geert_2

Forgot to say: due to the dissolving-effect of the solvent, the outer layer of the parts gets soft for a while. So that will leave fingerprints, and it may warp the whole part if it is thin. Also, if you print with little infill, it may evaporate the solvent inwards into the hollows, where it keeps working. I found that a mould I printed with 25% infill and of which I smoothed the inside (=where the cast comes), began to warp slightly after one month. I think that might be a result of the dichloromethane-solvent, because it never happened in similar unsmoothed parts in the same conditions (=room temp, average moisture).

It might also be a good idea to search on internet for demo-videos. Then you get an idea of how smooth the extruded sausage is, and if it fits your application. Most models that I saw gave a quite irregular extrusion, or maybe seemed irregular due to our hands moving irregularly? It would work very well for creating trees in miniature railroad landscapes, but not for smooth geometric objects.

At least for prototyping, I would begin with classic materials like PET or even PLA. But do some smoothing on their surface to reduce layer lines, so dirt and bateria have less grip. These do withstand desinfecting alcohol. For smoothing, have a look at the thread I did some time ago, with lots of pictures of the result. Search for: PLA and PET smoothing with dichloromethane. Chloroforme should also work, but I haven't tried that (too much hassle, requires special permissions here). Be aware that after smoothing, the parts will breath-out that chemical for several hours, so you need to give them time to dry completely.

I just read this now. An option might be to print it in nylon, but a bit too small, and then using a vice press it on the brass element with brute force? So it will seal well and won't fall off? Nylon may be able to handle this brute force.

Wouldn't you be better off using a lasercutter or waterjet cutter as base to start from? These are designed to go very slow, so they should have all the basic hardware ready. And they can run 2D DXF-files generated by most CAD-programs, I think.

There are people on the forum here who use PLA or dedicated mould-making materials to print a model, and then encapsulate that in sand (or whatever special mould-material), then burn the plastic model out, and cast metal into the cavity. I have seen videos of it, but I don't know the exact methods and materials. There are jewelmakers who do this with silver and gold, and also general hobbyists who do it with aluminum. Maybe you can find tutorials on Youtube?

Out of curiosity: have you tried, if you use different transparent colors, let's say blue and red, does it then get the subtractive mix color purple?

Catia, isn't that from the Dassault company? If yes, I think it might use the same engine as DesignSpark Mechanical (but then at full power, not feature-limited like the freeware DSM)? I rarely have this sort of problems in DesignSpark Mechanical. But what occasionally happens is that things do not want to merge (or round, chamfer, cut, extend, whatever,...) if the edges do *exactly* fall together. If they overlap 0.01mm or so, this is not a problem. It is the "just touching" that is the problem. I have also seen this in other software, and in computer games as well: then you get flickering or zebra-patterns when both surfaces "fight" to be displayed. Most of the time all those math operations just work fine in DSM, but if not, a solution is: make a small gap if you do not want (automatic/undesired) merging, or make a small overlap if you do want merging. In DesignSpark Mechanical, you can give each object a different color. If they do merge during editing, on purpose or by accident, you will notice it because they all get the same color suddenly. This is a visual check, sort of. You can also see in the model tree if it is one object, or lots of objects.

Also, printing in thinner layers normally gives a smoother surface. But twice as thin takes twice the time to print... These blocks are in PET, printed at 0.10mm and 0.06mm. The green model on the top right is in PET too. It's hard to see here, but the surface is quite smooth, letting a hollow watermark text shine through. Note that this model is small, see the ruler in mm and cm below. The red object shows 50% underextrusion (was part of an underextrusion test).

An STL-file describes a 3D-model in triangular surfaces. But a 3D-printer can not print triangles, it can only print single extruded lines. So the model has to be cut into thin slices first, and these slices then have to be cut into toolpaths, trajects that the nozzle can follow while extruding. This cutting into slices and toolpaths is done with a slicer-program like Cura. It outputs a gcode toolpath file ("somefilename.gcode"). This gcode-file is the file you need to put on the SD-card to print. So you need Cura. You usually have 3 file formats of each design: - the model in its native 3D-fileformat of your CAD-program (SKP-file for SketchUp, I think), - the model exported as STL-file, needed for slicing, - the toolpath GCODE-file made by the slicer, which can then be printed, - and I also add a JPG-picture of each model, so I can easily recognise the design in my Windows Explorer, without having to open all CAD-files But as said above, SketchUp is going to be an endless source of troubles, because it produces incorrect STL-files that can not print (at least not without lots of repairs). So you would better use another 3D-editor. DesignSpark Mechanical is free (requires registration) from RS-components, is easy to learn, and there are lots of good tutorials on Youtube. That might be a good alternative, but there are lots of others too.

