geert_2

I am not really a fan of browser-based GUIs, because that leaves you with an additional variable that you are dependent on but have no control over: the browser. Here are a few examples. Firefox was excellent until one day they changed the whole UI and concept, after which it broke all add-ons and became useless for 80% of its users. This broke a lot of people's workflow. Advanced users - like most people here are - tend to install a lot of add-ons in their browser, which may create additional dependencies and trouble. Some people - or some of these add-ons - may disable java, javascript, flash, silverlight, active-x, cookies, external fonts, third-party images, right-click functions, pop-up functions, resize-functions, and whatever else. If you have a good but not very common browser, like Pale Moon (=a Firefox derivation that has kept the old GUI-concept with menubar and statusbar), then this is often not recognised by the server. And then the server messes-up its webpages by *assuming* stupid things, for example that I have a micro-screen of 320x240 pixels instead of my real 1920x1080 pixels. So it sends me garbage instead of standard HTML: it sends fonts of 5cm high, so only a few lines fit on my huge screen. The bigger the organisation, and the more they are specialised in communication (e.g. news-sites and newspapers), the worse this gets, and the less they communicate. A lot of modern browsers even mess-up perfectly valid and simple standard HTML, which by design should reflow automatically in the available window. The browser should take the default font-settings if not specified, without changing them. But they don't. For example Google Chrome Mobile rescales some paragraph's font-sizes (sometimes making it larger, sometimes smaller), but not other paragraphs. And some browsers refuse to reflow text, so it falls off the screen. So you can't even limit yourself to old-school 1995's HTML and forms, because even these break today. You don't want that kind of trouble in a slicer GUI. Cloud-based computing is even worse: then you become dependent on a very unstable variable: the internet/network, coming with all its interruptions and its hazards (virusses, spyware, interception, industrial espionage...). It is unusable while moving (train, plane) or in remote areas: even Germany has no internet in lots of its eastern rural areas. And the data going over their monthly limit, and you going over your budget. Also, this creates GDPR and similar legal problems. So I prefer independent standalone applications installed on and running on the local computer. One application per function. Preferably with all user-settings stored in the same directory as the main program, or a subdirectory "user-settings". Not splattered all over the harddisk in unaccessible directories. So that it is portable. Although of course all programs should use generic and standard datafiles for smooth data-exchange. I am aware that my view may not be "politically correct", but this has proven to work best (for me).

I guess it must be something somewhere in your design that has such a big shape? Maybe a part way out of view on top? Or incorrect printer head-size settings, or an incorrect printer, so it thinks it has a very large print head? So it will stay clear of the edges, to not bang the head into the wall? Something along this line of thinking?

My guess is one of the following: - overextrusion, - blue tape coming off the bed, with air being entrapped below it, - the filament not sticking to the tape for whatever reason. If you watch closely while it is printing, maybe you can see what happens.

I once got wounded by a tiny 50mm computerfan running at overspeed, so the blades were ripped off and flew all across the room. And that was a standard injection moulded glassfiber reinforced nylon fan... (It was running *way* overspeed...) So, 3D-printing any propeller- or turbine-like blades looks like a risky business to me. Airplane propellers and turbines are carefully X-rayed for cracks and voids. FDM 3D-prints are by concept full of such "cracks" due to the way the molten "sausages" are laid down. Also, the surface of the blades needs to be very smooth to get a good airflow, and the airfoil shape needs to be very accurate, both which you won't get with a 3D-print. I think you would be better off making a mould from a real model (thus a duplicate), or if it is a new design, from a carefully post-processed and smoothed 3D-print, and then cast it in strong PU or whatever composite you want. If you cast in a transparent material, you can see any bubbles. If you select a slow-curing resin, you can evacuate the bubbles by vacuum and shaking or tapping. If the mould is big enough, you can also add high-strength fiberglass or other fiber reinforcements, around which you cast. Just like in real airplanes or good safety helmets. Plastic injection-moulded safety helmets are worthless crap, but the good helmets made from composite-resin impregnated fibers can be hit by a heavy hammer several times without fracturing. They don't come apart, which is what you might want.

