geert_2

I found that the most stable way to insert photos is by first making enough place by typing returns. This is important. Then drag and drop a JPG-file from the Windows Explorer into the desired empty line. Make new empty lines as required before dragging the next pictures into your post. For already existing pictures, first put the cursor at the desired location, then make enough empty lines, and then insert the pictures one by one. Inserting multiple pictures at the same time also works, but then it becomes difficult to separate them later on, to type text in-between. You can not position the cursor at the beginning or end of the line, and type returns. You need to click on the image so it is selected, then press the "End" key on your keyboard, and then type "Return". And even that sometimes fails. This is when using Windows 7 as OS and an up-to-date Pale Moon as browser (Pale Moon is a Firefox derivation that kept the old-style menu's and statusbar). Phenomena might be different on other systems of course. If you first include the photos at the bottom, and then try to move them to the desired location, things often mess up: duplicate photos, stuck (undeletable) photos, etc... Also, if you do not make room before by typing returns, and if you insert lots of photos by dragging and dropping at once, things sometimes mess up too. Also inserting emoticons - these are pictures too - messes up things. After typing the characters ": - )" (but without the spaces and quotation marks) then the cursor does not work anymore. And often when I start typing somewhere, the cursor jumps to another totally random location and starts typing there. Positioning the cursor in the correct place again with the mouse does not help, it keeps jumping away. So there definitely is a bug in picture- and emoticon-handling. But it can somewhat be circumvented by careful dragging and dropping, and by totally avoiding emoticons. The problem seems to be related to the auto-save or auto-upload function of the message, because the problems often start at the moment of that auto-save.

The teardrop-concept is not mine, I borrowed it from somewhere, but I don't remember where I saw it first. But it works well indeed. The basic idea is not going over 45° overhangs. For designing technical and geometric models (=with flat and cylindrical surfaces only), I would recommend DesignSpark Mechanical. This is free and only requires registration. It is very easy to learn, and there are good tutorials on Youtube. The user-interface is somewhat similar to the push-pull concept of SketchUp, but it produces good and printable models and STL-files (contrary to SketchUp). It is not suitable for organic models. I would say: search for it in Youtube, and see if it appeals to you.

It also took me quite some trial and error to get tight fits working. Layer lines always extend a bit; and sometimes there are little blobs. Also, ringing around corners gives inaccuracies; and elephant feet on the first layers do increase size. This makes fitting difficult. So, usually I provide a gap of 0.1mm to 0.3mm between parts that have to fit, the 0.1mm gap giving a very tight fit (really stuck), the 0.3mm barely contacting. Usually 0.2mm is best for my designs, giving a nice fit without play (after removing any defects like blobs or elephant feet). I tried printing threads, but they are so weak in PLA that they are worthless for practical purposes with some load and movement. Making a thread with thread cutting tools didn't work at all: it just melted everything, even when moving extremely slowly, and with good lubrication. At least in small sizes of M3 to M6; I didn't try large sizes like drink bottle caps. And the resulting threads were too weak anyway. So I mainly use alternative ways to fix things: screws, snap-fits, glue. Snap-fits lockings in PLA do work well if freshly printed. But after a year the PLA gets harder and more brittle, and then they break when trying to unlock or relock. Snap-fits in PET keep working. For printing *horizontal* holes for screws, I sometimes give them a teardrop shape to avoid sagging (see picture). For printing small *vertical* holes, I design them at the desired size + 0.5mm (e.g. 4.5mm to let an M4 screw through), and then drill them out with a manual 4mm drill to remove any blobs or so. Glueing with cyanoacrylate works very well on PLA, even on flat unprepared surfaces. The glue seems to chemically eat into the PLA: fractures are usually in the PLA, not in the glue. On PET or NGEN, cyanoacrylate glue works less well, and here the bonding-interface fails (=place where glue and NGEN meet). Often I also use nylon screws to fix things. I usually design a case around the nylon nut, so it doesn't fall out, as shown in two of the images below. With a nut-cage, nut and screw don't fall out when released. Teardrop shaped horizontal holes prevent sagging of the overhang. Another sort of nut-cage, now with tiny retention tabs. Even when the screw is totally removed, and the blue spoon is removed, the nut stays in place. Handy: this drill chuck gives a very good grip, and allows careful and manual drilling with good feeling, to avoid melting of the PLA. (Electric drilling nearly always results in melting.)

