geert_2

Ik maak in mijn ontwerpen in CAD alle lijnen 0.5mm dik, voor een 0.4mm nozzle. Dat werkt ook gemakkelijk op een 0.5mm ontwerpraster. Bij het slicen worden curves nadien vervangen door rechte stukjes: die snijden de bocht af, waardoor de dikte een beetje varieert tussen pakweg 0.48mm en 0.52mm. Als je dan als lijndikte 0.40 genomen had, varieert de dikte na slicen misschien tussen 0.38 en 0.42, en valt alles beneden 0.40mm weg wegens te dun. Ik vermoed dat de lijndikte instellen op 0.35mm tijdens het slicen, zoals hierboven gezegd, ongeveer hetzelfde effect heeft? Maar dat heb ik nog niet geprobeerd.

If you have them on a server, you have a chance of getting them back: on most servers there is "shadow copy" software running. When deleting or modifying a file, it keeps a copy of previous versions for up till one year, in a hidden folder. On our systems it is named "~snapshot". A bit similar to the Windows Recycle Bin, but then with different file-versions too, and not so easy to delete as emptying the bin.

Yes. These models were designed a couple of years ago for single-nozzle printers, and older slicers. So they are all printed with the exact same settings. Sometimes I provide gaps to make removal easier, so that parts are only kept in place by strings, not modeled connections. I consider these supports an integral part of the design, not a separate add-on later. Similar to when designing for injection moulding, then you also consider the fabrication method and aspects from the very beginning. These are all very small models, so reachability and removability of the supports, without doing damage to the model, is most important. I must be able to grab the supports, and wiggle or pull them out. For larger models, you could construct your own supports so that they consume very little material but are still quite strong. They don't have to be tree-shaped; you could also chose hollow cylinders or bars, or whatever shape suits the purpose. For stability, be sure to give it a huge custom brim, so it bonds wel to the bed and won't be knocked over. Make the side walls strong enough, so they don't wobble. This is why a hollow cylinder or pipe-system could be better than a thin tree in some cases. (Your concept-drawings might suffer from these: not enough brim to prevent falling over, and the stems may be too thin to be stable.) Include all required holes, slits and other features to get in there with a knife, scalpel, pliers, hooks, etc..., to remove the supports later on. Sometimes this requires a lot of thinking. For dual nozzle printers, you could design a stiff and cheap support in PLA (if het main model is in PLA too), and only use a thin dissolvable PVA-interface layer between support tree and model, like this concept drawing, with a dove-tail for good bonding of PVA-support to PLA-support:

I often design custom supports in CAD, as part of the model, and then turn the slicer-support off, like gr5 said. Then you have full control. General supports are great for general items. But special models may require special custom solutions. Yes, custom designing them may take quite some time, but you win that back when it prints well and does require less post-processing. For example, see the pink and orange supports in the first pic. This model is too small to get in with a knife or plier to remove standard supports. Now I can easily grab them and wiggle them loose. And see the free-hanging supports in the other pics, which consume very few material, are easy to remove, and do not damage underlying surfaces.

If you can't find the official oil, I think a fine oil for light machinery and house-hold stuff should be okay. For example the sort that is used for bikes, tools in the workshop, model trains, etc. As long as it does not dry out into gum, like mine did... But I didn't have a good quality fine oil, and I had 100 liter of hydraulic oil spare, so I gave that a try... Don't use thick oil (thus no grease, no high-temp oil, no car motor oil, etc.), and don't use thin dislodging oil (does not lubricate well). You can feel if the rods are too dry. Switch off the printer, and gently move the head around by hand, in both X- and Y-directions. It should go relatively smooth, with only the slight steps from the magnets in the stepper-motors, and the rubber belts. When the rods get too dry, or the bearings got clogged with dirt, moving the head around goes difficult and tends to stutter a lot.

I do it very similarly. First, moisten a tissue with isopropyl alcohol and wipe the rods to clean them: they collect lots of dust. Then moisten a tissue with a drop of oil, and wipe the rods to oil. Never pour oil on the rods directly: it will be way too much, and you don't want oil leaking on the glass, or on the rubber belts. Ultimaker recommends a light machine oil, if I remember well. However, I once had sewing machine oil from a sewing shop, but that dried-out fast and turned into gum. So, now I use thin hydraulic oil, because I have lots of spare of it from our hydraulic test bench, and it lubricates well, and it does not dry-out at all. Probably not the recommended solution, but it works for me.