Actually I have no idea of the percent: I take an old glass bottle, fill it half with water, and just pour in quite a bit of salt. Enough to make it taste really salt, but not so much that it does not dissolve anymore. It is not critical. And if kept in the fridge, it stays good for months. Concerning the rest of the settings: this is hard to say, since it is influenced by materials, model, circumstances and personal preferences. For me the standard settings work pretty well: 0.1mm layers, 60°C bed, 200...210°C nozzle, 50mm/s. For very small objects that need better quality, I print slower and cooler than standard: 190°C, 25mm/s. I rarely print large objects, but if so, I often use 0.2 or 0.3mm layer height, 50mm/s, 210...220°C. Also, overhangs come out better with thicker layers. Just try what works best for you.

If it is just a cap, why not leave out all internal extension and pin holes? So you only have the outside? But I think you would best print this in a softer material. An alternative might be to print a small cup, sacrifice a real connector and clamp it in that cup, and then pour silicone moulds (non-stick model-making silicone) of this? Then the silicone plug will have the correct dimensions, and it is flexible.

The first layer may be too cold. Hotter gives better bonding. Also, it needs to be squished well into the glass, thus into very thin flat layers, not round sausages. You did not mention if you used any bonding agent, so I guess not? Some people have excellent results with 10% dilluted wood glue in water (user gr5's method), some with the standard stick (sometimes with spreading the glue with a wet tissue), some with hairspray (user neotko). I prefer my "salt method": after cleaning the glass, then clean again with pure warm tap water only. And then wipe it with a tissue moistened with salt water. Gently keep wiping while it dries. This leaves a thin mist of an almost invisible layer of salt stuck to the glass. It greatly improves bonding when the glass is hot (compared to printing on bare glass), especially in moist weather when bonding is low otherwise. After cooling down, the models pop off by themself. For the next print, just wipe the glass again with the salt water. No need to take it out anymore. For me the ease of application and the ease of taking the models out, makes this attractive. It works very well for PLA, and for my low and wide models. But I do not recommend it for narrow and high models like statues or lantern poles: these tend to get knocked over. For these, use a glue that can absorb shocks better. Also, it does not work for ABS, PET (it makes PET a bit less sticky). For my old but still usable manual, see: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/ The model at the front is freshly printed. The one at the back was heated in an oven at elevated temperature (70...80°C), to see how it would warp and shrink, which it did obviously, thereby showing the huge internal stresses. It prints fine without lifting corners at 100% infill. Below is a small test model I use to try which bonding methods work well. It excerts huge warping forces due to the large overhangs and small bottom area, so good for testing and comparing. Stay with the printer, as this is likely to come off and produce "spaghetti". Dimensions of my model above, to give an idea.

Recently I tried glueing PP-straps back together after cutting them. It were injection moulded parts, not printed. You can bond them with cyanoacrylate, after roughening up the surface by sanding, and after using an "activator". These activators come in little tubes similar to a fluo pen marker with felt tip, except that the liquid is not ink but a chemical that partially dissolves the plastic (at least I think so, but I am not sure how it works chemically). Then apply the cyanoacrylate and *immediately* press parts together. The activator makes the bonding stronger and also speeds up curing, so you only have a few seconds before it is too hard. I found these in a normal super market (such as Delhaize, Carrefour) and hardware shops (such as Brico, Hubo, and similar here in Belgium). Bonding strength is comparable with older glues like contact glues and "universal glues". But it is *far less* than what we got used to today with cyanoacrylates and composite glues on good substrates. So under high loads and especially when peeling, it may fail. Depending on your application - if purely decorative, light loads or big surfaces - it can work. Under high loads, I think you would be better off screwing, clamping (e.g. snap-fits or dovetails), or welding parts together. An activator also improves bonding strength of PLA with cyanoacrylate, but the working time is reduced so much that it gets quite uncomfortable. You don't have time to align parts correctly anymore, they stick immediately. And if you are too slow, the glue is hard before even touching.

What I was thinking about, is to design and print a small item like this (the square baseplate is 50mm x 50mm). The 45° overhanging edges are likely to curl up (at least with PLA, at 100% infill, and on my printers), and cause the nozzle to bump into them violently, creating a hard test. Then dial up printing speed until it starts causing trouble and skipping steps. And then dial down until these go away. That will give you an idea of how fast you can maximally go. And then reduce further for a safety margin. Maybe that could work as a test?