I don't know how flash memory works internally either (apart from that charge is stored in insulated cells). But maybe the same effect as I describe could happen if a defective controller writes to the wrong memory blocks? Or if it writes to the right blocks, but the write-function is defective, so the old data can't be overwritten? If it still happens after formatting, I would discard the old SD-cards.

I also had designed something to prevent undesired unwinding near the end, to prevent tangling, and to clamp the end of filament while changing spools. But it was only for colorFabb spools. It is a sort of small clamp that freely slides around the edge of the spool, and prevents the filament from jumping off sideways. Can be fixed with screws when the spool is in storage. See here (scroll down a bit): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ But I also like both designs suggested above. They didn't exist yet (or I didn't know them) when I made my clamp, otherwise I would have considered one of the above.

That looks like a writing-action to the SD-card was interrupted half-way, for whatever reason. For example: the user pulled the card out of the computer while it was still writing data to it; or a software crash stopped the writing prematurely; or there was an unrealiable connection, e.g. dirt or oxydation on the connector pins. When you "delete" a file on a disk, its file-contents are *not* deleted. Only the reference to it in the index is deleted. But the old file itself is still intact. That is why you can recover deleted files with special software. So the disk space occupied by the old file is listed as free, although it is still full of the old data. A new file will then overwrite these old data. At the end of each file, there is a code indicating "end of file", so any software that opens the file knows where to stop reading. If the writing of the new file is interrupted half-way, then that damaged part of the file has no "end-of-file" code yet, of course. So the next software that is reading that file (here the printer) does not know that it should stop reading at the end of the damaged new file, and it continues reading the old data from the previous file which were still there. Maybe this is what happened? That would give exactly those phenomena you describe. It is a guess, but an educated one. :-) Writing to SD-cards may go slow, especially if there is an antivirus active, or if the SD-card controller has to move large blocks of data from unreliable memory cells to reliable cells. Or if the data are very fragmented (=spread all over the card in small pieces, instead of in big continuous chunks). This can cause unexpected delays. So that is why you always need to use the operating system's software to eject a card or portable hard disk, and wait for the message "This disk can now safely be removed", or something equivalent. Personally, I would completely reformat the old SD-card (after backing up any important data). And then do a thorough disk check, which also checks the hardware (=memory cells / hard disk surface). And then try the card with a couple of big video files or so.

Don't you have issues with edges of standard glass or mirrors curling up, due to the temperature differences between their upper- and bottom-sides? Standard glass has a much higher thermal expansion. Or is it more flexible, so its flexibility compensates for the curling up, and the clamps can easily keep it down? The PCB material is most likely fiberglass-filled epoxy indeed. Its natural colors usually range from brown-yellow, over pale yellow, to pale green. If the PCBs were manually cut to size, you should be able to see the glass fibers on the edges.

This is how the bottom generally looks when using the salt method. On top I focused the camera on the surface itself. At the bottom I focused on the reflections of a desktop fan in that surface. (Unfortunately I couldn't get them both in focus at the same time.) So the surface is quite glossy, with very small pits due to the salt particles. Inverted pyramids, thus models with a very small baseplate and huge overhangs, are likely to lift, due to the high shrinking forces and the nozzle bumping brutally into the edges that curl up. This was a test model specially designed to test how far I could go in inverted prisms. This is how the bottom usually looks in close-up. Nozzle size is 0.4mm, to give an idea of magnification. Often around corners a tiny area of 0.5mm lifts, but usually that is all. This is best visible in the top-right corner here. The pits caused by the salt are also well visible here: it looks a bit like corrosion. Another inverted test prism. Corners did lift, the center didn't, and the model could be completed. So this is what you could expect for normal PLA and PLA/PHA. With my printer and materials, bonding is best at a glass temp of 60°C. Usually I can lift the whole printer when pulling on the model. At 40°C models are likely to pop off suddenly in mid print. At 70°C the models stay too soft (this is too high above glass transition temp of PLA), and they tend to peel off. At room temp of 20°C bonding is nihil. So I always leave the glass plate in the printer and just wait untill it has cooled down. I would very much welcome feedback on other materials, especially those related to PLA, such as Polyalchemy Elixir, Tough PLA, high temp PLA, wood- and other particle-filled versions, and similar. Also on PET and polyesters, although I don't expect it to work well there.