The companies Materialise, Melotte, and Protolabs do metal 3D-printing. Maybe you can find the info you are looking for at their websites? Including guidelines for dimensioning? I think the models shrink a bit, but I have no experience and knowledge about this myself.

When making tiny filters for a vacuum pump, I had to print them very slowly and in very thin layers. Otherwise, if printed at 0.2 or 0.3mm layers, it had lots of tiny openings, through which tiny jets of water squirted out, similar to out of an injection needle. When printing at 0.06mm layers and 25mm/s in PLA, it was water-tight. But if you are going to print your huge model at such a low speed and layer height, it is going to take forever. So, I would make a *small* test model first, and see how that works under high pressure. Put tap water on it, or even via a high-pressure cleaner, and see how that works out (Use safety glasses and equipment!!!). If it is water-tight and can withstand high pressure from inside-out, which is the weakest, then it will also withstand high external pressure. If you would use paint, try to find one that bonds very well chemically, so it does not delaminate. The filters I printed: the bottom was printed with 70% infill if I remember well, to get the tiny holes for the filtering-effect, good enough for this purpose. Then both halves were glued together around the edges.

My first thought was: "dirty glass". If you clean it with soap or window cleaner, be sure to clean it again afterwards with pure warm tap water only. Best twice. Thus no more soap. The reason is that soap reduces bonding too. So you really need to remove all traces of soap and detergents. Don't ask how I know this. :-) After a closer look at your photos, it looks like the bottom "sausages" are quite round, not very much squeezed into the glass. That could also reduce bonding. So, in addition to cleaning, I would also check bed leveling, and bring that closer if necessary. Also check if you don't have underextrusion, thus too thin sausages, which could cause the same effect: the print not squeezed into the glass very well, but now due to lack of material, instead of the glass being too far away. If you have twice the contact area per extruded sausage, then you have twice the bonding strength. The bottom of my prints usually looks like this:

At first, I thought: are you really going to slide downhill in a 3D-printed sled? What if it breaks? But then I saw the dimensions: 12cm. :-) The first photo, it laying upside down, looks really convincing.

Thanks for the feedback, it could indeed help people in similar situations, even for other printers.

Do the walls need to be slanted for cutting cookies? For injection moulding plastics, for sure yes, for casting silicones also yes, unless the models are very thin. But for soft cookies or dough? If I had to print cookie cutters or plasticine cutters, I would probably first try straight walls, print with a standard 0.4mm nozzle (less risk of underextrusion and clogs), 0.8mm thick walls (=2 lines), print slow at 25...30mm/s, thin layers 0.06...0.12mm, and print cool at the lower edge of the temp range. Printing cool helps prevent the filament from decomposing in the nozzle due to the long transition time. Printing slow gives the molten filament more time to melt the previous layer and thoroughly fuse it. Thin layers give less indents, less entrapped air between the extruded sausages, a better fusion, and smoother surface. All that should give a quite solid and strong model. Then I would cut and sand away blobs and defects, and smooth the side walls with dichloromethane (for PLA and PET) or acetone (for ABS), so they get high-glossy and don't cause friction. Smoothing also makes cleaning much easier, and gives less grip for dirt and bacteria. After smoothing, I would let the models dry thoroughly, so they do not degass the solvents anymore while cutting cookies. Not sure if it would work, but that is how I would approach it. And then optimise further from there, if necessary.

This looks like severe underextrusion. It could have lots of reasons: user gr5 has made a good list and tutorial video on this. See if you can find it here on the forum (I don't know the links), and check each item on the list. I once made underextrusion tests. Your models look a bit like the 80-90% in my tests. This is PLA. (Lack of light, translucent filament, and narrow depth of field of the camera don't help the photo quality, but you get the idea.)

