geert_2
-
Posts
2,084 -
Joined
-
Last visited
-
Days Won
34
Content Type
Forums
Events
3D Prints
Posts posted by geert_2
-
-
A bit off-topic: before printing 230 copies, have you tested if these 3D-models are fully functional and can withstand the desired loads, plus any brutal abuse (dropping, slamming into things)? If they want 230 pieces, it probably is for a functional test, not just for a visual demo...
-
I modeled this test table in CAD in DesignSpark Mechanical (freeware, requires registration).
Dimensions in this test model:
- for reference: most plates are 1mm thick, text caps height is 3.5mm, text legs are 0.5mm wide
- horizontal gap between supports and model: 1mm
- tiny connection strands to keep support attached: 0.5mm wide (for nozzle of 0.4mm) x 0.2mm high (=2 layers of 0.1mm). These can easily be cut off, and cleaned.
- inverse staircase in supports: steps of 1mm. This staircase reduces the overhangs curling up, but does not eliminate it. But in my tests these stairs worked better than a 45° overhang slope.
- ribs on top of support: 0.5mm wide x 0.5mm high x 1mm horizontally separated.
- vertical gap between ribs and underside of model: depends on model, usually 0.2 to 0.4mm. Try what works best for you.
- supports extend sideways, bigger than the model: 0.5mm to 1mm. Making the support a bit bigger than the model improves the edges of the first layer of the model. Otherwise they sometimes sag and fall off the supports.
I think this was the STL-file of the model:
Also, for bigger models or difficult to reach areas: make very sure you implement enough features to insert pliers, hooks, knifes, and other tools to remove the supports. So you don't get stuck with unreachable and unremovable supports. This make take quite a lot of planning in complex models. Here too: the supports took way more time to model and test, than the table itself.
-
Did you write "alignment-marks" on the pulleys with a marker? To see which one, if any, is slipping?
Also check if your print head moves smoothly. If it would get stuck (e.g. no lubrication, dry bearings, or oil that got too sticky), it could cause the stepper motor to skip steps. I don't know if this could also happen if the steppers simply get too hot?
Printing too fast could also cause skipped steps.
-
Like tinkergnome said: also inspect the rods, bearings, rubber bands, nozzle, bowden tube, and rotating knob for wear, and the housing for scratches. My UM2s have 1000...1500 hours, and they still print fine. But you can see some wear on these parts. Depending on the material and temp printed, expect that the teflon coupler might also be worn (but this is hard to see, you would need to do a cold pull and inspect the outcome for deformations).
-
On 2/25/2019 at 11:58 AM, AndersK said:
As I said, very easy to get off once cooled but large flakes of glass stuck to the print.
Just like it chipped due to shrinking when cooling down.
Yes, I have also had a glass chip while cooling: I heard the normal little ticking sound of the model coming off gradually, and then suddenly I heard much louder snapping sounds...
Forcing the model off while the plate is still hot, also seems a no-no to me. That would rather increase the risk, I think.
Try a thicker layer of not too strong glue, that can absorb the expansion/shrinking.
Personally, I use my "salt method" here too: wipe the glass with a tissue moistened with salt water, prior to every print. Let dry into a thin mist of salt stuck to the glass. For PLA this greatly increases bonding to a hot plate, and gives no bonding at all on a cold plate, which is very handy. For PET it does not seem to increase bonding, but rather slightly reduces it, but at least it also reduces bonding on a cold glass a little bit. I haven't had any chips since then. If you print without cooling fans, or with very little fan, it might work. With full fan, bonding might not be good enough and parts might warp. Try this on small test pieces first, and stay around to see what happens, so you can stop the print if it comes off.
-
1
-
-
On 2/27/2019 at 3:17 PM, danielkrice said:
Just as an alternative idea, recess the lettering into the print instead of it being proud, then thickness becomes less of an issue.
Yes, but this might be ugly, since the printer outlines every character. And for letters like E, N, H, etc, the problem is the same: here it may not be able to fill the openings.
Some time ago I made character test sets with extruded text, recessed text, positive watermark text (=characters are solid, surrounded by voids, all below the surface of the model), negative watermark text (=characters are voids). Recessed text came out worst.
For opaque materials usually I prefer raised text of 0.2mm high, leg width 0.5mm, caps height 3.5mm. For transparent materials I use watermark text (=just below surface) of 0.5mm to 1mm high, leg width 0.5mm, caps height 3.5mm. Some of the pictures below show this watermark text.