Or maybe you could encapsulate the 3D-print in transparent PMMA (plexiglass)? At least, PMMA can be sanded and polished to a very smooth finish. And it is water-clear and water-tight. So it will be less likely to cause infections due to bacteria growth. Note that silicone is *not* water-tight: water vapours and other gasses go through it. And so do liquid oils, parafines and solvents. I don't know what effect that would have on a prosthetic eye, or on bio-compatibility? The only reason silicone looks water-tight for liquid water, is that it repels water. But it does not repel oils and solvents. I had liquid parafine (from candles) leak through silicone cups, to my surprise. This porosity is also the reason why you should thoroughly wet a silicone mould with silicone oil, prior to casting solvent-containing resins in it. If the mould is already saturated with silicone oil, less solvents can seep into it, and cure in there and destroy the silicone. So the mould will last longer. To make a silicone mould, you could start from a finely polished and smoothed 3D-printed eyeball as master. Google for: "encapsulate in acrylic -nails" (without the quotation marks, but with the "minus nails" part, otherwise you only get weird fingernails...) In a totally 3D-printed eye, you are definitely going to have trouble with bacteria growth in all the little holes inside the model, and in the layer-lines on the surface.

Nice print. Let us know how it sounds when finished. Out of curiosity: what is the purpose of the snake-house shape (NL: slakkenhuis vorm)? What does it do to the sound, compared to a traditional speaker in a square box?

When using rubber mats or pads, you need to find a mat that does dampen vibrations, but that does not make the printer come into resonance. The rubber should not act like a spring. The rubber density also needs to be appropriate for the printer weight: heavier equipment requires heavier mats, and vice-versa. A 3D-printer is rather light. I think it is best to try various consistencies from softer to harder rubber, and then compare the results. Maybe cork might also be an option. The damping mats I have under my heavy hydraulic pump (400kg) contain a layer of hard rubber with ribs, then cork, then rubber again. They give excellent damping for the pump, but they would be way too stiff for a 3D-printer. Also, some damping mats have ribs to improve grip and damping, so the equipment does not slide around.

I am not related to Ultimaker, and not involved in software development. So this is only a general answer. It is not only the slicer that has to be able to handle the file, but also the operating system and the computer hardware. If the computer can't get the whole file in memory, or if the operating system does not allocate enough memory, it needs to temporary swap parts of the file to harddisk. This may slow down things by a factor 100. Sometimes memory-requirements might be 2x or 3x file size, depending on the software and operation. For example if the software needs to store the original file in memory, plus the temporary calculations, plus the resulting file. Further, it makes no sense to provide more details than the printer's resolution, since microscopic details can't be printed anyway. Personally, I think reducing the file to a grid of 0.1mm or 0.2mm would make no visible difference. When designing small text, I found that all fine details were lost anyway, so I could as well leave them out from the design. I believe you could bring this file down to let's say 50MB, without much visible difference in printing. Maybe you could cut out a small piece of this model, and try to reduce that to different resolutions, and compare these in a test print?

If I remember well, user |Robert| did some tests and got good results; the best I have seen on this forum. Maybe you can find that post? With PET (brand ICE) I can only get a sort of "frosted glass" look, but not transparent. I got the best results when printing slow, so that the material has enough time to melt well in the nozzle, and then to flow into all corners while printing. And the material should be very dry, so you don't get entrapped bubbles because of steam formation. Printing in thick layers also helps, but then you need to print even slower to get it to flow well. Reducing fans also helps, so the next layers bond better to the previous layers: this makes it stronger and lets light through better. A bit to my surprise, printing at the cooler edge of the temp range, also improved transparency, but I am not sure why.