If you could redesign the overhanging rings to be at 45° or 60°, instead of 0° (where 0° is horizontal and 90° is vertical), they could print without supports. I don't know that machine, so I don't know if these rings are functionally required or not? Or if they could do with a shallower angle? That would seem the best option, if functionally possible, since you don't need supports then. If you do need the rings at 0°, and thus do want supports, but you don't want them to go all the way down, you could also design custom free hanging supports, like I did in these models. But test that on a small model first, because it will require some fine-tuning. They can easily be cut-off without doing much damage to the model, due to their tiny connection strands. Usually I make the connection strands ca. 0.5mm wide, 1mm long, and 0.2mm high (=2 layers of 0.1mm). In the bridge below they do touch the bridge's side panels. But in the slider-housing they are totally free-hanging, and are kept in place by the stringing of the filament, which makes them really easy to remove. But as said: test this on small models; it may require fine-tuning. Further, acetone-smoothing the model might also be an option, after sanding to remove the biggest defects: that should fill the gaps and create a nice gloss. Apart from that, ABS is known to cause underextrusion problems, because its extrusion temp is very close to its decomposition temp, causing it to degrade in the nozzle and partially clog it, if I understood that well. That could also be a reason for temporary gaps in the side. I am not familiar with high-temp polyesters, so I can't give much recommendations about using them as alternative materials. Maybe other people here can? Standard PET probably won't go high enough, I think ca. 80°C?

I don't know your printer, but it looks like it still needs calibrating steps, and maybe correcting slack and tolerances. You may need to find manuals and tutorials on that for your printer, or ask the manufacturer. After that, for finding the best settings for general printing: just stay with the printer and watch it closely, while printing small test models. Do lots of test models, one by one. Change speed on the fly, and see how that affects the print quality. Change temperature on the fly, and see what happens. Change cooling on the fly, change flow-rate, etc... Print the same thing in different layer-heights and try what settings each require: thicker layers (=more material) generally require higher temperatures and/or lower speeds to melt enough material. Change parameters in small steps up and down, so you find the window where quality is optimal. This will also be different for each material.

Yes, just print enough skirt priming lines around the object: 5 for a bigger object, 10 for a tiny one. And while it is printing these, turn the bed-adjusting screws until you get the desired thickness and squeezing of your first layer.

For sealing seams, maybe you could also make your own custom gaskets? Print a mould, and pour silicone in it, the kind of silicone that is used for, well, mould-making. Be sure to smooth the layer-lines of the mould, otherwise removal of the silicone will be difficult. Silicones come in various hardnesses, from flesh-like up to tire-like.

I wanted to fly RC-planes, but I can't even drive and park an RC-car: I keep getting left and right mixed-up. So I dare not try a plane where fast reactions are required, and in a roll even up and down get inverted. Maybe a huge RC-containership would go, because it is so slow I would have enough time to think the inversions through. But I am not into ships... Concerning the weight, if you visually compare old lightweight and new heavier prints, isn't there any visible difference? Are there added structures, or is just everything thicker? A 30% increase should be visible somewhere, I think? PS: if you print it in PLA, never let it sit in your car in the sun. Because it *will* warp, even in a very mild spring or autumn sun (="waterzonnetje" in Dutch, a dim sun shining through the haze or clouds). A had several PLA-parts warp in my car: clamps, hooks, and demo-models I forgot.

Just a question: this "increased weight", is that: (1) Calculated and indicated weight in Cura, prior to printing? Or (2) is it a measured increase on a scale, after printing? If (1), then could it be that Cura also calculates the weight of printing supports? Just a thought that crossed my mind, I don't know how realistic it is... About RC-planes: has anyone of you ever built one with a live camera, and a VR-headset, so you can fly it from a real pilot's viewpoint? Should be much easier to fly, I think? I never understood how some great pilots can accurately control an RC-plane from a kilometer distance, when I can't even see its orientation anymore? They must have some sort of sixth sense or telepathy with their plane. Some time ago there was a video of an RC-jet that reached almost 700km/h, and that went off to a little dot in the sky on each turn...

Is de draad versleten, of is hij volgelopen met gestold filament? Dat kan gebeuren als hij gelekt heeft. In dit geval: lichtjes opwarmen tot pakweg 60...80°C zodat het plastic terug zacht wordt kan helpen. Ik heb zelf geen printers met olsson blocks, dus verder kan ik geen concrete tips geven.

A tip: while slicing, always verify the model in layer-view mode, before starting a print. Then you can see all nozzle movements, and you can see layer by layer what is going to be printed (like brim, supports, model) and what not (too thin parts). Then this issue would already have come up during the slicing.