For my 0.4mm nozzle I prefer text legs to be 0.5mm wide. When converting a curved shape to STL, it all become straight segment lines. If the width of the curved shape is exactly 0.40mm, then in STL it may variate between 0.39mm and 0.41mm. In which case the 0.39mm part might not print. So, I prefer to be safe and take 0.5mm as width in the design.
See here for the character test set (STL-files and JPG-images):
https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/
-
1
-
-
Up till now, all parts printed in transparent PET (brand: ICE, from Trideus in Belgium) and in NGEN (colorFabb) have survived well in my car, even in the hot summer of 2018. These can be printed quite easily, although NGEN is more difficult to glue with my cyanoacrylate glues.
And indeed, parts in PLA warp, even in mild spring or autumn weather.
-
For best quality of the text, I would print it on its back, thus the text facing up. But I would design custom supports, so that I would use far less material, and waste less time. Similar to this test with support bridges shown below, after an idea of user smartavionics if I remember well. I don't know if this feature is already integrated in newer versions of Cura? (I am using an older one.) If yes, try that first: it will probably give a much better undersurface. I have UM2 (non-plus) single nozzle printers, so I can print in one material only (this is PLA).
-
If you suspect "not enough flow" is one of the causes, maybe you could try a (bigger) priming tower, or print a dummy block in PVA next to the real print, to keep more flow going through that print head? So it is always well purged? Try the idea on a small piece first, before doing a large job.
-
I don't have a dual nozzle printer, so just guessing:
When using dedicated support materials like breakaway or PVA, shouldn't the gap between support and model be zero? So that they are squeezed well together for a good bonding?
Maybe you could try this on a couple of small test pieces?
-
What about PET? This prints reasonably easy, although I don't know about compatibility with support material?
Anyway, wasted PET bottles seem to survive very well in the ocean, without much degradation.
-
For PLA, try wiping the glass with salt water and paper. For me this gives very good bonding when the glass is hot (ca. 60°C), but no bonding at all when cooled down to room temp.
Also, make sure the first layer is squeezed well into the glass: the bottom surface should be flat and shiny, like this. This PLA-surface reflects almost like a mirror (the bottom pic shows the cover and fan blades of a fan behind the model, the top pic shows the model itself. It is the same model, but I couldn't get both in-focus at the same time).
-
Wasn't there an option in Cura in which you could shift a second model into the first one, to modify the infill? So, if you would make a second model sitting on the glass, and change that to zero wall width, and x % infill? Then shouldn't that give an infill-only bottom plate?
(I don't have the latest Cura version, so I don't know this, but I vaguely seem to remember reading about it: was that you kmanstudios who used that technique in your gyroid-filled statues? Or was it gr5?)
Second option: maybe in CAD you could use the shell only of your model? Thus turn it into a surface instead of solid, then delete the bottom plate, and then thicken that shell to maybe 0.8mm? And then add a custom brim for stability while printing? I don't know if your CAD program allows this, but it might be worth trying?
Then you could just pour in gypsum. Normal gypsum won't work with a gyroid infill: it is way too viscous. You would have to use epoxy instead. Unless you have a special gypsum vibrator equipment to make it flow better. I am talking about powerfull dental gypsum vibrators here, which can bring a whole wall into resonance, not about the more gentle woman things. :-)
-
1
-
-
In DesignSpark Mechanical this can easily be done in CAD in the model itself:
- Make a backup and work on that (never on the original).
- Select and delete all internal geometry you don't need.
- Select and delete all screws etc. that you don't need. This may leave gaps.
- Select and delete the gaps.
- Select and pull any remaining walls that are too thin, into a thicker wall.
Other method:
- If the model would contain too much internal geometry to delete one by one, you can select only the outer surface and copy that surface only. That will remove all internals and leave you with an unprintable thin shell.
- Then delete all gaps until the interior is one solid block again. This is what I did in the quick test below, from left to right: original with lots of internal geometry, copied surface-only, deleting the gaps, after the last gap is deleted it becomes a solid block again. Since this consists only of selecting and deleting (often possible in bulk), it goes quite fast.
Maybe other editors allow a similar approach?
-
1
-
-
You will definitely need to chemically "activate" the surface before glueing, and use a glue designed for plastics and nylon. Otherwise bonding may not withstand high loads and repeated flexing.