For bonding PLA to the glass I use my "salt method". In short, it goes as follows: First, thoroughly clean the glass. Then clean again with tap water only (no soap, no window washer, nothing). Then gently wipe the glass with a tissue moistened with *salt water*. Gently keep wiping, until it dries into a thin, almost invisible mist of salt stuck to the glass. This takes less than a minute, and can be done while the printer is warming up for the next print. For PLA and PLA/PHA, this "salt method" gives very good bonding when the glass is hot, but no bonding at all when cold at room temp. So models come off by themself after cooling down, without any difficulty nor damage. For Ultimaker PLA and colorFabb PLA/PHA it works very well. But for ABS and PET it does not work. Also, for high but thin objects, like vertical lantern poles, it is not optimal: the salt is not flexible and does not absorb shocks very well, when the nozzle bumps into a model. But since my models are mainly large and low, this is of no concern for me. I have no idea if this would work for tough PLA, but it might be worth trying? If you try, use the damaged side of the glass for your tests. Feedback on the results would be very welcome. See the full text here (yes I know, the text is way too long and I should update it, but it is still usable): https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

The top side of the testmodel looks like this. It is some time ago since I did the smoothing with acetone, so I don't remember the exact details. Today I only did the heat gunning. But I sure did apply quite a lot of acetone back then, and it pooled in these pits in the arcs. So it is probably the excess and the pooling that caused this. I also saw the same effect earlier when using some glues, to glue silicone onto the model. Brushing off the white residu was quite easy, and I did it when the model was still somewhat soft, but not too soft anymore. I remember that I used a brush that was not too hard, otherwise it would still cause scratches. It was a hand brush with black hairs, like used in kitchens. (Those with white hairs, probably nylon, were too hard.)

Although it did work well on thick solid plastic parts like this dental appliance, and on PVC bottles, do *not* use this heat gun on 3D-printed parts, it will most likely destroy them. Heat travels way too fast and too deep into the part, and it makes entrapped air explode, causing bubbles and craters. See my tests and photos here, where I compare it to Cloakfiend's acetone smoothing: https://community.ultimaker.com/topic/8530-acetone-finishing-on-pla/?page=21 Seems nothing beats acetone smoothing, at least on my test parts... :-)

Finally I got my heat gun working again, so I tried "heat gun smoothing" a few test models to compare it to untreated and acetone smoothed parts. The heat gun is a small gas burner with catalytic combustion: it burns without flame on a platinum catalyst. Heat can be regulated from "rather hot" to "very hot"; I don't know exact temperatures. Tests were done on quite low setting. This type of gas heat guns is also used for soldering, and for shrinking socks in electronics. The translucent green 3D-printed model is ICE PET. All orange test models are colorFabb PLA/PHA, Dutch Orange. In the pictures, the dull left model is acetone smoothed, the shiny center model is heat gun treated, and the right one is untreated. The soft green items left and right of the orange parts are silicone impressions of the acetoned and untreated models (I don't have one of the heat gun treated part). Results: - The heat gun almost immediately melts the PLA/PHA. Tiny strands and edges tend to get brown after only one or a few seconds. - Contrary to most plastics, heat seems to travel very fast and very deep into the PLA/PHA, making the whole model deform very soon, after only a few seconds. These models are 10mm wide and 3mm high. So, on thin edges it will be worse. - Cloakfiend's method of "acetoning", does smoothe and round the edges, closes gaps, and gives a somewhat dull look to the model (this is after wiping the models with paper tissue and soft brush to remove the superficial white deposits). - Heat gunning does not close gaps at all, it adds a lot of gloss, and gives an irregular look to the surface. It does not improve edges, contrary to my expectations. (But the facts are what they are...) - Heat gunning melts the top layer, and then the entrapped air below suddenly expands and "explodes", causing bubbles and craters in the surface. This goes very fast. This is best visible in the green PET, although hard to photograph due to the high gloss. This is also contrary to heat gunning solid injection moulded or vacuum formed parts. - Heat gunning smoothes fine scratches from grinding, it gives some shine back to the dullness. Conclusion: Do *not* use a heat gun to smooth PLA and PET models. It is very likely to ruin the model, by melting, burning and deforming it almost immediately, even on low heat setting. Heat travels way too fast and way too deep. And it does explode the entrapped bubbles inside, a bit similar to printing nylon that is too wet. So, while we successfully used this heat gun to smooth dental models in soft plastic, it seems it can not be used for 3D-printed parts. Maybe the only use of the heat gun could be to smooth scratches from sanding. But only *very fine scratches*, showing up as dullness. It can't remove deep scratches. And even then the risk of deformation is high. Instead, use mechanical treatment first to remove big defects, such as cutting and sanding, and then try acetone smoothing. But try this on test parts first. Small gas heat gun with catalytic no-flame burner. Exploded bubbles in PET, similar to 3D-printing too wet nylon. Left = acetone smoothed | Center = heat gunned | Right = untreated. All colorFabb PLA/PHA. (The green parts are silicone impressions of the acetoned and untreated items.) Acetoning closes gaps, heat gunning doesn't. Showing the gloss, or lack of. The typical superficial white deposits on the acetoned part, were brushed off with a paper tissue and a soft brush. Focus on character edges. Burn marks on the heat gunned part, both in front and below the "t" in "bite". Damage showing up only after a day on the acetoned part. Acetone also seems to penetrate quite deep into this PLA/PHA, and to work quite long. The opposite side has indentations, which are shining through now. So testing and not overdoing seems important.