Yes, just do tests on tiny items, so you know from which point on the dummies become necessary. This takes only a few minutes due to their small size, but it can later save a big model. After some time you know and you can add the dummies from the beginning in the design. Below is a real application: the dummy is the green cube at the top right. Its bottom is hollow, its top is filled from the height onwards where the large flat areas in the blue object ends, and only the tiny high yellow part remains. Without dummy cube, the top of the yellow part would seriously deform. For reference: text caps height is 3.5mm, text legs are 0.5mm, thus all is quite small.

I am going to do some wild guessing: could this be a result of high pressure in the nozzle, after printing infill at high speed and temp, and then suddenly slowing down to the outer layer. Thus the built-up pressure and temperature has to leak away, sort of? Or a result of a move through the air, where it temporary stops extruding, with the same effect? Or both together? If the first, then setting all speeds and temps equal should minimise this effect. If the second, printing slower and cooler should minimise it (but printing cooler of course might reduce layer bonding). But as said, this is guessing.

Yes, but then it's a single use mould. Not the best option if you need a lot of casts. It sort of defeats the idea of moulding and casting. A good alternative to a fully 3D-printed mould, could be a silicone mould in a hard 3D-printed shell. It goes a bit like this: - Print the real model in PLA or whatever. - Design in CAD a shell that is sitting at some distance from this real model, maybe 5mm to 10mm (depending on the size of the model: bigger models need a bigger distance). This can be a very simple shell: two parts, with clamping flanges, alignment keys, stable baseplate, and big opening on top. Make sure to minimise details and undercuts in the shell. - Mount the real model in the shell, and verify there is some distance around the model. Bolt or glue the model to the shell. Removable glue like "rubber cement" might be a good choice. If not fixed, the model will go swimming around. - Carefully seal all seams, they must be absolutely watertight, or the silicone will leak away. - Pour silicone into the shell, around the model. Preferably transparent silicone, so you can see the model through it. Let cure (often overnight, depending on the silicone). - Remove the shell. This should be very easy if well designed, due to the simplicity of the mould and the flexibility of the silicone. - Now using a scalpel or very sharp knife, cut the silicone in half, or in multiple parts as required by the model, so the model comes out. While cutting, make zig-zag movements, so you get natural alignment keys in the silicone. It looks ugly, but it works very well. Cut along natural seam lines of the model. - Treat the mould with plenty of silicone release spray before casting, so it really soaks in. In this way, you get a reusable silicone mould, easy to clean, easy to demould, suitable for many many casts. There are lots of excellent tutorials on Youtube, search for: moulding and casting Pics: 3D-printed shell. Model (teeth) was mounted with plasticine. Alignment of both mould-halves is by M4-screws, also used for clamping the mould. Note the lots of draft in the mould, required for easy demoulding from the shell. Also note the zig-zag pattern in the silicone halves. Old-school mould, with hard shell of two-component epoxies. Note the zig-zag in the silicone, for alignment. Also note the bottom plate for positioning it upright stably, and the big pouring opening on top, and the alignment keys. Side-view. The flanges of the shell are for clamping the mould and for the alignment keys. This was made manually, long before 3D-printing. But you can use the same concept. This is fast-curing silicone: advantage is that no sealing of the mould is required: it cures before it leaks away. Disadvantage is that bubbles can not be evacuated: it cures too fast... For casting solvent-like stuff like polymethylmetacrylates (PMMA) or PU, lots of silicone release spray are required. Otherwise the solvent penetrates the mould and makes its life much shorter. Silicone is watertight, but not oil, parafine and solvent-tight: these seep into it, and would cure in there, thus hardening and destroying the mould. So, saturating the silicone with silicone oil prior to casting helps a lot.

