geert_2

Before printing hundreds, print one and carefully test if it really fits and closes well. Cover the front opening (filter opening) with your flat hand, and try breathing: it should be *absolutely impossible* to inhale, and your body should immediately go into panick-mode. However, if you can still inhale with the front opening closed, it leaks sideways, and would not protect well and cause a false sense of security. I have seen too many dust masks (like for painting) that did not work at all. The real gas masks we used in the chemical industry were a lot bigger, and they had to be from very flexible rubber to close all gaps around mouth, nose, and chin. Yes, the masks you see in war- and horror films, with huge filters and creepy glasses. :-) So I do have some doubt about the effectiveness of anything that is less soft and less big than those old real gas masks...

I have seen them on Youtube, maybe your best option is to search for reviews there? Could be on the CNC kitchen channel, or Makersmuse? Or was it that Bernaculus Nerdgasm guy, or something like that? I don't remember...

I would say: make the holes slightly smaller, and drill them out manually with a drill that is still 0.1 or 0.2mm smaller than the desired size. Maybe you could also design a simple 3D-printed retention clip into the case? Try various dimensions on a small test piece first.

For my UM2, I made a nozzle scraper from steel spring wire, after an idea from another user. The very hard inox wire that is also used in dental appliances. In 99% of cases, this works perfectly. Since it is a spring, it moves away when the nozzle moves over it. It doesn't cause any damage.

No, not worth it for me.

I also drill out the holes, with this tool. It only costs a couple of euro/dollar in any tool shop. This gives good feeling and control. Do it manually, not electrically, otherwise everything melts (don't ask how I know).

Hope you can read English? To let the model cool and solidify, the head needs to be moved away indeed. But while sitting elsewhere, it should keep printing at the same flow rate, so the material in the nozzle keeps the same temperture and viscosity. So I often print a dummy model next to the real one, that has a complementary shape. So the total printing time per layer is always the same. See the examples below: first the concept, then two parts of real models.

These are nice models, especially the highest one. Here this is called a "castle-villa" ("kasteelvilla" in Dutch). One question: why do most architects and urban designers use white as color for models? Why not more natural colors in various shades like sandstone, warm grey, etc.?

You should make one in which you can insert your smartphone, so it can actually be used. With a slider system. That would be really cool. :-) It might require a bit adaptation of dimensions, but as long as it is immediately recognisable...

The one I saw on photo, didn't have any attachments. The tube was cut to the correct length. And then by friction from unwinding, it was pulled towards the feeder, sort of, so it closed the gap. But I don't remember from who this was.

You could design a sort of venting hole from the top of the magnet, to the hole in which you pour-in the resin. Or better two holes: one filling-canal, and one venting canal. Similar to those used in metal casting. However, if the magnet is totally surrounded by plastic, and that is impregnated with resin, I don't see how the magnet could be damaged. It might rattle a bit, but metal is harder than plastic. Another approach might be if you design a sort of springs or vanes into the hole. So, when you push-in the magnet, these vanes keep it centered and gently clamp it. Sort of snap-fit mechanism. That might be easiest? Try this fitting on small test pieces first, before doing a big model.

In daily life, I don't clean it. Well, I only clean it less than once a year. First wipe with alcohol, and then thoroughly wash again with pure handwarm tap water only. Then wipe dry with a paper towel. So, in daily life, I don't clean it for 99.7% of the time. I don't use soaps as they reduce bonding. Also, cheap industrial solvents might leave residues that reduce bonding: there is a lot of variation. I tried these things in the very beginning, but now I don't use them anymore. They didn't work for me. But before each print, I do wipe the glass with a tissue moistened with salt water, and I gently keep wiping until this dries into a thin almost invisible layer of salt stuck to the glass. For PLA, this greatly improves bonding compared to printing on bare glass. And after completion and cooling down, the models come off by themself. This "salt method" works excellent for my typical long flat models. But it is not optimal for narrow high models like lantern poles and towers: these might get knocked over. I also use this for printing PET, although in this case it does not improve bonding. But it makes removal easier after cooling down, without chipping the glass. PET sticks well enough from itself. It does not improve bonding for other materials like ABS (which has poor bonding by itself), so it is not suitable for this. This inverted pyramid is the limit of what can be printed with the salt method: the edges do warp a bit due to the huge overhangs. Most of the time the model can be completed, with edges lifting, but occasionally it comes off and produces spaghetti. Stay with the printer when doing tests like this, or when printing narrow high models like poles and towers.