I am totally happy with the current answer, so for me you don't have to make another Zbrush-video. Thanks. I was just curious because in the Skulptris version I once tried, going back from asymmetrical to symmetrical didn't work well indeed. And I believe Skulptris is related to Zbrush. That is why I was wondering. But currently I have neither of both, and no immediate plans. For the moment all of my models are technical models, so I use DesignSpark Mechanical only now. Maybe I will try to model airplanes later on, but then it will probably in Blender, to do the animations and unwrapping too.

This one looks way bigger than he is. Impressive. Out of curiosity: when sculpting faces, do you usually keep left/right symmetry on, as in the video a few posts above? If you switch symmetry off, can you later switch it on again, without destroying the model? How does it react then? Then just a remark: on the test models I acetoned a month ago (also orange colorFabb PLA/PHA), at the beginning I noticed nothing special, apart from the nice smoothing of course. But a week later, I saw that in very thin areas the bottom had gotten deformed a little bit, with slight indentations and wrinkles. Like model airplanes or HO-scale houses can sometimes deform too after using too much glue containing xylene (a long time ago, I don't know if they still contain xylene, as it may cause cancer). Also, my testmodels tended to curl up a tiny bit, only 0.5mm, but still visible. They were sort of rulers, 10cm long x 10mm wide x 3mm high. So the acetone seems to penetrate deeper and work longer than we might think. Haven't you had problems with that?

If you are talking about sort of irregular ribs along the height, could it have to do with the Z-axis (dirt, play,...)? Or with the heat of the nozzle (and thus changes in viscosity)? Or related: with other currents influencing the temperature measurements (e.g. the high current of the build-plate heater causing a shift in ground level of the heat sensor, and thus tiny errors in measurment)? Or with the filament sliding smoother or less smooth while unwinding (thus less or more resistance, and flow)? Or something along this lines? This is guessing, but somewhat educated guessing. The deformation on the bottom seems to be elephant feet, probably due to the heated buildplate and the first layers sagging a bit, or to the squeezing of the first layer onto the glass and spreading a bit. Or both.

The SMD heat gun can melt solder, so it could work for this too. Definitely worth trying. Try setting it at maximum temperature, and a low airflow (if adjustable), and apply from a very close distance, but for a very short time. The blue material we used was fairly soft, and feels a bit like vinyl, or like soft PVC tubing. The gun worked well on old PVC drink bottles too. Not sure if it will work for PET, since PET bottles are much harder to melt and form: it doesn't melt as locally as PVC. PVC-bottles had a small transition temp range from hard to molten, so it would very suddenly melt where heat was applied, while the rest stayed stiff. But PET has a much wider transition temp range in which is it rather rubber-like, and even when molten it doesn't really flow like a liquid. Polystyrene might also be an option, like in model airplanes and HO-scale trains and houses. This is commonly used for packaging and drink cups too. PLA could also be an option, if you can find that in sheets. Some of the drink cups we had here, were PLA, so it must exist somewhere. Probably you know, but for good vacuum forming, provide enough holes in the mould model, so the air can escape when sucking vacuum. Otherwise the molten plate will not get into the details and you get bubbles. Don't ask how I know. :-)

Found a picture of the old heatgun. On this sort of plastic, which is vacuum thermoformed over a gypsum model, you could get a very nice gloss with the heat gun. The edges were first cut with special scissors, then shaped with a Dremel-like tool, and then heat-gunned. Our model-edges were a lot thicker than yours. So in your case the balance between smoothing the edges only, and melting and destroying the whole thing, will be much more critical. This was about the distance we used, ca. 5mm. This brand of gun was not very good: it was too hard to light. Today, I can't even get it started anymore, even not with a flame or bunsen burner. Seems like the catalyst has burned up...

