geert_2

If the cause is not clear, maybe you could try along this line: first try to use a "known good" filament, which gave no problems in the past. Carefully watch what happens while printing, at exact which moment the blobs appear, and why. Sometimes you can see what is going on. And then, on the fly adjust all sorts of parameters a bit up and down, one by one: temp, speed, fans, material flow? Set each one back to default before changing the next. Of course, also check if the nozzles are clean (do atomic pulls), if the feeder wheel and traject are clean, if feeder pressure is correct, and if the whole feeding traject flows smoothly. And check filament diameter. I could imagine that this sort of problems could be caused by irregular feeding: pressure somewhere in the feeding traject building up, and then suddenly jerking forward. Or a piece of filament stuck in the knurls of the feeder wheel, causing an irregular diameter. If none of this helps, maybe you could design a new model with the same outer dimensions in a simple editor of your choice? I would suggest DesignSpark Mechanical: easy to learn, powerfull, and free. But do *not* use SketchUp. See if that has the same problem? This in case weird model geometry might also play a role?

Yes indeed, that's a good point.

For structural integrity, I think it is best to do a test print: make a ruler of 10mm x 10mm x 200mm or so, and print it with various infill-levels: 50%, 70%, 80%... I do understand the desire for low temperatures by your supervisor. Otherwise students are going to dial temps up all the way, to get a faster flow. And then he has to spend all his day cleaning out burnt plastic and blocked nozzles... If I had to run a huge lab with printers, probably I would assign materials to each printer: this printer only for PLA, this printer only for PET, this only for abrasive stuff, this for messy things and experiments, etc... And I would interrogate all students before starting to print, on all aspects of it; like sort of an exam, to get a "printing license". Only students who would pass with +90% would be allowed to print. The others would first have to restudy the manuals, and redo the test until they pass.

Have you tried printing much slower and cooler? And keep the material flow as constant as possible (=print all things with equal speed)? So that there is no excess pressure build-up in the nozzle? On my UM2 I would aim for 25...30mm/s, and the lowest temp that does not give underextrusion, maybe plus 5°C. Do a test and adjust speed and temp on the fly. In my experience, some materials tend to have this more than others, especially somewhat rubbery materials like PET. The blobs often come about when the printer travels and then lands again, when the nozzle leaks due to internal pressure build-up. Or when material accumulates on the outside of the nozzle, sags, and is deposited on the print.

Cool. What is the purpose of these models? To show the internal components? Or as part of a scale model for a new plant? I have worked in a huge chemical plant long ago, in the port of Antwerp, so I have seen and operated quite a lot of distillation columns, heat exchangers, valves, coolers, pumps, and stuff. These models look really realistic, with all the bolts and flanges that I would expect, and even the eyes for lifting it with a crane. Are those little tubes hollow all the way through? Wonder how that is going to come out of the printer, and how you are going to get the support material washed out of them...

I like the city-idea. This really shows how much creativity they have. Did you require the buildings to be on an exact scale, like HO (=1:87), so the designs fit in a standard landscape with HO-scale trains, cars, and figures?

Maybe you should give that same model another try (=same gcode file), but now closely watch what happens? Vibrations could make things move, I have seen noisy transformers vibrate off a table, but I have never seen this in my two UM2. And since the head moves both directions, I would guess that the frame should move in both directions equally too, if it moves at all? Unless accelerations are way different in both directions, or if it rotates like a helicopter blade, or if the table isn't horizontal? Anyway, to prevent it from falling off the table, you could use wood or metal bars, and glue them to the table-edges, or attach them with powerfull glue clamps if you don't want to glue. So there is a ridge. Then it can move around, but not run off.

Instead of using a printer, maybe you could also try a couple of incandescent lamps or spots as load, for about the same wattage? Then you can easily see from a distance when they go out, and it doesn't cause a failed print.

If the nozzle is clean (= if you can look through it from above after an atomic pull), and the feeder knurled wheel is clean, maybe it could be the teflon coupler? Have you checked if this white teflon coupler on the nozzle isn't worn out? On an UM2 (non-plus) this has to be replaced every couple of 100 hours, depending on materials and temperatures used. On an UM2+ is should live much longer, but not endlessly.

