecting everyone (both the company, dealers and honest customers) from malicious threaths. That can be a difficult balance. I do indeed see the risk of problems with home-made filament or cheap garbage, used by accident, or used maliciously to sue a competitor out of business. So, concerning wording, something like the next phrase might be better for the warranty? "... Ultimaker filaments or filaments from reputed manufacturers who have thoroughly tested their filaments on Ultimaker printers. For example but not limited to: ..." This will encourage non-listed manufacturers to thoroughly test their products too, and to publish the data and recommended settings and practices for use of their filaments on UM-printers. The more versatile the UM-printers are, the more documentations there is, and the more filament there is available for it, the more reasons people have to buy an Ultimaker. Which is good for everyone.

One more comment: I do understand the Ultimaker fear that some people may abuse the printer or materials in an attempt to sue and to get a huge damage payments. Like those idiots who wash their pets (pet animals, not PET plastic) in a dish-washer, and then dry them in the microwave. And next, of course, they sue the manufacturers and demand 100 billion dollar. Criminal lawyers working on a 25% margin of these wins, and idiot judges, are often all too happy to join in. This tendency is now coming to Europe too. So I do understand that any company should legally avoid that huge risk. But the wordings in any warnings and warranty should be more in the style of: "This device should not be used by people with reduced mental or physical abilities, or by people who do not understand the technology and the inherent risks." Ultimaker already uses some of these wordings. The disclaimers and warranty should leave any qualified user, technician or artist the freedom to use any filament (only excluding blocked nozzles due to wrong filament from warranty), and the freedom to experiment with the printer and to improve it. As long as good common sense, good care, and good technical practices are used. It should state that open development and testing are encouraged, not discouraged. It may take a bit of thinking and creativity to find a well balanced wording that is legally valid, but I believe this is the direction to go.

ColorFabb PLA is a PLA/PHA blend (the opaque filaments) and is a bit more flexible than pure PLA without additives. But still harder than ABS and PET. So it is by no means "flexible". I don't know about the real "flexible PLA", I have never used that (so I also don't know if that stays flexible, or if it degenerates and gets brittle like standard PLA). PET sticks way better to the glass build plate than nylon, so that may be a reasonable choice indeed.

Just a few things that came to my mind: - What about the heat generated? Is that contained in the engines, or does it spread out to the 3d-printed housing? If any heat is spread out, obviously PLA would be a bad choise. Maybe all plastics. Especially on a vertical take-off from the ground there is a possibility that the wings would melt and catch fire. - What about flutter of the wings? I have heard this is a very common problem in airplanes, even at relatively low speeds (a few 100 kt)? It already is a problem when driving with open windows in a car at 100km/h. When using flexible materials, this could be a problem too, unless you would use very thick wings with internal stiffening structures. But that would greatly increase the frontal area and thus drag. - How do you plan to stear the rocket? Does it have active controls and do the vanes move like in military missiles? If not, the tiniest imbalance in the engines or wings could bring it out of course. Concerning survivability, I think not much is going to survive a drop from 1500 feet. Unless you attach a sort of parachute or balloon that automatically deploys. Or speed brakes that fold out, activated by springs, or so. But that adds complexity and weight. Most model airplanes are destroyed on a crash even at low speeds and heights. Personally, I think I would go for thin aluminum plates to make the wings. Or if they were steerable, maybe use a sort of "gratings" instead of wings, like in modern military missiles. I am not sure why they use these? Maybe better control? More compact? Easier to fold and transport? Less drag? But you see it more and more. Anyway, I think this is an interesting experiment. I am curious about the final results.

In addition to Labern's suggestions: print slow and warm enough, and minimise cooling, so the layers do bond as well as possible. I have made things like that in colorFabb PLA, and they worked too, if not used frequently. And only for a few months, until the PLA got harder and more brittle, after which they broke. So it may work in colorFabb PLA for a prototype to show off, or for a single snap-on fixing that is never to be released again. But not as long term solution for frequent use. PET is much more flexible, but may also break along layer lines. You will have to try. Editing your design to increase the traject is also an option, like adding some U-turns or sine-waves. But in my experience PLA will always get brittle over time (a year or so).