I haven't printed with any other PET, so I can't really compare with any other PETs. It appears similar to NGEN, except that NGEN is better in bridging gaps. But strength and flexibility are similar. Compared to PLA: - For PLA, I use my "salt method" for bonding to the glass (=wipe the glass with a tissue moistend with salt water, that is all). But this does not really work for PET if I need a lot of cooling, so I need to use another method. - When using dilluted wood glue as bonding method, this PET bonds very well, but tends to tear pieces of glass out of the glass plate. Already while cooling, even before I start pulling it off... - So I usually print it on bare glass, which works well, but then I need to reduce cooling to avoid edges lifting on big models. - Without cooling, it is difficult to bridge gaps: the filament tends to curl up in a ball on the nozzle, instead of pulling a nice string to bridge the gap. So top layers of hollow models are difficult to close. - It tends to create a glossy residu in the nozzle, sort of hard "varnish", which is hard to clean. While PLA tends to create a black carbon layer in the nozzle, which comes off easier. - It has a nice glossy finish. - Its transparancy (frosted glass look) is good for embedding logos and text as watermarks in the model. - Layer bonding is good, even at the lower edge of its temp range. I usually print at 220...225°C, 25...30mm/s. - It is more flexible than PLA, so it works a lot better for snap-fit lockings, key rings, and similar things that need to flex a bit. In PLA they break after a few months as it gets harder and brittle. Not so in this PET. - PET models survive in a car in the sun, contrary to PLA. - Strength is quite similar to PLA and NGEN, not great, not bad. - Bottom layer towards the glass is almost perfectly flat, high gloss, and reflects like glass. - Seems to have a bit less creep when under continuous load than PLA, but still does deform, like almost all plastics. So in general I think it is okay, not great but not bad either. I use it mainly for models that need a higher temperature resistance, that need some flexibility, or that need some transparancy. But the majority of models I print in PLA or PLA/PHA. If it wouldn't leave the hard residu in the nozzle, and if it would close gaps better, I would use it more. These items are in ICE PET (all text are hollows totally inside the model, text characters are 3.5mm caps height x 2mm wide x 1mm deep, sitting 0.5mm below the surface): Actually, the "top side" is the bottom side while printing: this side is on the glass.

The weird "worms" or "insect-antennas" that you see, probably come from the nozzle leaking while traveling through air. Upon reaching the other wall, the leaked blob is deposited on the side of that wall. Then next time, the blob is deposited on the already existing blob. So you get a sort of worm. Brown spots are usually molten material that accumulates on the outside of the nozzle, and then burns a little bit and discolors. Then it sags down and is deposited somewhere on the print in a big brown blob. At least, that is what happens in my printer when printing PET. The holes, I don't know. Maybe material that has leaked away during traveling through air, so there is not enough on the first moments of printing again? Or too high retraction settings? But this is just guessing... Try a small test piece, and carefully watch what happens. In my prints, I can reduce the first two phenomena by printing slow and cool, near the bottom of the range. Typically 220...225°C and 25...30mm/s, but I can not totally eliminate it. But of course, it could be different with other brands and colors of PET, and other printers. I have UM2, and the PET is from the brand ICE.

I found that when setting the bed temp to 90°C (for PET), it only reached 80°C in the area where it was printing. But it did reach 90°C in the other corners, where it was not printing (it was small model). The cause was that the air from the nozzle-cooling fans cooled the glass too much. And glass being a poor conductor, it seems the heat can not keep up traveling upwards from the aluminum baseplate fast enough through the glass. This was measured with an infrared thermometer gun, which I found in a hardware shop for ca. 35 euro if I remember well. I am not saying that this is your problem, but it might be worth checking if you have an IR-thermometer available.