I am not part of the Ultimaker team, nor software development, so I can't say what you should do. However, I think it would be best if you get more familiar with your 3D-editors and STL-export first, so you can always produce technically correct and "water-tight" STL-files (=no gaps between the STL-triangles). As soon as your STL-files are error-free solids, and you still feel there is room for improvement in the slicers, then of course you could write a report. This gives you a lot more chance that it will be welcomed and considered.

Quite often you will need a couple of standard texts, such as a logo and a copyright notice. I make these beforehand and save them as separate designs, so I can always re-use them. Then, for surface text (=raised or recessed) I load the text-file and move it into the right location of the design. For raised text I union it with the model, so text plus model are only one solid. For recessed text I subtract it from the model, which again leaves me with only one solid, with cut-out characters. But I do the fusing only at the very end, after I am sure the text and model are fully correct. Otherwise it is harder to edit. It is best to keep them separate as long as possible. And save the intermediate model before fusing everything, so you can easily revert. For watermark text, thus hollows sitting in a model, I just move the text into the model, without subtracting. This works well (at least in DesignSpark Mechanical), because the text is totally sitting in the model, without any intersections of text and surface. In this way I can easily access and modify the text later on, by just hiding the main model, so only the inside text remains visible. If I would fuse everything, this would not be possible, and it would be very hard to select and edit parts of the text. A few examples: Watermark text as separate solids in a model. Not fused with the model, nor subtracted from it. It are all separate solids. In real life, you cannot move objects into each other like this, but you can in a 3D-editor. So now I can access and edit the red text easily, by temporarily hiding the blue ruler. Upon exporting to STL, DesignSpark Mechanical correctly subtracts the text from the model, and it slices and prints correctly. Raised text (but in a recessed surrounding area, so it does not protrude above the model). Here the text is fused together with the model, and it is just one solid. Fusing is done with a "union" or "combine" math-function, or whatever it is called in your editor. This is the separate copyright-text used for the above model. The text caps height is 3.5mm, leg-width is 0.5mm, and height is 1.0mm. But usually I want the text to be only 0.2mm raised, so I sink it 0.8mm into the model, prior to fusing it.

Indeed, the first layer is crucial. For me, 0.1mm is too thin and gives occasional blank areas (=uncovered), 0.3mm is too thick and reduces bonding, and 0.2mm is best: this gives good coverage and good bonding (glass bed). But for other materials, beds, and printers, it could be different. Normally the bottom looks like this (ruler is in mm and cm): For PLA and printing on glass, wiping the glass with a tissue moistened with salt water greatly improves bonding, compared to printing on bare glass. See my old manual here (and then scroll down a bit): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ For me this works very well on long flat models like rulers and calipers. Not very suitable for narrow high models like poles: they tend to get knocked off. So, stay with the printer until you know by experience how well it works for you.

I would probably use the vertical position, and indeed use a very wide brim for bonding to the glass. Or design a custom brim in the CAD model. If you want to make custom supports in CAD for a single-nozzle printer, you could try free-hanging supports. Without supports the arcs might look a bit like grapes. These below are free-hanging supports. Note that this is a very small model: the opening where the supports are hanging in, is only 5mm wide. The ribs on top of the supports are 0.5mm wide and high. Between support-ribs and the underside of the roof is a tiny gap between 0.2mm and 0.5mm (adapt this to what works best for your model: narrower gaps give more accuracy, but support is harder to remove). The supports attach to the side-walls by the stringing alone. This is enough to keep them in place, but makes it easy to remove them later on. The 1mm staircase underside improves stability and reduces curling-up of the support while printing (compared to triangular undersides). The tiny connection-tabs on the side-walls (shining through the supports) improve connecting to the walls, make removal easier, and minimise damaging the walls. These tabs are ca. 0.2mm high and wide, and 0.5mm long. It took me some time to come up with all this, but it works well for my designs. If you have a dual-nozzle printer, maybe you can still use this concept: make the supports in the same material as the model, and provide a support-layer in-between this support and the model. So that would fill up the gap that is present here. Then you get a nearly perfect underside of the model, without using too much support material. See the support-sandwich below. Try-out and optimise these concepts on a small test part only, before doing the whole print, so you don't waste too much time and material. You may need to adapt them to your models and dimensions.