Your image didn't come through. Normally, in Windows you can just drag and drop an image from the Explorer into this editing box. Or copy and paste from image viewers like IrfanView, or from Windows Clipping Tool. Could you try that again?

If the automatic functions don't work well for your particular situation, you can design your own supports in CAD, so they become part of the model. Then you can implement anything you want: holes to insert hooks and pliers for better removal, gaps, ribs, hanging supports,... I only have single nozzle printers (UM2), so my supports are in the same material as the model. But for dual nozzle machines, you could design for separate materials. A few concepts I have used or considered in the past:

Making a fine print from the original gcode, and then sand, smooth and paint that. And then 3D-scan and digitize that 3D-model? Would that be an option? Then at least you have a nice original to start from. But even when people leave a company in bad terms, you might be able to get them to cooperate, if you were not personally involved in the conflict too much. Most people can make that distinction. Also, doing a thorough search on STL, OBJ, and whatever else files on servers and local computers might give some results. Do this from a bootable stick, so you get around Windows read/write permissions. Temporary directories often contain copies of old files. Undelete-software might also help. On most servers there is a "shadow copy" service running: deleted files can be recovered for up to several months or years, depending on the setup. This is because most data-loss comes from accidental deleting the wrong directory of file, and people not finding out before they are some weeks further. So, contact your admin and ask about this. We here can recover those deleted files ourself, by making hidden files and directories visible, and then go to the shadow-copy directories.

You can always design (one of) them in CAD, and consider it part of your model. But for printing, using both has no practical meaning as far as I know.

These too thin horizontal areas are underextrusion. This basically means that the printer can not deliver enough filament, for whatever reason: too much friction in the feeding traject, (partially) blocked nozzle, dirty feeder, filament windings under other windings on spool, worn-out teflon coupler, non-working little cooling fan for nozzle, nozzle too cold, flow too high, worn-out or incorrectly installed bowden tube,... There is a video and extensive list of possible causes somewhere on this forum, but I don't know exactly where. Maybe you can find it? The vertical lines, is that infill shining through? Or is that part of the design? Looks like they too shouldn't be there? But that must have a different reason.

I do not have a dual nozzle printer, so no personal experience. But I do vaguely remember that supports might be printed at double layer-height, or beginning from a thicker minimum, or something like that, since it does not melt so nice as PLA. So, if the PVA would print at 0.2mm layers, and PLA at 0.1, you might get this effect. Could be an explanation. But I am not sure, it is just a vague memory, you need to check. Maybe you could search for this and find more info?

If you bend PLA and then keep it bent under stress, then microcracks may grow, and they may keep growing until it snaps. This can happen when it is sitting in the feeding traject for a longer time. So it is best to unload it after printing, and store it under no load. As it gets older, it can become harder and brittle too. Moisture also degrades PLA, it breaks down the molecules, so that could also speed up the brittling-effect. But I haven't had it snapping while printing or while in the feeding traject yet. So, in your case, it could be a bad batch or bad brand too. Wood, stone, carbon, or metal-filled filaments are also more brittle. These are microcracks after I bent the filament in the opposite direction for a few seconds, and then released it again (=removed stress), to make bending radius less tight and reduce friction in the bowden tube and nozzle. Idem, seen through a microscope These are cracks growing in a keychain carabiner hook, due to repeatedly opening the hook. A few times more, and it would totally snap.

Yes, you should definitely print cooler. On my UM2 the standard for PLA is 210°C, but for tiny models printed slowly and in thin layers, I can go down to 190°C. On a test piece, while printing, try going down manually in steps of 5°C, and see what happens. At some point it will start to underextrude severely. Then go up a little bit until it flows well again. Then try that value for a whole tiny model.

When selecting silent fans, keep in mind that they usually have a *much lower* flow rate and cooling capacity. So you might need to provide a lot more or a lot bigger fans. And indeed, adding rubber dampers, mats, or washers between fan and housing, can greatly reduce transmitted noise and resonant vibrations. Also adding a sound-absorbing mat on the opposite sides of a printer (or room) might help: this does not eliminate the source of course, but it damps the reflections and resonance. Similar to the covering of walls in a sound studio, but then in miniature. If your printer design allows for this, of course, it might not fit everywhere (I don't know the CR10).

And what is the effect if you make the wall thicker? (Or thinner, to exaggerate the effect?) I also have noticed that any sudden changes in layer-area to print, or in wall thickness, do show up as horizontal lines. Try this on a small cut-out of your model, at the place where the problem appears. So you don't waste too much time and material.