As you say, what about making a negative, and then again positive of it? Maybe it could be done as follows? First make a copy of the design, and only work on that copy, so you do not destroy the original. Then make sure your house has no gaps: so no open windows, no open doors, no open chimneys; all must be nicely closed. Put a box around your house, a bit larger than the house. Subtract the house from the box. This will give a complex thing, since all interior details will also be subtracted. But you also have the negative shell. Move the negative shell (box) out of the way, and keep it. Delete all the rest. Now you have only the negative shell, not the interior anymore. Make another box of the same size. Subtract the negative shell from the new box. And there you have a positive house in one solid block. Print with some infill as desired for mechanical strength. If your software supports it, maybe an easier way would be to fill everything: select all, or select the shell, press "Fill", and hops, you have one big solid. Might work with some 3D-packages?

I found that keychains and ID-card holders in my pockets survive longer in PLA than in NGEN and PET, contrary to my expectations. At least, as long as the PLA does not have to flex! In a year time PLA gets harder, stiffer and more brittle, but it does not crumble apart, also contrary to my initial expectations. So, as long as its function does not require flexing, and does not require warmer temperatures than 45°C, standard PLA might be good enough. But I don't know what it will do in 5 years time? However, if flexing is required, for example in snap-fit locking mechanisms, then the tabs in PLA will break after a year when forcing them in or out, because they lost their flexibility. In that case NGEN or PET is better. I haven't printed in ABS, apart from a few test pieces, so I can't compare these to 3D-printed ABS. A disadvantage is that you can not make screw threads in PLA, nor drill, because it just melts. And like almost all plastics, PLA suffers from severe creep deformation and stress cracks, if subjected to continuous loads. I would suggest that you print a couple of things in various materials and keep them in your pockets all day. Keychains are a good test item. And see how they survive when in contact with hard metal keys, when getting stuck and getting bent when you move or go sitting, etc.

For bonding PLA to PLA, I use cyanoacrylates. Just the plain standard liquid 3ml tubes, as found in any shop, Loctite at this moment. These work very well on Ultimaker PLA and colorFabb PLA/PHA. When breaking the bond, often it breaks not in the glue-interface, but in the PLA itself, which means the bonding is good. Roughening the PLA-surface a little bit (but not too much) with a file also helps. Since cyanoacrylates are said to bond to glass too, it should work. Keep in mind that for cyanoacrylates, thinner layers of glue give stronger bondings: the best is a layer of only a few molecules thick. So, do not make the surface too rough. For NGEN and PET, cyanoacrylates do not work so well in my experience: they do bond, but can not withstand as much force as PLA. I haven't tried composites on PLA yet.

I would also like to see photos of the casting process: the mould, venting channels, the furnace, pouring the melt in, rough casted model without post-processing, etc... Although I am not into metal casting, I think it is interesting and educative. The main difference between a metal part and a plastic part, is that the weight makes it feel more valuable, even if they look identical.

This model can have layer lines and print lines, I think. It gives it a nice "industrial look".

For the cause I would as well look into: vandalism, pets, kids playing and accidentally hitting it, someone who got stuck behind the power cord and pulled the printer off the table (and dares not admit it), and similar? Or maybe the spool got stuck on something (e.g. on a book, a ridge, a heavy box), and when unwinding filament while printing, it sort of rolled the printer off the table, like a cart? For the nozzle covered in PLA, maybe this could also have happened as a result of the crash, the models coming loose by the shock, but the printer continueing printing anyway? Just guessing...

This is a smooth model. It would be good if you could post a summary of the most important settings: material, printing speed, temp, nozzle diameter, layer height. Might be helpfull for other railroad enthousiasts. One remark: if this is PLA, do not place it on a hot radiator, as in photo 1, and don't leave it sitting in the sun in a car. It will warp. Don't ask how I know...

Ja, daar heb je een goed punt. Mijn bekommernis was dat wanneer je op 100°C per vergissing de nozzle aanraakt, dan heb je nog een fractie van een seconde tijd om je vinger terug te trekken voordat hij verbrandt. Bij +200°C heb je bijna ogenblikkelijk brandwonden. Mijn PLA is bij 100 à 120°C al behoorlijk zacht, maar veel andere materialen niet.