If your printer has a front panel, then maybe you could change the speed while it is printing, without having to go into the gcode? Like we can do on our Ultimakers? But I don't know your printer, so I don't know if it has this functionality?

I don't know the cause of the slicing problem, so I will leave that to others. But even if you could get it to slice correctly, the 6 tiny holes will be way too small to print, I think: they will get closed down. Also the circular gap around the holes will become too narrow to be functional, if it has to be functional and not only visual. (It looks like a cap for a DIN-connector or something similar?) Unless you would use a 0.25mm nozzle at low speed, temp and layer heights, then you might have a chance.

Have you tried designing a custom brim in the CAD file itself, and maybe some sort of tree-like or inverse-tree supports? Just to add more stability? I guess it now starts wobbling around due to the brim or supports not being stiff enough? An added issue could be not enough cooling, due to the hot nozzle sitting continuously on a small area, which tends to increase this sort of effects. So, adding a dummy cooling tower or printing multiple at once might improve this. But it might also worsen the situation if during traveling around, the nozzle would hit the legs...

If you are printing near your printers maximum, then the slightest bump can cause skipping steps. Overhangs often tend to curl up, so that can be enough to trigger it. At slow speeds the force required to accelerate and move is far less, so the printer has more margin before it skips steps. Maybe try a small test print with overhangs and other difficult stuff, and on the fly change settings up to the point where it starts skipping (or underextruding or whatever else problem comes first)?

That is very severe underextrusion, which can have a ton of causes: partially blocked nozzle, worn-out teflon coupler, bad little nozzle fan, dirty feeder, too wide filament, friction in feeding traject, printing too cold, too fast, etc... Search for the whole list made by user gr5 on this forum (I don't know it by head), which has excellent info.

Yes, I have seen it, good. Reachability is key. :-)

At thicker layer heights, and at higher speeds, you have to extrude a lot more material in the same time, so internal nozzle pressure will be a lot higher. So there will be a lot more "overpressure" stored in the feeding traject that has to be released when the printer slows down. When printing slow and in thin layers, there is very little pressure build-up in the nozzle. I guess...

I have tried annealing PLA (Ultimaker) and PLA/PHA (colorFabb), by very gradually increasing temperature during the course of several hours: 50...60...70°C in my laboratory oven (=Binder: incubator with range up to 99°C). Especially the Ultimaker Pearl filament gets clearly harder and stiffer, and the sound when dropping it changes in pitch: it gets a higher and less dull pitch. This gave maximum 10°C higher temperature resistance, thus still not enough for use in the car, nor for letting it sit in the car in hot weather. That is why I tried it, to see if PLA prints and demo-models would survive transport and storage in a car in summer. Not. Also, during heat treatment, the models tend to warp, and shrink in length. So they need to be clamped down. And mating parts will no longer fit. It might be usefull for artwork to releave some of the internal stresses, but not for mechanical objects with precise dimensions. So for PLA it isn't worth the effort for me. Printing in PET is a better option. I have no experience with nylon or high-temp materials. Light-curing 3D-printing materials used in optical printers (with lasers or beamers) get a lot stiffer after post-treatment. During the initial curing phase in the printer, only a portion of the material reacts to the light and cures. During the post-curing treatment (sometimes heat, sometimes UV-light) a lot of the remaining uncured resin also gets cured. The result is that it is far less susceptible to creep deformation under load, but it can get very brittle. If you do not have an oven, you can use your 3D-printer for annealing by putting the model under a cover, and let it sit overnight with the bed at elevated temperature. If you are not sure if annealing works for you, it might be a good idea to try it in this way first, before investing in an oven. If not done very carefully and slowly, or if going over the limit, the models will soon warp (the top one was at 80°C, PLA, just for testing). Sometimes they first warp upwards, and then after a couple of hours start warping in the other direction (second one). Weird, and I have no idea why. This sort of models will no longer fit and slide well after annealing: the stem of the spoon shrinks in length, but expands slightly in width and height. Also, the ruler in mm is no longer correct. The Binder oven I use:

My personal viewpoint on software - any software - is to keep using it as long as it is stable and I am happy with it. I can do everything I want with an old MS Office 2000, so if I wouldn't need to be compatible with others, I would still be using that since I like that way more than the horrible "Ribbon" mess from Office 2007 onwards. Idem for a very old image editing program and a couple of others. I keep using old hammers and screwdrivers too. :-) Only things like VLC and browsers need to be updated to get the best performance (video formats) and best page display and security (browsers). Concerning slicers, it might be a good idea to keep a couple of stable old versions on your computer, and install a newer version parallel to it, if you need some of the latest features for some prints. Then you have the latest stuff, plus a fallback. Especially if you have a printer with dual nozzle, the area where the most progress is. But that is just my personal preference, based on my needs. Feel free to see things differently.

This is fairly common with polyesters (PET, CPE,...). Some people advise to use glue, in the hope that the glue layer will break instead of the glass. But for me that didn't help, my glue was too good. So, just like you, I also heard the glass cracking while it was still cooling down. Now I use my "salt method": wipe the glass with a tissue moistened with salt water prior to printing, which leaves a thin almost invisible layer of salt stuck to the glass. For PLA the salt method greatly increases bonding (compared to printing on bare glass) when the glass is hot, but there is no bonding at all when cold. But for PET the salt method slightly reduces bonding, so the glass doesn't crack anymore. Now the model separates from the glass in the salt layer, as desired. Disadvantage is a bit more risk of warping, so I print with no cooling fans. I only use glue now if I need a lot of cooling fan for printing overhangs.

I don't know your printer, but this looks like too much mechanical play in the system? Things wobbling around? But I would rather guess the X or Y-axis.

If you want very accurate temperature control, you might want to go to laboratory equipment: there do exist lots of ovens and incubators (=breeding machines) in several temperature ranges. I have one that goes up to 99°C, and one that goes up to 200°C. The point is that they shouldn't have a too big overshoot, which would melt the model. Some models have computer control, where you can set desired curing times and temperatures in stages. Also, dental lab equipment might work, such as the ovens used to cure prostheses which can be pressurised. But I don't know how stable control is. But these are all expensive. Maybe there do exist hobby ovens for moulding and casting too? Try googling for that? If have seen big vacuum equipment and mixers for hobby use, so it would surprise me if ovens would not exist. Some silicones, epoxies and poly-urethanes need heat to cure.

I do have experience with moulds for silicone printed in PLA, but not in PVA or breakaway. The biggest problem, even with "correct" moulds with slanted sidewalls and no undercuts, are the layer lines. The silicone gets stuck into these, causing a very firm grip, which makes it hard to remove the model from the mould. You might also have this problem with breakaway. Most silicone can withstand 250°C for a while, so another option would be to use PLA, and then heat it up and peel away or melt away the PLA at around 100...150°C? If you would use PLA, "smoothing" is highly recommended. This reduces the undercuts in the layer lines, and tends to fill and close tiny gaps. See the thread of user cloakfiend on this forum, where he has done an enormous amount of tests, with excellent results. I don't know if breakaway can be smoothed. If using PVA, smoothing should be possible with water, so you might want to try that on a test piece: it will probably be required to close the gaps and remove irregularities, because PVA seems to print less smooth than PLA. Further, there do exist dedicated mould-making filaments, but I have no experience with them. This you will already know, but I add it for other people not familiar with mould making or silicones: if you make a mould from multiple parts (the A and B side of the mould), make sure to seal all gaps, because the silicone will leak away through the tiniest openings, even if only microns. And provide a way to let air escape while pouring or injecting the silicone. Entrapped bubbles are another common problem.

I haven't used TPU yet, but I vaguely remember others saying that soft plastics are hard to print, because they tend to get compressed and stuck in the bowden tube. Some say you need to print it *very* slow (to not compress it), and some oil the filament (to reduce friction). But I don't know the details, so try searching on the forum. Also, a nozzle temp of 135°C seems very low? Could that be a typo, and should have been 235°C? If you would really print at 135°C, that of course would cause underextrusion.