For us "open source" and an "open filament" system (=standard spools + multiple manufacturers) were absolute requirements. This among other breakpoints like: sufficient size for our models (of course), good printing quality, within budget, good service (both concerning spareparts, and help, thus including the forum),... 3D-printing is a very volatile market at this time, with unmatured products in full development, where the users have to understand the technology and actively work on the technical inside aspects. It is like cars around 1900: you couldn't drive a car if you didn't understand the technology, and you had to be able to repair it yourself. We do not want to be stuck with devices for which there are no more spare parts, or no suitable filaments, or no technical documentation. So we need multiple suppliers, and we need sufficient technical info for maintenance and upgrade work. And we also need enough freedom (both legally and practically) to improve the device and to experiment with different filaments, so that we can select the best quality filament for our needs. It is clear that a lot of new filaments will be developed. We also need enough different colors, which may not be available from only one manufacturer. We, users, are all technicians, or artists with good technical understanding, and we are all aware of the need for backup parts and sources. This is why a Zortrax and a few other printers of similar size and quality were out of question for us: they were closed-source, and/or "closed-filament". Of course there may be a few UM-clones due to the openness. But they don't hurt sales, since those buyers probably wouldn't have bought an UM anyway. If they wanted the real thing, they would have bought a real UM. And if there was no UM-clone, they would have bought another cheap clone from an unreliable source... So I think that ultiarjan is absolutely right that these values and openness should be preserved at all cost. They are the basis of the Ultimaker quality and sales success. Without these values, we would not have bought our two UM2. And yes, openness includes freedom of speech. If you can not analyse and discuss problems, you can not solve them. A forum that only allows praising the leader, like in some Middle-East countries, would soon be no forum at all anymore. Of course I also do understand Ultimaker's point that they do not want to be liable for destructive idiocies done by unqualified users. Such as "repairing" fragile things with a hammer, and then complaining that they are bent and don't fit anymore... So there has to be a good balance, but it should definitely be towards the openness. (Besides: I have no idea who runs the Ultimaker company, so this is not personal. But in general I think that companies should be run by enthousiasts and developers; thus by the very people who create and use their own products. Not by legalese people who have no feeling with the product or the market. Of course, commercial and legalese people are very valuable, but only as advisors in lower positions, not at the top. They should be servants, not the desicion makers. This applies to all companies, including car companies, computer companies, whatever... Bookkeepers and lawyers may be required to prevent a company from collapsing, but whenever they take over, the products go down and then the company goes down too.) My viewpoint, so feel free to see things differently.

In addition to searching the forums and reviews, I would suggest that you design a few small test models that include all your specific questions or items. Then go to a club, 3d-hub, or dealer who has those printer, and have the models printed while attending it. Have the owner explain stuff while watching the process. Of course you will need to pay for the materials and time this takes, but it will be a good investment. You can see if it matches your expectations or needs.

If you buy cyano-acrylate glue for glueing parts together, try to find one with an activator. That is a separate tube with felt tip (like a marker pen) with which you need to wipe the plastic, then let the "ink" dry for a minute, and then glue as usual. This activator improves bonding of plastics. In Belgium you find them in the local supermarkets like Delhaize (among the school things), or in Brico-shops.

My first idea too was to try nylon. But maybe PET or PETG would have a better layer bonding, and they are still relatively flexible? You don't easily break a PET bottle when dropping it. And PET is easier to print and has a better bed adhesion. I think it will also depend a lot on the shape, whether that is able to absorb any shocks, and to distribute the load, without high local concentrations of forces. Try smashing the part into a concrete wall to test it. Would be interesting if you could post a few photos of the results.

I use Pale Moon 26.0.0 as browser. This is a Firefox-derivation that kept the classic menu bar and status bar. No add-ons. Operating system: Windows 7 professional, English. After logging-in, I usually open several tabs by middle clicking on interesting topic titles in the forum list. And then I open each tab and read the topic. Before I begin reading a topic or typing a reply, I reload the current tab to avoid this auto-logout (which is a bit inconvenient by itself). But sometimes that obviously doesn't work. I haven't timed how long it takes before I get logged-out. It feels like 20 minutes, but time goes fast when doing things, so it could be more. I don't have remind-me enabled. If I remember well, there was another user recently who also complained about the auto-logout, so it seems I am not the only one.

You could try DesignSpark Mechanical as modeler: it is freeware (requires registration) from RS Components, and is actually a limited version of SpaceClaim. Its method of direct modeling by pulling on things is the same as in SketchUp, so the learning curve is relatively low. And with a bit of searching, you can find several good instruction videos.

Hoi Sander & co, Could you increase the auto-logout time of the forum? Now it regularly happens that I get logged-out and lose a reply that I have been typing. Especially when in-between I had to look up something, or had to service my printer for a few minutes. I would suggest a logout-time of minimum 1 hour. Or better: make it user-adjustable if possible, with settings between 15 minutes minimum and 12 hours maximum, or so? This would make it much more comfortable.

Thanks for trying to help, but my problem is not the letters! It's the visible lines on the surface. Those I need to get rid of. Regards, Uwe Yes, I had understood that. You can not get totally rid of these lines, but hollow characters inside a transparant model do reduce the ugly effect. Instead of the weird outline-patterns that you have now, you would get nice straight diagonal lines, which are hardly visible. The fact that the whole model also gets a sort of frosted glass look, or a carbon or textile look when using a transparant or translucent material, even masks them more. And as mentioned before by others, your nozzle needs to sit close to the build plate, so the first layer is squeezed hard. Printing slow also helps, then the filament has more time to melt and to flow into the corners. Even in high pressure injection moulding you see these "weld lines" where flows meet each other. It won't get better than that, unless you chemically smooth, or sand and paint the model.

Like ultiarjan said: no brim. I always print this sort of models without brim, raft, "ears", or whatever. Even for solid models. You could use dilluted wood glue for most materials, or my "salt method" (=gently wipe the glass plate with a tissue moistend with salt water) for PLA. Reducing fan speed or switching them off may also help for some materials (eg. PET).