If I had to do this, I would probably do the mould-making in CAD: subtract the model from a solid block. And then cut the block to pieces along natural seam lines, so the cast can be demoulded. Next add flanges to clamp mould parts together, add alignment features ("keys"), add a stable baseplate so it does not fall over or slide away when pouring heavy plaster in it, add air venting holes if required, add pouring holes if required, add features to lift the whole mould, add features to insert a screw-driver in-between mould-parts to wiggle them apart later on so you can open the mould, etc... Requires lots of work, and lots of thinking, but it gives you way more control and understanding. Making a mould is not that difficult. But afterwards opening it, and getting the cast out undamaged is... :-)

Not NGEN, but for other similar products (PET): I have to print them very slow, in very thin layers, and around the lower edge of their temp-range, in order to get them reasonably transparent. After printing, sand and polish to remove layer lines. It is the entrapped air in-between the sausages that causes the whiteness, due to reflections and diffractions. Like sugar crystals look white, although they are transparent. If sitting in the nozzle for too long, due to the slow printing speed and thin layers, PET starts to decompose and discolor brownish. So, lower temp to minimise this effect. Make a simple test block like this below, and print it with various settings. Top row: speed 50mm/s, bottom row 10mm/s. Layer thickness from left to right (out of a standard 0.4mm nozzle), both rows: 0.40mm, 0.30mm, 0.20mm, 0.10mm, 0.06mm. At 0.06mm and 10mm/s: left as printed, and right after sanding and polishing a bit. The model: 20mm x 10mm x 10mm, with floating watermark halfway (text caps height 3.5mm, legs 0.5mm, thickness 1mm): Printed in 0.40mm layers:

A few tips (maybe you already know them, maybe not): If you would go the moulding and casting route, be sure to post-process your mould very well: remove all layer lines as much as possible by sanding, coating/painting, or chemical smoothing. Otherwise the cast may be very hard to remove, as each layer line acts as a tiny undercut. Don't ask how I know... :-) I do the smoothing with dichloromethane now. (See my separate post on: chemical smoothing PLA and PET with dichloromethane, should come up in search.) And I make my moulds in silicone from 3D-printed models, and in 3D-printed hard shells (unless I want to cast silicone). If casting PU or epoxies, (1) use plenty of release coatings in plastic moulds, or it will glue like hell (that is why I use silicone moulds, and even then plenty of release spray). And (2) use low exotherm epoxies, or else a 3D-printed mould will melt, and the cast itself might catch fire or explode, if too much epoxy is used at once, since the heat can not radiate out quickly enough. I have had plastic cups and even metal (lead) cups melt... For 500V, I wouldn't worry too much about arcing or breakdown, if the material is thick enough. I don't remember, but isn't the general rule in air that you need a safety gap of 3mm per 1000V? Or was it 6mm? If you can find that rule, and multiply it a few times, you should be safe.

Normally, a well-tightened connection of a pipeline is absolutely gas- and watertight, even up to pressures of 200bar, as in my hydraulic machine. This is brass on brass, or brass on inox. However, the tiniest damage can cause it to leak. So I could imagine that one of both parts has an uneven mating surface from earlier contamination, or from careless machining, or from dropping or hitting something. Or one of the treads could be too short. Or, if the thread-cutting is done after machining the part to length, it will cause deformation and an uneven surface close to the cuts. That sort of things could also be a cause...

I would have some doubts about carbon-filled materials. The leads of pencils are carbon-filled too, and are *highly conductive*: short-circuit a battery with a pencil lead, and you get a really nice welding arc... Black anti-static mats get their properties from the carbon too. Make sure you measure conductivity on a small test print, and also the breakdown voltage. So the glass-filled looks like a better idea to me. Maybe PET should also do, at least for normal house-hold voltages? But I have no clue about its high-voltage properties. Also, have a look at the working temperature at high loads: some equipment might run at 100°C or more, too high for most 3D-printing materials. In that case, you might consider printing a mould, and casting it in a thermoharder or two-component material. Maybe that is why they used bakelite in the first place?