Maybe you could also look into mechanical bondings, in addition to glue? For example dovetails (I hope that is the correct word in English), recessed screws, recessed rivets flush with the surface, sewing (like in the old days of book-binding), or so?
Another option that just comes to my mind: polyurethanes are usually easier to glue than nylon, and some of them are also flexible and durable. Maybe you could ask a couple of samples?
It will take some experimenting and pioneering, but it looks like an interesting project. Lots of people have looked into wings that can deform, but I hadn't seen your approach yet. So, feel free to post the results of how you solved the problems, and what the efficiency of the wing is.
-
1
-
-
You could also try DesignSpark Mechanical, by RS-components. This is a free and limited version of SpaceClaim, and only requires registration. It's user-interface is somewhat similar to SketchUp, and easy to learn, but it is way more powerfull. And I *never* got any problems when exporting to STL. Go to Youtube and have a look at the lots and lots of tutorials, to see if it appeals to you.
Learning DSM will cost you less time than trying to repair a single corrupt STL-file produced by SketchUp, so it is worth the time-investment.
DSM is excellent for mechanical designs based on geometric shapes. But it is not suitable for organic designs.
SketchUp was never made for 3D-printing; it was ment for visual 3D-representations of buildings only (before Google invented its own methods of converting aireal photos automatically into 3D). And indeed, it produces sort of cardboard models that are glued together, with gaps in the seams. Zoom in on text made in SketchUp and you can easily see it.
Example of design in DSM:
-
1
-
-
No experience with nylon. However, I regularly print carabiner hooks in PLA and PET, and use them daily, just to learn how the materials flex and behave on the long term. The PLA ones always crack in the same place. The PET ones are more flexible and survive. Unless I pull too hard, and then they fracture too, usually at random places (not necessarily in the corners).
So, if you would already have PLA or PET, try these materials first, and see how the model flexes and where it breaks. It gives an idea of the stress concentrations and weak points. Then, nylon will just be stronger and easier to flex than PET, and unlikely to break.
Would a real full-scale model also be in nylon? Or rather in composite, thus much harder and stiffer than nylon (in which case it might behave more like PLA or PET, althoug much stronger)?
Anyway, it looks like a very interesting project, which might also be usefull for composite airplanes. A question: how much turbulence do the little ribs create? And doesn't that disturb the smooth airflow and spoil the efficiency of the wing? Or will these ribs be covered with an extra plate later on? Would this flexible shape be more efficient than the crude profile changes achieved by traditional flaps and ailerons
in airplane wings?
Left and center: PLA/PHA-blend (colorFabb). Right: PET (ICE). The PET is more flexible and survives repeated opening well. The PLA/PHA show the usual point of faillure.
-
I don't have an UM3, and never ran out of filament while printing, so I can't comment on that.
But if the printer has been "printing" dry, then it is possible that the last piece of molten filament has been sitting at hot temperature in the nozzle for a long time, and that it got burned, and may have clogged the nozzle. So I would first make sure that the nozzle is not clogged, and thoroughly clean it, before restarting any print.
-
On 2/17/2019 at 3:57 PM, tinkergnome said:
...
- Are you sure that you want to print with only a single top and bottom layer (0.06mm)?
...
I agree on thinkergnome with that question: usually top and bottom should be about the same thickness as the walls, but it depends on the model and material. If too thin, it can't bridge and close the top gaps. So I would suggest you begin with 0.8mm, same as the walls.
Concerning printing speed, start a test piece with 50mm/s, and adjust up and down in steps of 10mm/s on the fly (if your printer allows that), to see how it affects print quality.
Idem for temperatures: start with a safe average recommended for your printer (may differ from brand to brand, model to model, and material to material), and gently adjust up and down in steps of 5°C while watching it.
Idem for bed temp: too hot and the model will sag, or peel off. Too cold, and the model won't bond well and may suddenly pop off. Try to find the point in-between where bonding is optimal, without deformation.
-
I think you would best contact the manufacturers directly, such as for T-glase (is that Eastman Kodak?).
Most of us have tried printing transparent parts, but with very limited success. Due to the inevitable entrapped air, both in-between the extruded sausages as in the center of the sausages, it is very difficult to get a better clarity than "frosted glass" or shattered glass. This is good enough to show a watermark, and let some light shine through. Suitable for model trains, and lenses for rear lights of bikes. But far from usefull for long real light guides and optical datacables. Another problem are surface defects due to layer lines, which act like lenses and which deform the light path. All these defects combined will probably play a much greater role in the disturbance of the ligth than the pure material characteristics. For good transparency, you will have to mould and cast, I think.