Hoi Sander, If I select "Insert other media > Insert existing attachment" (at bottom right), then inserting images at the desired cursor location always works fine, as expected. However, if I select "choose files" (at bottom left) and then browse to the desired file on my harddisk, then it gets attached at the bottom, not in the text at the cursor position. Next, when I click that image, or when I drag it into the text to have it at a desired location, it sometimes gets added twice when uploading the message, sometimes not. When I try to delete the duplicate, by editing the message (after sending it), I can not remove it. Or better: I can remove it from my editing window, but after uploading, it appears again. While editing, I have to remove both images from the message, to get rid of the duplicate (and the other one too). And then I have to insert the image again, but now from the "Insert existing attachment" function. Then it works fine. (Windows 7, most recent Pale Moon browser.) But this problem does not always happen, sometimes I can add 2 images well, and the 3th gets duplicated.

Ik ken de Ultimaker Original (Plus) niet, dus hier alleen een algemeen antwoord: - Heb je al geprobeerd om "atomic pulls" te doen? Voor een zachtere methode, zie mijn handleiding: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ . En daar een beetje naar beneden scrollen. - Probeer ook eens om bij warme nozzle (ca 100°C) zeer voorzichtig met een zeer dunne naald doorheen de nozzle te prikken, om eventueel roet eruit te duwen. Maar echt wel zeer voorzichtig, zodat je de nozzle niet beschadigt. - Verwijder ook eens de bowden tube, warm de nozzle op, en probeer een stukje filament manueel doorheen de nozzle te drukken. Doe daarna hetzelfde maar dan doorheen de hele bowden tube (=dus losmaken achteraan feeder). Dan kan je makkelijker voelen of daar ergens een probleem zit. - Kijk ook eens of het feeder wiel niet slipt op de feeder as?

Yes, indeed. Buy a second glass, buy a cheap electronically controlled oven, and preheat the second glass in there when the other is almost finished printing...

Lasercutting out of a plate of colored plastic (plexi glass, polycarbonate, etc...), metal or wood, could also be an option. This might go a lot faster for larger sizes.

There seems to be a bug in the image uploading system? If I select an image from the already existing images, there is no problem. But if I want to upload a new one, this sometimes gets published twice, sometimes not. Editing the message again, and deleting the duplicate, does not work. I have to delete both copies, and then re-insert it from the existing images (to which it is added when uploading); then it works fine. This on Windows7, newest Pale Moon browser (=Firefox split-off which keeps the menu-bar and statusbar, like in the old Firefoxes).

Most of the time I simply can't remove my models as long as the glass is above 30...35°C, or this would damage the model and decalibrate the glass bed. However, you can speed up cooling quite a lot by putting a 25cm desktop fan in front of the printer, as soon as it is finished.

If I had to print this on my UM2 in PLA, I would try 20...25mm/s, and 180...185°C first. And then watch what happens and adjust on the fly as required. The slow speed is to get as little nozzle pressure as possible, thus even flow and less leaks during travel, and the low temp to improve cooling and reduce stringing as much as possible. And of course I would print a dummy model next to it, so the nozzle has more cooling time, to avoid heat-deformation. Less heat deformation with a dummy (square tower) next to the real testmodels (cones). Numbers are cm, little lines are mm. The cones are 20mm diameter x 20mm high; the blocks are 10mm x 10mm wide, and higher than the cones. Tips for dummies. Example of real dummy and part of real model. Another dummy in a real print.

Doesn't Paint3D have an export option to save as STL-file? First save in the native format, so you don't lose the design. Then export to STL if that options exists. Otherwise, in addition to converting, you might also want to try other simple CAD programs? Maybe Tinkercad? I haven't tried it yet, but I heard that some schools use it. Maybe other programs too? For kids of +10y, you might try DesignSpark Mechanical. It is more technically oriented, but still very easy to learn. And there exist a lot of good tutorial videos. Google or Youtube for these programs, and watch a few demos, to see if they appeal to you, and to the kids.