If I had to do that, I would consider printing a mould, and then cast some sort of rubber (PU? - which exists in various hardnesses) into it. Or print a real model, print a shell, pour silicone in-between model and shell, and thus make a silicone mould. And then pour rubber into that silicone, to prevent it from sticking. Then you have the advantages of both 3D-printing and casting.

I would print a dummy "cooling tower" next to the real model. So the hot nozzle is moved away from the model, while it is busy printing the dummy, and the model has time to cool down and solidify. Where the bottom of the dummy is empty, since the real model is big enough, and the top of the dummy is 100% filled for the extra cooling time. My standard pictures on the subject: Printing with and without dummy: Concept: inverse shaped dummy, to keep printing time per layer constant: Part of a real model, with pink dummy. The dummy is only needed when printing the tiny top area, so the rest is empty: Another real model, with dummy (red). Here too, the dummy is empty, except for the top where it is needed:

I occasionally print PET. In the beginning I tried printing on bare glass, which gave mixed results. Then I tried dilluted wood glue, which gave *very good* bonding. Way too good, because at one time, it chipped the glass. This already happened while cooling down, I heard it crack violently. When I removed the print, it came off without any force, but with the glass chip stuck to the print. Now I use my salt method: apply a few drops of salt water to the glass bed, and wipe that with a paper tissue until it dries into a very thin mist of salt. For PLA, and as long as the glass is hot, this salt method greatly increases bonding. But it has no bonding at all after cooling down. However, on PET this slightly *reduces* bonding, but also makes removal much easier, compared to printing with dilluted wood glue or bare glass. Disadvantage is that I have to switch off the cooling fan, otherwise corners tend to lift a bit. No cooling does reduce the quality of overhangs and bridges a lot. But most of my models have no overhangs anyway, so not really a problem. The ease of application is the biggest benefit: just wipe the glass, and ready to go. No need to take the glass out of the printer. When trying new bonding methods: stay with the printer, and watch what happens. So that you see how it works, and you can stop the print in case models come off and produce spaghetti or slide around under the nozzle. These models are in PET, bottom-side up: notice the little pits from the salt crystals, and the nice flat surface, squeezed well into the glass. For reference: ruler is in mm and cm. This is what the glass looks like after applying salt water, and wiping it dry. This is a bit too much salt. This is a better amount of salt. (Don't mind the spaghetti in the print: these are free-hanging supports to print a bridge, and will be removed later. This print is PLA, not PET.)

I don't see any underextrusion or broken lines any more. A couple of blobs and what appear to be "insect antennas" (at the right, but hard to identify in this picture), which is what you can expect with PET in my experience.

If the rest of the part was still okay, thus not worn-out and not brittle or crumbling apart, and if it was my machine, I would probably consider modeling only the damaged areas. And then cut these off the original part, drill a couple of holes, and bolt the new 3D-printed parts on, using a M4 or M5 bolts and nuts. That is, if there would be enough room for that in the machine and in the part, of course. But this won't work if the original is crumbling and too brittle (it looks a bit like that?). It seems to be an ABS-blend, as I noticed the words ABS on the side.

I have manual leveling on my UM2 printers. I did that according to the official procedure in the beginning. Since then I only adjust it occasionally on the fly. I start a print with a thick skirt (5 lines or so), or when printing the brim if I use brim (rarely), and then I watch closely how the first layer is, and adjust on the fly by moving the screw a 1/8 of a turn, and then watch again, etc. I think it is a little bit closer now than originally, indeed. But I only adjust that maybe once a year... My bottom layer is 0.2mm thick: I found that 0.1mm is too thin and a bit uneven, and 0.3 is too thick and gives lesser bonding than 0.2, especially for objects with very small holes like in my photo above. Everyone has his own taste: I like the simplicity of manual leveling, like manual shifting in a car, and a manual handbrake. :-)

Wow, these are really impressive paintings. They look like the antique art you would find in museums. When priming your paint spray cans, maybe you could do that on a plain sheet of white paper, or sheet of wood, and keep this sheet with all color patches, and make photos of it? This would give an impression of each of the colors and effects on its own. Question: before painting, do you chemically or mechanically prepare the surface of 3D-prints, for a better bonding? And do you use primers that chemically dissolve and penetrate the plastic, or do they just coat it?