See these blocks (20mm x 10mm x 10mm), with a hollow watermark halfway inside, printed at different speeds (50mm/s and 10mm/s, top and bottom) and different layer heights (0.40; 0.30; 0.20; 0.10 and 0.06mm left to right). This material is waterclear PET (brand: ICE, from Trideus in Belgium), but yellowed due to printing too slow at the bottom, thus subjected to heat for too long.
In the second pic, one of the blocks is polished, to remove the outer defects.
-
1
-
-
I have no experience with nylon, but from the recommendations I read here before: did you keep the spool in a dry-box with lots of desiccant during printing? And with a teflon tube from that box to the printer? So that no normal moist air comes in contact with it? It is known that especially in a moist environment, in a few hours nylon can absorb so much water that is can not be injection moulded anymore. I guess the same would apply to 3D-printing?
Of course, theoretically there could be other causes too, such as dust accumulated on the filament and pulled into the nozzle, or the spool tangling up after some time of printing, or filament burning in the nozzle and gradually clogging it, etc..., but I guess you checked for these.
-
2
-
-
I also had this once too. Got it out with fine tweezers. Otherwise, use a stick with chewing gum or sticky wax attached to it. Maybe it doesn't hurt the machine to leave it in there, but it does hurt your peace of mind.
And then print a couple of these. See here, and scroll down a bit:
https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/
-
I have no experience with soap models, but with moulding and casting in general with silicones, yes.
First, I would suggest you start with a simpler model, a dolphin or shark, that can be made in a two-part mould. And then, as kmanstudios said, go to Youtube and search for: silicone mould making and casting
This gives a wealth of information, with hundreds of very good tutorials, and product info.
What I often do is: first print the model in PLA. Then in plasticine (a no-sulphur variant) I model a box closely around the part. And then I pour liquid silicone into the box. After curing, I remove the plasticine, and very carefully cut the silicone on appropriate places. These seam-lines are important: they should be in a place that does not deform the model, and they should allow you to remove the model from the silicone mould. Also, the seams should make it possible to assemble and align the mould correctly later on.
Another option is to use paste-on silicone. Paste that onto the model and let cure. Then paste shell-material onto it: this is a paste that hardens into a very hard plastic, like polyester. Some people use gypsum too. Make the shell so that it can be removed, thus in multiple parts. Include alignment features in the shell, so it fits back together in the correct position later on.And provide channels in which you can pour the soap, and through which entrapped air can escape.
There are lots and lots of things to consider and keep in mind. Making a good mould takes a lot of planning and studying, even for fairly simple models. Hence the importance of the Youtube tutorials: even if you spend a whole day studying them, you will soon win that time back.
-
2
-
-
On my printers overhangs work better with thick layer heights: 0.3mm and 0.2mm are more beautiful than 0.1mm (haven't tried 0.06mm with huge overhangs). In 0.1mm layers, the edges of overhangs tend to curl up.
Try to optimise the settings and design on *small test pieces* first, so you don't waste too much time and material.
And indeed, a round underside might need a brim, or a custom flange designed in CAD, to make it stick.
If it has to be a demo-model, I do understand your concerns about elegance: square tubing doesn't look as nice as round tubing. And especially a house-shape (=square with roof) might look ugly due to the unsymmetry.
Maybe you could try a couple of small test pieces in different shapes next to each other? A round shape with brim, a house-shape (=square with roof), a true pentagon shape, triangular shape? Although personally I wouldn't use a triangular shape, due to the restricted volume and huge inner surface area, causing more friction to the flow.
Then the students can see all the alternatives, and the benefits and disadvantages of each.
To create mold for concrete using PLA
in UltiMaker Cura
Posted
You need to make sure you have no undercuts, and that all side walls have draft (=wider opening at the top). Otherwise the models get stuck.
If in doubt, make the walls out of multiple parts, and provide strong flanges with holes to screw them together.
Provide big openings for pouring in the liquid, and openings for venting air. Make sure no air can get trapped.
Weight could also be a problem in big models: make the walls thick enough, and provide extra supports where required (like the beams on the sides of castles and cathedrals).
And heat development: I don't know how hot concrete gets, but gypsum gets warm and could melt PLA moulds.
For release agents, you will have to google. Not only do you need to find one that works well in combination with concrete, but it also should not inhibit the curing. Try on small test pieces first. Silicon oil might work.