Like gr5 said: this is a cooling problem: the hot nozzle keeps sitting on same spot on top of the model, so it can't cool down and solidify. This is a limitation of this printing-concept. I often print small models that have this problem. Then I print multiple at once, so one can cool while the other is printing. Or I print a dummy tower next to the real model. Ideally, the dummy should have the complementary surface area as the real model, so that the total area per layer is constant. Brutal changes in layer area show up as horizontal lines on small models. Printing slow and cool also helps. I am no fan of moving away the nozzle without extruding: then the filament keeps sitting in the nozzle for a longer time, and gets a different viscosity. This also shows up as horizontal deformations in the model. And the nozzle may leak while sitting and waiting aside, causing blobs. A dummy tower reduces the effect, but does not eliminate it: Theoretical model of a dummy with complementary shape: Part of a real design: Tiny models, each ca. 20mm high, printed at different speeds and temperatures. Printing tiny models on their back -if the model allows it - also increases layer cooling time:

@thorsenrune: I don't know what this "@-thing" does, other than drawing visual attention, so I rarely use it. :-) And yes, you can quote me on the Onshape phylosophy. I am used to very wel organised graphic design packages, with easy navigating through logical menus and toolbars. Functions should be where you expect them to be. But in Onshape all functions seem to be randomly splattered all over the screen, and often hidden, without any logic (or at least I can't find it). I had to consult the manual *every time*, again and again, even for the simplest functions like saving or exporting a file. Further, I couldn't orient and rotate the models on-screen: they would tumble upside-down and then I couldn't get them upright again. While in other software rotating works so smoothly and intuitive. Also, often I couldn't select the parts of a design that I wanted, and I kept losing the selection, or it selected other stuff than I wanted. And so on and on, endlessly... Thus, even though Onshape theoretically had a lot of possibilities, I couldn't find and use them. And it was too unpleasant. By the way, I had similar problems with older SolidWorks versions, which came from the same engineers I think, although to a lesser degree (I don't know the newer versions). I think this is a pitty, and the Onshape-engineers should really reconsider their user-interface. It should be very intuitive, and very straightforward and logical. It should adhere to the UI-standards and logical menus that were set from the very beginning in graphic design computers, in the 1980's, even before Windows, Mac and Linux came along. Then it could have great potential. It is like adhering to the standard way pedals and the steering wheel in a car do work. What I like in SketchUp is the simplicity of its user-interface, with its push-pull concept. It is very straight-forward and logical. For me, DesignSpark Mechanical is close enough to this, but it is better for engineering parts, based on geometric shapes, and it never caused any printing problems. If you would prefer a more architectural- and interior-oriented CAD-software, that can do both push-pull and numeric design, have a look at "Form Z" from AutoDesSys. They now have a freeware but feature-limited version too. The user-interface is close to that of SketchUp. A unique feature of Form Z is that it can do very smooth nurbs (=3D-splines and organic curves). For a demo, see: I have played around with their beta-versions in 2014-2015. But this is offline-software (or at least it was back then; I don't know how it evolved since). Also, I don't know how solid and error-free their 3D-models are, for 3D-printing. But its free, so you can try. :-) For teaching, indeed ,I can see the value of online-software, compared to having to install it on all classroom- and student computers. Here at our university all teaching-software is installed and maintained on one master-computer only. And then this master is mirrored automatically to each student-pc in all classrooms. This combines the advantages of both online- and offline software. Maybe that concept might also be an option for your classes? On our laptops and in our offices, we can install what we want, as long as it is legal and virusfree, of course.

I have tried Onshape, in the beginning when you could still have 10 private models (all above 10 would become public, accessible to the whole world). But I don't think you can still have 10 private models? Maybe 4? Unless you pay big money per month. Anyway, I could not find my way around in its user-interface: I can't find anything and can't get anything done. Its phylosophy seems to be incompatible with mine, so I stopped using it. This could be my problem of course, since some other people *can* make great designs in Onshape. :-) I have a very different graphics background than most other people. Now I use DesignSpark Mechanical, freeware from RS-components (only requires registration), and a feature limited version of the commercial SpaceClaim. Its phylosophy is close to that of SketchUp, so the transition is easy. And there are lots of good video-tutorials on Youtube. You can make exact dimensions, but you can not set variables and stuff like in Solidworks. Others here use Fusion360, but I haven't tried that yet. I would suggest that you watch a couple of demos on Youtube of all candidate-programs, and see how they appeal to you. Everyone has his own taste and requirements.