Actually, I sanded a standard injection needle of 0.41mm diameter down to 0.39mm, so I could use it to poke into the nozzle from the bottom, to clean it. That was its original purpose. After sanding, that needle's diameter gradually transited from 0.39 to 0.41mm. Measured with a good quality Mitutoyo digital calipers, +-0.01mm. Also I cut off and rounded the sharp tip, to prevent damaging the brass nozzle (see photo below). Even the soft steel of this injection needle is much harder than a brass nozzle, so you could easily damage it. The ability to measure the nozzle-diameter was a nice side-effect. :-) First I measure the needle diameter at several spots, and then gently (!!!) push the needle into the nozzle and see how far it gets. At the very beginning, I could only get the nozzle in once it was 0.39mm. Even 0.40mm didn't go (+-0.01mm). Now, a couple of years later, it easily goes in all the way up to 0.41mm. So the nozzle clearly has worn-out a bit, from mainly printing white PLA, which seems slightly more abrasive than other colors, maybe due to filler particles? But it still prints fine. But any thin wire strands, like copper wire, should also work. However, for calibrating the flow, I just dial the flow-rate in until it prints nicely, and that is my setting (e.g. 105% for a certain material). I am not saying this is a good approach, but it works for me. :-) So I don't actually usethe nozzle diameter, nor do any calculations. At the time when I studied electronics (ages ago...), transistor component specs could easily be 100% off, capacitor specs also, and resistor specs were usually 5% or 10% off. So we learned not to care about accuracy while calculating things, it made no sense anyway, but to get the order of magnitude correct. And then we provided good feedback for stabilisation (most important), and a potentiometer for calibration if that would be desired. I still think that is a workable philosophy for subjects with lots of often unknown variables, like 3D-printing. You could calculate flow up to 0.01% accuracy. But then if there is some unknown resistance somewhere in the feeding traject, the indents that the feeder wheel makes into the filament could stretch maybe 20%, from square-shape to diamond-shape, causing an error of 20% in the feed rate. If filament diameter varies with 0.1mm (3.5%), cross-section surface varies with 12%, and so does your feed-rate. So whatever method you use, if it works for you and you are happy with it, I think it is a good method. See the photo for the needle (ruler is in mm and cm): While we are at it: to clean the internals of the brass nozzle, I have a soft brass M3 thread that I use as a file to gently scrape the side-walls. This too has a rounded tip at the bottom to prevent damage. This doesn't really measure anything, but if it has difficulty going through the teflon coupler and/or the brass area of the nozzle, there is accumulated dirt, or a deformed teflon coupler, which could cause extra resistance and negatively affect the flow rate. So this gives a rough indication if the feeding traject in the print head is clear enough to let filament pass easily.

@yellowshark: You can overcool by blowing compressed air directly on the nozzle tip. This cools down the nozzle too much so it can not keep up heating, and generates a "temperature error" (I don't remember the exact description). I had that once in my overhang-tests, where I was testing the effect of additional cooling on the quality of bridges and overhangs. This was on a standard UM2 (non-plus). I think layer-bonding could also suffer, similar to printing with ABS without front door and too much fan, but I haven't tested bonding strength, so this is a bit guessing. However, putting a 30cm desktop fan (like those used in hot summers) at 2m in front of the printer, at its lowest speed, gave excellent extra cooling without triggering errors for PLA. @paulrhee2002: I do see some horizontal gaps, but I am not sure if this would be underextrusion, or the lower layers sagging due to still not enough cooling? I would guess the last. While it is printing, maybe try poking into the model with a tiny pin or screw driver, to see if it is hard or still weak? Keep watching what happens (don't let this print unattended). I still do think two top layers is not enough, unless you use +70% infill. But then that would greatly increase cooling problems, since all that infill-heat has to be removed too. At such thick layers you are faced with the dilemma that you need lots of heat to get enough plastic molten in time, but at the same time due to the thick layers, you can't remove the heat from the print fast enough. When using only two top layers, the first one is going to sag a bit, and the second one can't fill up all the gaps. I would suggest you try 4 top layers and compare. Since this is an area where we have no experience, you will be pioneering a bit. So keep posting your results, they are very welcome.